Review of
Marketing Research
Review of Marketing Research Volume 6
Naresh K. Malhotra Editor
M.E.Sharpe Armonk, ...
139 downloads
2985 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Review of
Marketing Research
Review of Marketing Research Volume 6
Naresh K. Malhotra Editor
M.E.Sharpe Armonk, New York London, England
Copyright © 2010 by M.E. Sharpe, Inc. All rights reserved. No part of this book may be reproduced in any form without written permission from the publisher, M.E. Sharpe, Inc., 80 Business Park Drive, Armonk, New York 10504. Print ISSN 1548-6435 Online ISSN 1944-7035 ISBN 978-0-7656-2127-6 Printed in the United States of America The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences Permanence of Paper for Printed Library Materials, ANSI Z 39.48-1984. ~ IBT (c) 10 9 8 7 6 5 4 3 2 1
REVIEW OF MARKETING RESEARCH Editor: Naresh K. Malhotra, Georgia Institute of Technology
Editorial Board Rick P. Bagozzi, University of Michigan Russ Belk, University of Utah Ruth Bolton, Arizona State University George Day, University of Pennsylvania Morris B. Holbrook, Columbia University Michael Houston, University of Minnesota Shelby Hunt, Texas Tech University Dawn Iacobucci, Vanderbilt University Arun K. Jain, University at Buffalo, State University of New York Barbara Kahn, University of Miami Wagner Kamakura, Duke University Donald Lehmann, Columbia University Robert F. Lusch, University of Arizona Debbie MacInnis, University of Southern California Nelson Ndubisi, Monash University, Malaysia A. Parasuraman, University of Miami William Perreault, University of North Carolina Robert A. Peterson, University of Texas Nigel Piercy, University of Warwick Jagmohan S. Raju, University of Pennsylvania Brian Ratchford, University of Texas, Dallas Jagdish N. Sheth, Emory University Itamar Simonson, Stanford University David Stewart, University of California, Riverside Rajan Varadarajan, Texas A&M University Michel Wedel, University of Maryland Barton Weitz, University of Florida
v
AD HOC REVIEWERS Sunil Gupta, Harvard University Dominique Hanssens, University of California, Los Angeles Scott Neslin, Dartmouth College Shuba Srinivasan, Boston University Russ Winer, New York University Lisa Bolton, Pennsylvania State University Meg Meloy, Pennsylvania State University Bill Ross, Pennsylvania State University
vi
CONTENTS Review of Marketing Research: Analyzing Accumulated Knowledge and Influencing Future Research Naresh K. Malhotra Contents, Volume 1 Contents, Volume 2 Contents, Volume 3 Contents, Volume 4 Contents, Volume 5
ix xix xx xxi xxii xxiii
1. A Review of Prior Classifications of Purchase Behavior and a Proposal for a New Typology Hans Baumgartner
3
2. Measuring Customer Lifetime Value: Models and Analysis Siddharth S. Singh and Dipak C. Jain
37
3. Learning Models S. Sriram and Pradeep K. Chintagunta
63
4. Customer Co-Creation: A Typology and Research Agenda Matthew S. O’Hern and Aric Rindfleisch
84
5. Challenges in Measuring Return on Marketing Investment: Combining Research and Practice Perspectives Koen Pauwels and Dave Reibstein
107
6. Service-Dominant Logic: A Review and Assessment Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
125
7. Marketing in a World with Costs of Price Adjustment Shantanu Dutta, Mark E. Bergen, and Sourav Ray
168
About the Editor and Contributors
189
vii
REVIEW OF MARKETING RESEARCH Analyzing Accumulated Knowledge and Influencing Future Research Naresh K. Malhotra
Overview Review of Marketing Research, now in its sixth volume, is a fairly recent publication covering the important areas of marketing research with a more comprehensive state-of-the-art orientation. The chapters in this publication review the literature in a particular area, offer a critical commentary, develop an innovative framework, and discuss future developments, as well as present specific empirical studies. The six volumes feature some of the top researchers and scholars in our discipline, who have reviewed an array of important topics. The response to the first five volumes has been truly gratifying, and we look forward to the impact of the sixth volume with great anticipation. Publication Mission The purpose of this series is to provide current, comprehensive, state-of-the-art articles in review of marketing research. Wide-ranging paradigmatic, theoretical, or substantive agendas are appropriate for this publication. This includes a wide range of theoretical perspectives, paradigms, data (qualitative, survey, experimental, ethnographic, secondary, and so forth), and topics related to the study and explanation of marketing-related phenomena. We reflect an eclectic mixture of theory, data, and research methods that is indicative of a publication driven by important theoretical and substantive problems. We seek studies that make important theoretical, substantive, empirical, methodological, measurement, and modeling contributions. Any topic that fits under the broad area of “marketing research” is relevant. In short, our mission is to publish the best reviews in the discipline. Thus, this publication bridges the gap left by current marketing research publications such as the Journal of Marketing Research (USA), International Journal of Marketing Research (UK), and International Journal of Research in Marketing (Europe) that publish academic articles with a major constraint on the length. In contrast, Review of Marketing Research will publish much longer articles that are not only theoretically rigorous but also more expository, with a focus on implementing new marketing research concepts and procedures. This will also serve to distinguish this publication from Marketing Research magazine, published by the American Marketing Association (AMA). ix
x
Review of Marketing Research
Articles in Review of Marketing Research should: • • • • • • • • • •
Critically review the existing literature Summarize what we know about the subject—key findings Present the main theories and frameworks Review and give an exposition of key methodologies Identify the gaps in literature Present empirical studies (for empirical papers only) Discuss emerging trends and issues Focus on international developments Suggest directions for future theory development and testing Recommend guidelines for implementing new procedures and concepts
Articles in the First Volume The inaugural volume exemplified the broad scope of the Review of Marketing Research. It contained a diverse set of review articles covering areas such as emotions, beauty, information search, business and marketing strategy, organizational performance, reference scales, and correspondence analysis. These articles were contributed by some of the leading scholars in the field, five of them being former editors of major journals (Journal of Marketing and Journal of Consumer Research). Johnson and Stewart provided a review of traditional approaches to the analysis of emotion in the context of consumer behavior. They reviewed appraisal theory and discussed examples of its application in the contexts of advertising, customer satisfaction, product design, and retail shopping. Holbrook explored and reviewed the concept of beauty as experienced by ordinary consumers in their everyday lives. His typology conceptualized everyday usage of the term “beauty” as falling into eight categories distinguished on the basis of three dichotomies: (i) extrinsically/ intrinsically motivated; (ii) thing(s)‑/person(s)‑based; and (iii) concrete/abstract. Xia and Monroe first reviewed the literature on consumer information search and then the literature on browsing. They proposed an extended consumer information acquisition framework and outlined relevant substantive and methodological issues for future research. Hunt and Morgan reviewed the progress and prospects of the “resource-advantage” (R-A) theory. They examined in detail the theory’s foundational premises, showed how R-A theory provides a theoretical foundation for business and marketing strategy, and discussed the theory’s future prospects. Bharadwaj and Varadarajan provided an interdisciplinary review and perspective on the determinants of organizational performance. They examined the classical industrial organization school, the efficiency/revisionist school, the strategic groups school, the business policy school, the PIMS paradigm, the Austrian school, and the resource-based view of the firm. They proposed an integrative model of business performance that modeled firm-specific intangibles, industry structure, and competitive strategy variables as the major determinants of business performance. Vargo and Lusch focused attention on consumer reference scales, the psychological scales used to make evaluations of marketingrelated stimuli, in consumer satisfaction/dissatisfaction (CS/D) and service quality (SQ) research and proposed social judgment-involvement (SJI) theory as a potential theoretical framework to augment, replace, and/or elaborate the disconfirmation model and latitude models associated with CS/D and SQ research. Finally, Malhotra, Charles, and Uslay reviewed the literature focusing on the methodological perspectives, issues, and applications related to correspondence analysis. They concluded with a list of the creative applications and the technique’s limitations.
Review of Marketing Research
xi
Articles in the Second Volume The second volume continued the emphasis of the first by featuring a broad range of topics contributed by some of the top scholars in the discipline. The diverse articles in the second volume can all be grouped under the broad umbrella of consumer action. Bagozzi developed a detailed framework for consumer action in terms of automaticity, purposiveness, and self-regulation. Mac Innis, Patrick, and Park provided a review of affective forecasting and misforecasting. Ratchford, Lee, and Talukdar reviewed the literature related to use of the Internet as a vehicle for information search. They developed and empirically tested a general model of the choice of information sources, with encouraging results. Miller, Malhotra, and King reviewed the categorization literature and developed a categorization-based model of the product evaluation formation process, which assists in the prediction of set membership (that is, evoked, inert, or inept). Lam and Parasuraman proposed an integrated framework that incorporated a more comprehensive set of various individual-level determinants of technology adoption and usage. Recently, marketing has come under increased pressure to justify its budgets and activities. Lehmann developed a metrics value chain to capture the various levels of measurement employed in this respect. Finally, Oakley, Iacobucci, and Duhachek provided an exposition of hierarchical linear modeling (HLM). Articles in the Third Volume Bolton and Tarasi described how companies can effectively cultivate customer relationships and develop customer portfolios that increase shareholder value. They reviewed the extensive literature on customer relationship management (CRM), customer asset management, and customer portfolio management, and summarized key findings. They examined five organizational processes necessary for effective CRM: making strategic choices that foster organizational learning; creating value for customers and the firm; managing sources of value; investing resources across functions, organizational units, and channels; and globally optimizing product and customer portfolios. Chandrasekaran and Tellis critically reviewed research on the diffusion of new products, primarily in the marketing literature and also in economics and geography. While other reviews on this topic are available, theirs differs from prior ones in two important aspects. First, the prior reviews focus on the S-curve of cumulative sales of a new product, mostly covering growth. Chandrasekaran and Tellis dealt with phenomena other than the S-curve, such as takeoff and slowdown. Second, while the previous reviews focus mainly on the Bass model, Chandrasekaran and Tellis also considered other models of diffusion and drivers of new product diffusion. Eckhardt and Houston reviewed, compared, and contrasted cultural and cross-cultural psychological methods. They presented the underlying conceptions of culture that underpin both streams and discussed various methods associated with each approach. They identified the consumer research questions best answered using each perspective and discussed how each approach informs the other. Finally, they examined how consumer research can benefit from understanding the differences in the two approaches. While cultural and cross-cultural perspectives adopt distinct views about culture and psychological processes, it is possible to view them as complementary rather than incompatible. Several suggestions by Malhotra and his colleagues can be useful in this respect (Malhotra 2001; Malhotra, Agarwal, and Peterson 1996; Malhotra and Charles 2002; Malhotra and McCort 2001; Malhotra et al. 2005). For example, one can start with an etic approach and then make emic modifications to adapt to the local cultures. Alternatively, one can start with an emic perspective and then make etic adaptations to get an understanding across cultures. This systematic theory building and testing process is illustrated by Kim and Malhotra (2005).
xii
Review of Marketing Research
Grewal and Compeau synthesized research from consumer behavior, psychology, and applied economics to address how price as an information cue affects consumers’ responses in the context of other information cues. They developed a conceptual framework, using adaptation-level theory and transaction utility theory, that synthesized prior research on price, reference price, and other information cues and their effects on consumers’ price expectations, evaluations, and behavioral intentions. Their conceptual model contributes to our understanding of the way imperfect information affects consumers’ decision processes, goes well beyond the original price–perceived quality paradigm, and integrates knowledge from consumer research, psychology, and applied economics. Sayman and Raju provided a review of research on store brands. Their review focused on integrating research in key areas and identifying directions for future research. There is limited theoretical and empirical research regarding optimal counterstrategies of national brands against store brands; studies tend to focus on one aspect, and national brand quality is typically assumed to be exogenous. Researchers have, by and large, focused on me-too-type store brands. Future research should consider premium store brand products as well. Merunka and Peterson examined an intrapersonal aspect of language, namely, whether the structure of a language, per se, influences the thoughts of those who speak it. They reviewed empirical research conducted over the past half-century on the effects of language structure on a variety of mental activities. They found support for the weak form of the linguistic relativity hypothesis, the notion that the structure of a language does indeed influence (but not determine) cognition. The estimation of independent and joint effects of language is difficult at best. We need comprehensive studies that incorporate the order in which bilinguals acquire their respective languages, how they acquire their languages, and when they acquire their languages. Future research should also compare the possible influence of a single language on mental processing across different cultures. Belk discussed the implications of getting visual for research, teaching, and communicating. He identified basic opportunities, threats, and consequences of becoming visual. Several techniques for collecting visual data were discussed in the realm of interviewing as well as observation. We might well be entering a Golden Age of visual and multimedia marketing research, and Belk helps us to get a good handle on it. Articles in the Fourth Volume Consistent with the first three volumes, the fourth volume also featured a broad array of topics with contributions from some of the top scholars in the field. These articles fall under the broad umbrella of the consumer and the firm. Louviere and Meyer considered the literature on behavioral, economic, and statistical approaches to modeling consumer choice behavior. They focused on descriptive models of choice in evolving markets, where consumers are likely to have poorly developed preferences and be influenced by beliefs about future market changes. They called for a better alliance among behavioral, economic, and statistical approaches to modeling consumer choice behavior. Economic and statistical modelers can constructively learn from behavioral researchers and vice versa. Folkes and Matta identified factors that influence how much an individual consumes on a single usage occasion by drawing on research in consumer behavior as well as allied disciplines. They developed an integrated framework to understand how, and at what stage, various factors affect usage quantity based on Gollwitzer’s (1996) “action goals” model. Initially, factors such as a product’s price and social norms influence consumption-related goals and their perceived desirability
Review of Marketing Research
xiii
and feasibility. In the next phase, factors such as self-control strategies and product instructions influence the implementation of the goal. Finally, the consumer’s motivation to use feedback, and the type of feedback about consumption, has an influence on subsequent goal setting. Kumar and Luo also examined consumption, but from a modeling perspective. In order to allocate scarce marketing resources efficiently and effectively, it is important for a firm to know what to sell, when to sell, and to whom. Kumar and Luo reviewed how the purchase timing, brand choice, and purchase quantity decisions have been modeled historically, as well as the issues within each decision that have been addressed. A vast majority of these studies use scanner data or transaction data. Since recent research has shown that common method variance may not be a serious problem (Malhotra, Kim, and Patil 2006), surveys can also be a useful source of such data and should be increasingly employed. Despite the interest in global branding, studies involving brand extension strategies in foreign markets remain very limited. The fact that so few studies exist limits our understanding of effective brand extension strategy in a cross-cultural context. Merz, Alden, Hoyer, and Desai proposed a new conceptual framework and several propositions regarding effective global brand extension strategy in a cross-cultural context. In doing so, they first reviewed more commonly examined antecedent variables of (national) brand extension evaluation. Then they proposed a definition of culture and subsequently reviewed the existing cross-cultural brand extension research. Given the growing importance of visual marketing in practice, Wedel and Pieters reviewed eye-tracking research in marketing and evaluated its effectiveness. Specifically, they reviewed eyetracking applications in advertising (print, TV, and banner), health and nutrition warnings, branding, and choice and shelf search behaviors. Finally, they discussed findings, identified current gaps in our knowledge, and provided an outlook on future research. Singh and Saatcioglu reviewed different approaches for examining role theory implications for boundary spanners such as salespeople, frontline, and customer contact employees. They focused on universalistic and contingency approaches and developed the configural approach by extending configurational theory principles to role theory. They compared and contrasted different approaches and reviewed literature that has remained less accessible to marketing researchers. John considered price contract design templates governing procurement and marketing of industrial equipment. He argued that price formats choices precede the selection of a price level. These price formats are an integral aspect of the institutional arrangement devised to govern an exchange. John reviewed institutions, that is, rules of interaction that govern the behavior of actors in dealing with other actors, with a focus on their pricing elements. Articles in the Fifth Volume The existence of two discrete, parallel, interactive cognitive systems underlying human judgment and reasoning has been postulated in several psychological and behavioral disciplines (Agarwal and Malhotra 2005; Malhotra 2005). One system is relatively unconscious, based on associations, and tends to be rapid. The other system is consciously guided, based on symbolic manipulation, and tends to be slower. The two systems generally operate in parallel, contributing interdependently to decision outcomes. Bond, Bettman, and Luce reviewed recent developments in consumer behavior in terms of this dual-systems paradigm. They first examined a variety of frameworks that have been proposed, focusing on both their commonalities and their application domains. Then, they applied these frameworks to review selected topics from the recent marketing literature, including persuasion, metacognition, and immersive experiences. The “chasm” is a well-accepted paradigm among new-product marketing practitioners that has
xiv
Review of Marketing Research
taken root in the last decade. According to this paradigm, the market for new products is composed of “early” and “mainstream” markets with a “chasm” in between them. A fundamental premise of such an approach is that there is a communication break, at least to some degree, between the consumers in the early adopters and the mainstream market segment. Libai, Mahajan, and Muller examined empirical support for the existence of a communication break in the diffusion of innovations using aggregate product growth data, typically used in the diffusion of innovation research. They reviewed three alternative models due to Bass, Rogers, and Moore. Their results provide some support for the dual-market phenomenon and show the existence of a partial communication break. As the authors pointed out, aggregate adoption data are not sufficient for answering these questions. More in-depth and disaggregate investigation across various time points should be conducted (Kim and Malhotra 2005). Rajagopalan and Bayus explored two of Eric Raymond’s key open-source product development principles embodied in the bazaar community development model involving developers and users. They empirically examined the relationships between project community size (“eyeballs”) and development activity, and between development activity and product adoption. Their analysis supports the premise that “developer eyeballs” are positively related to development activity and that product development activity is significantly related to the speed of product adoption. Thus, they find support for some key principles of the open-source bazaar. However, some of their results are contrary to the bazaar model. Therefore, Raymond’s bazaar community development model involving developers and users should be revised to accommodate the more typical open-source development project. Future research should explore the applicability of different new-product diffusion models to open-source innovations. The Segmentation-Targeting-Positioning (STP) process is fundamental to the formulation of marketing strategy (Malhotra, Charles, and Uslay 2005). DeSarbo, Blanchard, and Atalay briefly reviewed the STP framework and optimal product positioning literature. Then these authors presented a new constrained clusterwise multidimensional unfolding procedure for performing STP, in which the brand coordinates are a linear function of product characteristics. Their method simultaneously identifies consumer segments, derives a joint space of brand coordinates and segment-level ideal points, and creates a link between specified product attributes and brand locations in the derived joint space. Generalizing the proposed methodology to the analysis of nonmetric and three-way data would extend the range of applications for this approach. Conjoint analysis is one of the most versatile methods in marketing research. Although this method has been popular in practice, one serious constraint has been dealing with the large numbers of attributes that are normally encountered in many conjoint analysis studies. Rao, Kartono, and Su reviewed thirteen methods for handling a large number of attributes that have been applied in various contexts. They discussed the advantages and disadvantages of these methods. Based on their analysis, three methods (self-explicated, partial profiles, and upgrading) seem to stand out and merit consideration by researchers in this area. Yet, no single study has systematically evaluated these potential alternative methods in the context of a specific applied problem. It would be worthwhile to conduct large-scale empirical and simulation studies to compare the methods. Laddering is a qualitative research technique that has great potential to uncover the factors underlying consumer decision making. However, this potential has not been realized because the time and costs of this qualitative technique as well as the lack of standard statistical measures to assess data and solution quality have been obstacles. Reynolds and Phillips assessed the laddering research practices of both professional and academic researchers. They proposed a set of quality metrics and demonstrated the use of these measures to empirically compare the traditional faceto-face interviewing method to an online one-on-one interviewing approach.
Review of Marketing Research
xv
The Internet provides marketers with an expanded set of communications vehicles for reaching customers (Kim and Malhotra 2005; Malhotra, Kim, and Agarwal 2004). Two of the important and fast-growing elements of this new communications mix are online advertising and electronic word-of-mouth. Bucklin, Rutz, and Trusov reviewed recent research developments in marketing that are most relevant to assessing the impact of these communications vehicles. They first discussed the two major forms of Internet advertising: display ads (also known as banners) and paid search. Online communities, social networking sites, online referral programs, product reviews, and blogs all allow word of mouth (WOM) to spread faster and farther than in the past. Research has shown how electronic records of online word of mouth (for example, product reviews) can be connected, via models, to performance outcome variables such as product ratings and sales levels. Articles in This Volume In keeping with the earlier five volumes, this one also reflects an eclectic mixture of theory, measurement, data, and research methods, reinforcing the mission of Review of Marketing Research. The purchase of products is at the heart of much of consumer and marketing research. Baumgartner provides a review of prior classifications of purchase behavior, discussing their strengths and weaknesses. He proposes a new, empirically derived typology based on purchase motives. A classification of forty-four different purchase behaviors reflecting various purchase motives yielded a typology of eight distinct types of purchase behavior based on three underlying dimensions. These dimensions are functional vs. psycho-social purchases, low vs. high purchase involvement, and spontaneous vs. deliberate purchases. The typology was replicated using a task in which respondents rated how well the forty-four different purchase behaviors described the purchase of forty different products, and the interpretation of the three underlying dimensions was also validated. Baumgartner’s typology better captures the important dimensions underlying different forms of buying behavior. His review provides rich insights and contributes to a greater understanding of people’s purchase behavior. Singh and Jain focus on the literature related to the measurement of customer lifetime value (CLV). They highlight the issues related to the context of CLV measurement and propose a contextual framework for understanding and categorizing models of CLV. They also review the major models for measuring CLV in different contexts and discuss their comparative strengths and weaknesses. Finally, they identify the key issues that impact CLV but have not been adequately considered in its modeling. These factors include network effects (such as word-of-mouth effects), cost of customer acquisition, cost of managing customer relationships, cross-selling, competition, forecasting and planning, and endogeneity of CLV drivers. Sriram and Chintagunta discuss learning models in the context of consumer choice. Typically, choice models assume that agents know the utility they would derive from various alternatives with certainty. Such an assumption is likely to be violated, and consumers may experience uncertainty when the agent is new to the context or the choice set has new alternatives. Consumers resolve uncertainty regarding products or their characteristics in such contexts by making use of learning models. Sriram and Chintagunta provide a critical review of the learning literature in marketing and economics, with a focus on models in which consumers update their beliefs in a Bayesian fashion, with the extent of updating being related to their perceived precision of the signals that aid in such learning. They discuss several possible extensions of the learning literature with an emphasis on biased signals, changing value of the unknown entity, and integration of Bayesian and alternative learning mechanisms. They also identify some directions for future research in this area. O’Hern and Rindfleisch discuss customer co-creation in the context of new product development. Traditionally, new product development (NPD) has been viewed largely as an internal firm-
xvi
Review of Marketing Research
based activity, with the customers being relatively passive buyers and users. They challenge this traditional paradigm and discuss a new perspective in which customers are active co-creators of the products they buy and use and, in some cases, are capable of creating new products with little help from firms. They identify the origins of this paradigm shift and present a conceptual typology of four different types of co-creation activity. Customer co-creation involves two key processes: contribution by way of submitting content and selection by choosing which of these submissions will be retained. Using these two processes as a foundation, the authors offer a conceptual typology of four different forms of customer co-creation. Based on this emerging paradigm, they offer an agenda for future research. Their agenda focuses on the impact of customer co-creation on six distinct domains of inquiry: (1) organizational culture, (2) organizational learning, (3) organizational dynamics, (4) resources and capabilities, (5) customer valuation, and (6) brand communities. The Return on Marketing Investment (ROMI) metric holds promise in increasing the accountability for marketing spending. However, many organizations experience several roadblocks to measuring ROMI and using it to make better marketing decisions and achieve higher performance. Pauwels and Reibstein discuss the challenges in measuring return on marketing investment. They define ROMI as the incremental margin generated by a marketing program divided by the cost of that program at a given risk level. They discuss ten such roadblocks, give examples, and critically examine how research has addressed and should further address these issues. The service-dominant (S-D) logic shifts the focus of marketing away from the production and distribution of goods (goods-dominant logic) toward service, the application of operant resources (knowledge and skills), as the basis of exchange. The central tenet of S-D logic is that reciprocal service is the fundamental basis of economic exchange, that is, service is exchanged for service. Vargo, Lusch, Akaka, and He give a review and assessment of the S-D logic. They present an S-D logic perspective of the market and marketing and summarize its current state of development. They clarify major theoretical misconceptions and review the extension of S-D logic and its integration with existing knowledge. They provide an assessment of the role of S-D logic in the evolution of academic marketing, and identify directions for future research in this area. Initially, S-D logic was not developed as a testable theory, and there is a great need to further develop testable hypotheses based on the service-centered mindset. Moreover, these hypotheses should be empirically tested in a variety of settings so that a wealth of findings could accumulate. Dutta, Bergen, and Ray deal with costs of price adjustment in marketing. They review the literature in marketing and economics to summarize what we know about the nature, magnitude, and broad impact of these costs. The literature on the nature and scope of these costs has been evolving from simple menu costs to richer decision-making, organizational, and customer-based costs. These costs have substantial implications for research in pricing; they influence the magnitude and frequency of price changes, asymmetric pricing, pass-through in channels, and price synchronization. The authors also identify some areas of potential interest, where consideration of price adjustment costs are likely to yield greater insights into marketing decisions for both researchers and practitioners. Their basic conclusion is that there are significant domains of pricing decisions that are under-researched from the perspective of price adjustment costs. An explicit consideration of these costs should lead to greater understanding of pricing and also to better pricing decisions. It is hoped that collectively the chapters in this volume will substantially aid our efforts to understand, model, and make predictions about both the firm and the consumer and provide fertile areas for future research. Review of Marketing Research continues its mission of systematically analyzing and presenting accumulated knowledge in the field of marketing as well as influencing future research by identifying areas that merit the attention of researchers.
Review of Marketing Research
xvii
References Agarwal, James, and Naresh K. Malhotra. 2005. “Integration of Attitude and Affect: An Integrated Model of Preference, Intention, and Choice.” Journal of Business Research 58(4) (April): 483–493. Gollwitzer, Peter M. 1996. “The Volitional Benefits of Planning.” in The Psychology of Action, ed. Peter M. Gollwitzer and John A. Bargh. New York: Guilford Press. Kim, Sung, and Naresh K. Malhotra. 2005. “A Longitudinal Model of Continued IS Use: An Integrative View of Four Mechanisms Underlying Post-Adoption Phenomena.” Management Science 51 (5) (May): 741–755. Malhotra, Naresh K. 2001. “Cross-Cultural Marketing Research in the Twenty-First Century.” International Marketing Review 18 (3): 230–234. ———. 2005. “Attitude & Affect: New Frontiers of Research in the Twenty-First Century.” Journal of Business Research 58 (4) (April): 477–482. Malhotra, Naresh K., James Agarwal, and Mark Peterson. 1996. “Cross-Cultural Marketing Research: Methodological Issues and Guidelines.” International Marketing Review 13 (5): 7–43. Malhotra, Naresh K., and Betsy Charles. 2002. “Overcoming the Attribute Prespecification Bias in International Marketing Research by Using Nonattribute Based Correspondence Analysis.” International Marketing Review 19 (1): 65–79. Malhotra, Naresh K., Betsy Charles, and Can Uslay 2005. “Correspondence Analysis: Methodological Perspectives, Issues and Applications.” Review of Marketing Research 1: 285–316. Malhotra, Naresh K., Sung Kim, and James Agarwal. 2004. “Internet Users’ Information Privacy Concerns (IUIPC): The Construct, the Scale, and a Causal Model.” Information Systems Research 15 (4) (December): 336–355. Malhotra, Naresh K., Sung Kim, and Ashutosh Patil. 2006. “Common Method Variance in IS Research: A Comparison of Alternative Approaches and a Reanalysis of Past Research.” Management Science (December): 1865–1883. Malhotra, Naresh K., and Daniel McCort. 2001. “A Cross-Cultural Comparison of Behavioral Intention Models: Theoretical Consideration and an Empirical Investigation.” International Marketing Review 18 (3): 235–269. Malhotra, Naresh K., Francis Ulgado, James Agarwal, G. Shainesh, and Lan Wu. 2005. “Dimensions of Service Quality in Developed and Developing Economies: Multi-Country Cross-Cultural Comparisons.” International Marketing Review 22 (3): 256–278.
CONTENTS, VOLUME 1 Review of Marketing Research Naresh K. Malhotra
ix
1. A Reappraisal of the Role of Emotion in Consumer Behavior: Traditional and Contemporary Approaches Allison R. Johnson and David W. Stewart
3
2. The Eye of the Beholder: Beauty as a Concept in Everyday Discourse and the Collective Photographic Essay Morris B. Holbrook
35
3. Consumer Information Acquisition: A Review and an Extension Lan Xia and Kent B. Monroe
101
4. The Resource-Advantage Theory of Competition: A Review Shelby D. Hunt and Robert M. Morgan
153
5. Toward an Integrated Model of Business Performance Sundar G. Bharadwaj and Rajan Varadarajan
207
6. Consumers’ Evaluative Reference Scales and Social Judgment Theory: A Review and Exploratory Study Stephen L. Vargo and Robert F. Lusch
245
7. Correspondence Analysis: Methodological Perspectives, Issues, and Applications Naresh K. Malhotra, Betsy Rush Charles, and Can Uslay
285
About the Editor and Contributors Index
317 319
xix
CONTENTS, VOLUME 2 Review of Marketing Research: Some Reflections Naresh K. Malhotra
ix
1. Consumer Action: Automaticity, Purposiveness, and Self-Regulation Richard P. Bagozzi
3
2. Looking Through the Crystal Ball: Affective Forecasting and Misforecasting in Consumer Behavior Deborah J. MacInnis, Vanessa M. Patrick, and C. Whan Park
43
3. Consumer Use of the Internet in Search for Automobiles: Literature Review, a Conceptual Framework, and an Empirical Investigation Brian T. Ratchford, Myung-Soo Lee, and Debabrata Talukdar
81
4. Categorization: A Review and an Empirical Investigation of the Evaluation Formation Process Gina L. Miller, Naresh K. Malhotra, and Tracey M. King
109
5. Individual-Level Determinants of Consumers’ Adoption and Usage of Technological Innovations: A Propositional Inventory Shun Yin Lam and A. Parasuraman
151
6. The Metrics Imperative: Making Marketing Matter Donald R. Lehmann
177
7. Multi-level, Hierarchical Linear Models and Marketing: This Is Not Your Adviser’s OLS Model James L. Oakley, Dawn Iacobucci, and Adam Duhachek
203
About the Editor and Contributors Index
229 231
xx
CONTENTS, VOLUME 3 Review of Marketing Research: A Look Ahead Naresh K. Malhotra Contents, Volume 1 Contents, Volume 2
ix xv xvii
1. Managing Customer Relationships Ruth N. Bolton and Crina O. Tarasi
3
2. A Critical Review of Marketing Research on Diffusion of New Products Deepa Chandrasekaran and Gerard J. Tellis
39
3. On the Distinction Between Cultural and Cross-Cultural Psychological Approaches and Its Significance for Consumer Psychology Giana M. Eckhardt and Michael J. Houston
81
4. Consumer Responses to Price and Its Contextual Information Cues: A Synthesis of Past Research, a Conceptual Framework, and Avenues for Further Research Dhruv Grewal and Larry D. Compeau
109
5. Store Brands: From Back to the Future Serdar Sayman and Jagmohan S. Raju
132
6. Language, Thought, and Consumer Research Dwight R. Merunka and Robert A. Peterson
152
7. You Ought to Be in Pictures: Envisioning Marketing Research Russell W. Belk
193
About the Editor and Contributors
207
xxi
CONTENTS, VOLUME 4 Review of Marketing Research: Taking Stock Naresh K. Malhotra
ix
Contents, Volume 1 Contents, Volume 2 Contents, Volume 3
xv xvii xix
1. Formal Choice Models of Informal Choices: What Choice Modeling Research Can (and Can’t) Learn from Behavioral Theory Jordan J. Louviere and Robert J. Meyer
3
2. How Much to Use? An Action-Goal Approach to Understanding Factors Influencing Consumption Quantity Valerie S. Folkes and Shashi Matta
33
3. Integrating Purchase Timing, Choice, and Quantity Decisions Models: A Review of Model Specifications, Estimations, and Applications V. Kumar and Anita Man Luo
63
4. Brand Extension Research: A Cross-Cultural Perspective Michael A. Merz, Dana L. Alden, Wayne D. Hoyer, and Kalpesh Kaushik Desai
92
5. A Review of Eye-Tracking Research in Marketing Michel Wedel and Rik Pieters
123
6. Role Theory Approaches for Effectiveness of Marketing-Oriented Boundary Spanners: Comparative Review, Configural Extension, and Potential Contributions Jagdip Singh and Argun Saatcioglu
148
7. Designing Price Contracts for Procurement and Marketing of Industrial Equipment George John
183
About the Editor and Contributors
201
xxii
CONTENTS, VOLUME 5 Review of Marketing Research: The First Five Volumes Naresh K. Malhotra Contents, Volume 1 Contents, Volume 2 Contents, Volume 3 Contents, Volume 4
ix xvii xviii xix xx
1. Consumer Judgment from a Dual-Systems Perspective: Recent Evidence and Emerging Issues Samuel D. Bond, James R. Bettman, and Mary Frances Luce
3
2. Can You See the Chasm? Innovation Diffusion according to Rogers, Bass, and Moore Barak Libai, Vijay Mahajan, and Eitan Muller
38
3. Exploring the Open Source Product Development Bazaar Balaji Rajagopalan and Barry L. Bayus
58
4. A New Spatial Classification Methodology for Simultaneous Segmentation, Targeting, and Positioning (STP Analysis) for Marketing Research Wayne S. DeSarbo, Simon J. Blanchard, and A. Selin Atalay
75
5. Methods for Handling Massive Numbers of Attributes in Conjoint Analysis Vithala R. Rao, Benjamin Kartono, and Meng Su
104
6. A Review and Comparative Analysis of Laddering Research Methods: Recommendations for Quality Metrics Thomas J. Reynolds and Joan M. Phillips
130
7. Metrics for the New Internet Marketing Communications Mix Randolph E. Bucklin, Oliver J. Rutz, and Michael Trusov
175
About the Editor and Contributors
193
xxiii
Review of
Marketing Research
Chapter 1
A review of prior classifications of purchase behavior and a proposal for a new typology Hans Baumgartner
Abstract The purpose of this chapter is twofold: (1) to critically review prior typologies of purchase behavior, recognizing their strengths and weaknesses; and (2) to propose a new, empirically derived typology based on purchase motives that better captures the important dimensions underlying different forms of buying behavior. Although the purchase of products is only one aspect of consumer behavior, it is at the heart of much of consumer research, and it is hoped that a comprehensive review of previous attempts to classify purchase behaviors and the proposed new typology will contribute to an improved understanding of the whys and wherefores of people’s buying behavior. Prior Typologies of Purchase Behavior Many different types of purchase behavior have been distinguished in the consumer behavior and marketing literatures. In this section we will review the major typologies of consumer buying behavior that have been proposed. We will not discuss work on individual forms of purchase behavior such as impulse buying, or general models of consumer decision making and choice such as the Nicosia (1966), Engel, Kollat, and Blackwell (1968), Howard and Sheth (1969), and Bettman (1979) models, because they are not concerned with classifying buying behavior. The review will be structured into three sections: (1) approaches that distinguish purchase behaviors along a single dimension of variation (unidimensional typologies), (2) approaches that differentiate among purchase behaviors using multiple dimensions of variation (multidimensional typologies), and (3) approaches that present an unstructured collection of distinct categories of purchase behavior (categorical typologies). Unidimensional Typologies of Purchase Behavior Unidimensional approaches distinguish purchase behaviors based on one dimension of variation. The underlying dimension may be unipolar (for example, degree of involvement) or bipolar (such as thinking vs. feeling purchases). Frequently, the dimension is dichotomized and only the low and high poles of the underlying continuum are distinguished in order to simplify the discussion and highlight the differences (for example, low and high involvement). The approaches to be discussed include functional versus psycho-social bases of decision making, Howard and Sheth’s (1969) stages of decision making, and involvement as a determinant of purchase behavior. 3
4
Hans Baumgartner
Functional vs. Psycho-social Bases of Decision Making Based on the traditional distinction in philosophy and psychology between cognition, reason, intellect, mind, and logic on the one hand and affect, emotion, passion, soul, and intuition on the other hand, many researchers have attributed purchase decisions to either thinking or feeling processes. As early as 1924, Copeland proposed a rational versus emotional classification of consumers’ buying motives (Copeland 1924). The distinction is also implicit in the literature on perceived risk, where several authors have contrasted choice situations involving functional (economic) risk and purchases involving psycho-social risk (e.g., Taylor 1974). The most influential recent discussion of the two positions was provided by Hirschman and Holbrook (1982; Holbrook and Hirschman 1982), who argued that, after some early interest in motivation research (Dichter 1960) and product symbolism (Gardner and Levy 1955; Levy 1959), consumer research had narrowly focused on the rational aspects of decision making and that it was necessary to reorient the field toward the more experiential aspects of consumption. Specifically, they suggested that the dominant information processing view of consumer behavior could be contrasted with an emerging experiential or hedonic perspective of consumption. While the former sees the consumer as a “logical thinker who solves problems to make purchase decisions” (Holbrook and Hirschman 1982, p. 132), the latter is concerned with the “multisensory, fantasy and emotive aspects of one’s experience with products” (Hirschman and Holbrook 1982, p. 92). According to Hirschman and Holbrook, in the information processing view the consumer focuses on the verbal elements of products and marketing communications, defines tasks as consumption problems to be solved, is cognitively involved in the purchase, acquires relevant information in order to evaluate products based on how well they perform utilitarian functions, and is more interested in product purchase than consumption. In contrast, the consumer in the experiential perspective attends to the sensory elements of products and messages; is driven by pleasure seeking; gets emotionally absorbed in tasks; engages in diversive exploration; experiences fantasies, feelings, and fun; and is more interested in consuming a product than buying it. Similar to Hirschman and Holbrook, Mittal (1988) distinguished between the information processing mode of consumer choice and the affective choice mode. In the former, the consumer is viewed as a rational information processor who acquires pertinent information, evaluates products on relevant choice criteria, and then chooses on the basis of some decision rule. In the latter, judgments are holistic, implicate the self, and are difficult to explicate. Mittal argues that consumption might be motivated by either utilitarian/functional goals or hedonic, symbolic, or psycho-social goals. He calls the latter expressive goals and regards sensory enjoyment, mood-states attainment, social goals (such as impression management), and self-concept fulfillment as subdomains of expressive concerns. Although he suggests that many product choices involve both functional and expressive considerations, he argues that some products are primarily functional or expressive in nature and thus promote either an information processing or affective choice mode. Mittal, Ratchford, and Prabhakar (1990) also proposed that brand evaluations could be based on functional or expressive attributes (or both). Functional attributes are objective, instrumental, performance-oriented, and utilitarian; expressive attributes, in contrast, are subjective, experiential, image-oriented, and hedonic. Examples of functional attributes are: gets rid of dandruff, cleans hair, gives hair shine (for shampoo); and masks body odor, convenient to apply, does not irritate the skin (for perfume). Examples of expressive attributes are: it is a high-fashion brand, the bottle is attractive, it is for beautiful people (for shampoo); and it will impress people, it is feminine, the fragrance is flowery (for perfume). The authors argue that the relative importance of functional and expressive attributes depends on whether the consumer’s buying motive is functional or expressive.
purchase behavior and a proposal for a new typology
5
When functional motives govern a purchase, consumers are mostly concerned with the physical function of the product and try to satisfy physiological and safety needs. When expressive motives are dominant, consumers are focused on how they feel and how others feel about them, and they try to satisfy esteem, social, and self-actualization needs. In two studies, Mittal, Ratchford, and Prabhakar showed that attributes could be classified into functional and expressive sets and that brand attitudes for shampoo were mostly influenced by functional attribute beliefs, whereas brand attitudes for perfume were primarily driven by expressive attribute beliefs. Johar and Sirgy (1991) draw a similar distinction between utilitarian and value-expressive attributes. They argue that with utilitarian attributes, the consumer evaluates the congruity of performance-related product beliefs with referent (for example, ideal) attributes (so-called functional congruity). On the other hand, with value-expressive attributes, self- and social-congruity— or the match between value-expressive attributes and the self-concept—matter. The self-concept consists of actual and ideal forms of the self- and social-image, and self-congruity thus involves assessments of the degree of self-consistency (with one’s actual self-image), self-esteem (with one’s ideal self-image), social consistency (with one’s actual social-image), and social approval (with one’s ideal social-image). The distinction between cognitive and affective bases of decision making is also reflected in research on attitudes. Batra and Ahtola (1991) distinguish between hedonic and utilitarian sources of consumer attitudes (see also Ahtola 1985). The hedonic component is assumed to be “based on the consumer’s assessment of how much pleasure he gets,” while the utilitarian component is “based on his assessment about the instrumental value of the brand’s functional attributes” (p. 161). These two determinants of an attitude correspond to the two reasons for purchasing and consuming products: “(1) consummatory, affective (hedonic) gratification (from sensory attributes), and (2) instrumental, utilitarian reasons concerned with ‘expectations of consequences’ (of a meansend variety, from functional and nonsensory attributes)” (p. 159). Batra and Ahtola constructed scales to measure the two components (for example, pleasant/unpleasant, nice/awful, and happy/ sad for the hedonic component, and useful/useless, valuable/worthless, and beneficial/harmful for the utilitarian component; see Crowley, Spangenberg, and Hughes 1992 for a replication; and Crites, Fabrigar, and Petty 1994 for a similar attempt) and showed that (a) the hedonic (utilitarian) attitude component was more strongly related to sensory (instrumental) attribute adequacy composites; (b) the overall attitude of consumers buying toothpaste for emotional/feel-good/ pleasure (rational) reasons was primarily determined by the hedonic (utilitarian) component; and (c) for some behaviors (such as test-driving a Mercedes Benz), the hedonic component was the more important influence on overall attitudes, while for others (such as having dental checkups) the utilitarian component had a greater impact (see also Millar and Tesser 1986). More recently, Voss, Spangenberg, and Grohmann (2003) reported the development of a ten-item instrument that assesses the hedonic and utilitarian dimensions of consumers’ product and brand attitudes (called the HED/UT scale). The scale includes such items as fun/not fun and exciting/ dull for the hedonic dimension and effective/ineffective and helpful/unhelpful for the utilitarian dimension, and the authors provide evidence of superior reliability and validity of their instrument vis-à-vis the Batra and Ahtola (1991) scale. Using the functional approach to attitudes as a theoretical guide (Katz 1960; Smith, Bruner, and White 1956; see also Lutz 1979), Shavitt (1989, 1990; Shavitt and Lowrey 1992; Shavitt et al. 1994) proposed that two important functions served by attitudes were the utilitarian and the social identity (or symbolic) functions. Utilitarian attitudes are based on rewards, and punishments associated with an attitude object. They are generally linked to outcomes derived from intrinsic product attributes and include sensory experiences the consumer has had with a product (which
6
Hans Baumgartner
are usually classified as hedonic). However, the most prototypical examples involve a product’s performance on functional characteristics, such as providing protection from the sun for sunglasses or obtaining relief from the heat for air conditioners. Social identity attitudes, on the other hand, reflect a desire to express one’s central values or social identity and to gain social approval. Shavitt argues that some products generally engage the utilitarian function (for example, aspirin), others tend to trigger the social identity function (for example, class ring), and still others may serve either function (for example, watch). She also showed that an advertising appeal that was relevant to a product’s primary function (such as a functional appeal for a functional product) led to higher ad and brand attitudes and purchase intentions than a function-irrelevant appeal, and that the correspondence between attitudes and behavioral predictions was stronger when the function that was salient at attitude expression was consistent with the function engaged by the brand. The distinction between functional and psycho-social purchase motives has also been considered in the context of involvement (see the next section for a more detailed treatment of involvement). Park and Mittal (1985) conceptualize involvement in terms of both its causes (type of involvement) and its strength (level of involvement). They distinguish cognitive involvement from affective involvement depending on the motives underlying consumers’ involvement: utilitarian motives or value-expressive motives. In the former case the consumer is mainly concerned with the functional benefits and costs of a product, whereas in the latter case self-concept congruity and self-image enhancement are the reasons for a consumer’s involvement (see also Park and Young 1986). Claeys, Swinnen, and Vanden Abeele (1995) studied consumers’ cognitive structures associated with low-involvement “think” (for example, batteries, glue, detergent) and “feel” (for example, birthday cards, indoor plants, liquor) products. They define think products as those for which the motive for buying is utilitarian and cognitive, the mode of processing is logical, rational, and sequential thinking, and the focus of concern is functional performance, costs/benefits, and tangible features. In contrast, feel products are those for which the motives for buying are valueexpressive and affective, the mode of processing is holistic, synthetic, and image-based thinking, and the focus of concern is enhancement of the self, subjective meanings, and intangible features. They conducted laddering interviews with one hundred subjects, in which respondents were repeatedly asked why certain attributes of think and feel products were important to them. They then classified responses into the following levels: concrete attributes, abstract attributes, functional consequences, psycho-social consequences, instrumental values, and terminal values. The major finding was that the cognitive structures associated with think products were dominated by functional consequences and concrete attributes (80 and 70 percent of subjects mentioned product meanings at those levels, respectively), whereas the cognitive structures associated with feel products were dominated by psycho-social consequences and abstract attributes (80 and 60 percent of subjects mentioned meanings at those levels, respectively). According to the authors, this supports the notion that buying motives for think products lie at the functional consequence level and that reasons for the purchase of feel products are at the psycho-social level. In the context of research on shopping motivations, Babin, Darden, and Griffin (1994) argue that shopping can produce either utilitarian or hedonic value. Utilitarian value is based on whether the goal that motivated the shopping trip was attained. When shopping has a utilitarian purpose, it is extrinsically motivated by the need for a product or the acquisition of product information, and efficiency of shopping is desirable. Utilitarian shopping is perceived as work, and successful completion of the shopping task is the goal. In contrast, hedonic value derives from the intrinsic enjoyment of shopping as an activity. Hedonic shopping is like play, gives the consumer pleasure, and may sometimes be a means of escape from a stressful situation. The authors constructed a scale (the personal shopping value scale) that assesses the two kinds of shopping value with items such
purchase behavior and a proposal for a new typology
7
as “I accomplished just what I wanted to on this shopping trip” (for functional value) and “This shopping trip was truly a joy” (for hedonic value). They also showed that the hedonic dimension of their shopping value scale was positively correlated with consumers’ experiential shopping motivations, compulsive buying behavior, and unplanned spending, and that the correlation with the degree of pleasure and arousal experienced in a store was stronger than for the utilitarian dimension. Both dimensions were equally strongly related to perceptions of the number of bargains available in the store, amount spent, and overall satisfaction with the shopping trip. The difference between utilitarian and hedonic products has also figured prominently in some of the recent choice literature (e.g., Dhar and Wertenbroch 2000; Okada 2005; Shiv and Fedorikhin 1999). These studies have focused on the conditions that favor consumer choice of utilitarian and hedonic products, but the important point for the present purposes is that functional and psychosocial purchase motivations are a key dimension differentiating buying behaviors and choices. Finally, in recent work by Chitturi, Raghunathan, and Mahajan (2008), (lack of) delivery of utilitarian product benefits is shown to lead to prevention emotions of satisfaction (anger), while (lack of) delivery of hedonic benefits is shown to lead to promotion emotions of delight (dissatisfaction). In turn, the postconsumption emotions are related to word-of-mouth and repurchase intentions. Although dichotomies between functional and psycho-social bases of decision making have been the most common approach, several researchers have considered three classes. Park, Jaworski, and MacInnis (1986, p. 136) distinguish between functional, symbolic, and experiential needs. Functional needs are “those that solve consumption-related problems (such as solve a current problem, prevent a potential problem, resolve conflict, restructure a frustrating situation; see Fennell 1978).” Symbolic needs are “desires for products that fulfill internally generated needs for self-enhancement, role position, group membership, or ego-identification. . . . A brand with a symbolic concept is designed to associate the individual with a desired group, role, or self-image.” Finally, experiential needs are “desires for products that provide sensory pleasure, variety, and/or cognitive stimulation.” The authors argue that researchers have generally classified products as functional, symbolic, or experiential depending on which need they tend to satisfy, but point out that in principle a brand could be positioned with any of these images. Park, Jaworski, and MacInnis’s distinctions are an elaboration of an earlier proposal by Woods (1960) that consumers demand products for ego-involving, hedonic, or functional reasons. According to Woods, ego-involvement is characteristic of prestige products (which are used to project a certain self-image), maturity products (which signal that the consumer has reached a certain state of maturity, as in the case of cigarettes or alcohol), status products (which symbolize class membership), and anxiety products (for which ego-defense, instead of ego-enhancement, is the primary motivator). Hedonic products are those for which demand is based on sensory appeal. Finally, functional products have little cultural or social meaning and are simply bought because of the function they perform. In summary, the differentiation into functional (utilitarian) and psycho-social (hedonic, experiential, value-expressive, and symbolic) bases of decision making is an old one, and it is reflected in various distinctions made in the information processing, attitude, and choice literatures. The basic implication is that purchase (and, more generally, consumption) can be motivated by either functional (cognitive) or psycho-social (emotional) considerations and that this leads to a corresponding interest in utilitarian or hedonic products, or a focus on utilitarian or hedonic attributes of products. Furthermore, researchers have speculated about and, less frequently, studied the characteristics of cognitive and emotional decision processes. Although the distinction highlights important differences between the two choice modes and therefore constitutes a relevant dimension for differentiating between alternative purchase behaviors, by itself it provides a limited account
8
Hans Baumgartner
of the ways in which purchase behaviors may differ, as will become clearer once other dimensions and typologies have been described. Howard and Sheth’s Stages of Decision Making Howard and Sheth (1969) presented a dynamic model of buying behavior according to which purchase decision making depends on a consumer’s familiarity with both the product class and the brands in the product class, and the strength of preference for particular brands. The three stages in the decision process are called extensive problem solving, limited problem solving, and routinized response behavior. In the first stage, extensive problem solving, the consumer knows little about the product class and does not have strong attitudes toward the various brands. Information search is extensive because the consumer has to acquire a product class concept, and the speed with which the decision is made is slow. Many products are actively considered for purchase (that is, the size of the evoked set is large), and consumers carefully deliberate which brand will best suit their needs. In the second stage, limited problem solving, the consumer has developed a basic familiarity with the product class and knows what the important choice criteria are but is still unsure about how the different brands compare and does not yet have clearly defined brand attitudes. Both the amount of information searched and the speed of decision making are moderate. The size of the evoked set is smaller than in the case of extensive problem solving, but there is no strong preference for any one brand. Apart from situations in which consumers gradually become more familiar with a product category and decision making becomes less extensive, limited problem solving is also characteristic of cases in which a new brand is introduced into an established market so that existing buying habits are disrupted. In the third stage, routinized response behavior, the consumer knows both the product class and the brands within the product class well and has a clear preference for one or two brands in the evoked set. Information search is minimal and purchase decisions are made quickly, often in response to environmental cues such as a product display in a store. For example, impulse purchases are mentioned as an instance of routinized response behavior. In subsequent work, Howard (1977, 1989, 1994) extended the dynamic model of buying behavior in two ways. First, he explicitly related the three decision-making stages to the product life cycle. Extensive problem solving is assumed to be characteristic of the introduction stage, limited problem solving of the growth stage, and routinized response behavior (or routine problem solving) of the maturity and, to some extent, the decline stage. Second, he developed the concept of the decision-making model (consisting of the constructs of information, brand recognition, attitude, confidence, intention, and purchase) and considered modifications of the basic model when it is applied to buyer behavior in each of the product life cycle stages. Here we will discuss only those aspects of the conceptualization that are relevant to a discussion of typologies of consumer buying behavior. Howard views extensive and limited problem solving as rational information processing tasks during which consumers first form a product category image (that is, what the product is, what benefits it offers, and so forth) and then attain a brand concept (how the brands rate on the choice criteria, and so forth). The major difference between the two stages is in the amount of information needed and the deliberateness of the decision-making process involved. In routine problem solving, several distinct purchase patterns emerge. Some consumers buy the same brand habitually (or loyally), usually with little thinking and low involvement in the purchase task. Others get bored with buying the same brand again and again and engage in variety seeking (or boredom problem
purchase behavior and a proposal for a new typology
9
solving). This means that consumers complicate the buying process by considering other product benefits besides price and availability. In effect, this means that they return to the stage of limited problem solving. Finally, emotional factors are more important in routine buying tasks because, in the maturity stage of the product life cycle, brands become more and more similar on functional characteristics, and emotions are used to differentiate parity brands. A major contribution of the Howard and Sheth model is that these authors recognized early on the importance of familiarity in influencing a consumer’s decision-making process. Familiarity, conceptualized in terms of both product category and brand familiarity, is, in combination with the concept of strength of brand preference, the key driver determining the amount of information searched and the speed with which purchase decisions are made. The model provides a useful perspective on the temporal dynamics of decision making over the product life cycle and has many strategic implications for the management of brands from the introduction to the decline stage. The major shortcoming of the model is that because of its unidimensional nature, it ignores or minimizes important differences between various forms of buying behavior. This problem is most glaring in the routinized response behavior stage. Habitual purchase behavior, impulse buying, variety seeking, and emotional buying are all regarded as instances of routine problem solving. Furthermore, buying behavior in the early stages of the life cycle is restricted to utilitarian information-processing tasks, in which consumers rationally acquire product information and assess the benefits of different product features. Emotional factors are given insufficient attention early in the product life cycle. Involvement and Purchase Behavior It is now well established that an important determinant of how consumers make purchase decisions is their level of involvement. The concept of involvement was first introduced into consumer behavior to explain the persuasive effects of advertising (Krugman 1965), and many of the subsequent discussions of the construct dealt with the role played by involvement in the communication process (e.g., Andrews, Durvasula, and Akhter 1990; Celsi and Olson 1988; Greenwald and Leavitt 1984; Park and Young 1986; Petty, Cacioppo, and Schumann 1983; Ray et al. 1973). Here we will focus on the literature that has discussed the influence of involvement on purchase behavior. As mentioned by Kassarjian (1981, 1994), the increased interest in involvement as a determinant of consumer behavior in the early 1980s was a counter-reaction to the then dominant view of the consumer as an active information processor who carefully deliberates each purchase before making a choice. In his presidential address at the Association for Consumer Research Annual Conference, Kassarjian (1978) suggested that while “[i]mportant or expensive, high risk, high involvement, psychologically or socially or ego related products, may under certain conditions, lead to highly sophisticated actions and decisions . . . the world is mostly full of insignificant decisions and unimportant solutions” (p. xiii). Olshavsky and Granbois (1979) similarly argued that “for many purchases a decision process never occurs, not even on the first purchase” (p. 98). Kassarjian (1981) specifically proposed that there might be two fields of decision making, one applicable to high-involvement decisions and the other to low-involvement decisions, and he admonished consumer researchers to be cognizant of this distinction. We will refer to a consumer’s involvement in purchase decisions as purchase involvement and, consistent with current conceptualizations of involvement (Celsi and Olson 1988; Greenwald and Leavitt 1984; Park and Young 1986; Zaichkowsky 1985), define it as the personal relevance of a purchase task. Personal relevance means that the consumer views the purchase as important, is willing to expend cognitive effort on the purchase decision, and/or gets emotionally absorbed in the purchase process. Other terms used to refer to purchase involvement are importance of the
10
Hans Baumgartner
purchase (Howard and Sheth 1969), decision importance (Clarke and Belk 1979), and purchasedecision involvement (Mittal 1989). Purchase involvement is a function of the consumer, the product, and situational factors. As suggested by Kassarjian (1981), some consumers find purchasing activities highly involving, regardless of the product or situation in question. Building on this idea, Slama and Tashchian (1985) developed a scale of purchasing involvement that assesses individual differences in the extent to which a consumer is interested in shopping, values making careful purchase decisions, and enjoys getting a bargain. They also showed that consumers high in purchasing involvement tended to be women with children, moderate incomes, and relatively high levels of education. Purchase involvement also depends on the characteristics of the product to be bought (Clarke and Belk 1979; see also Hupfer and Gardner 1971; Lastovicka and Gardner 1979; Laurent and Kapferer 1985; Ratchford 1987; Zaichkowsky 1985). Across consumers and situations, some products tend to generate more involvement than others. Laurent and Kapferer argue that product involvement will be high when the product is important to the consumer (interest), when it has emotional appeal (pleasure), when it says something about the product user (sign value), or when the purchase of the product is risky, either because the likelihood of negative consequences is high (risk probability) or the negative consequences are severe (risk importance). Finally, purchase involvement is influenced by situational factors. To begin with, the nature of the buying task impacts consumers’ involvement in the purchase. Clarke and Belk (1979) argue that a task will be involving if it relates to important goals, and they show that a product bought as a gift leads to greater expenditure of effort than the same product bought for personal use. Furthermore, the circumstances under which the purchase occurs may affect the consumers’ involvement. For example, time pressure is likely to decrease involvement, whereas the presence of relevant others in either the purchase or usage situation should increase involvement (Houston and Rothschild 1978). It is thus important to know not only who makes the purchase and what is being bought, but also the task and context factors surrounding the purchase. Previous research on the consequences of involvement on purchase decisions has generally focused on the effects of cognitive forms of involvement, where a consumer is primarily concerned with the functional performance of products, rather than affective forms of involvement, where consumers focus on the hedonic and symbolic qualities of the product (see Park and Mittal 1985; Park and Young 1986). The basic hypothesis is that involvement leads to more extensive decision processes in terms of the amount of internal and external information consulted and the number of choice alternatives evaluated. The empirical evidence is consistent with this prediction. For example, Beatty and Smith (1987) showed that purchase involvement was associated with greater retailer, media, and interpersonal search, and similar results for search effort were also obtained by Clarke and Belk (1979) and Zaichkowsky (1985), among others. Zaichkowsky found that consumers who were involved with products were more likely to compare attributes across brands, perceive differences among brands, and have a most preferred brand in the product category. Gensch and Javalgi (1987) showed that low-involvement consumers tended to use an attribute-based, noncompensatory choice heuristic, while high-involvement consumers were more likely to use a brand-based, compensatory choice heuristic. In summary, there is ample support for Kassarjian’s (1981) claim that low- and high-involvement decision making differ, and the distinction between low- and high-involvement purchase behaviors is well established. However, similar to the differentiation between functional and psycho-social bases of decision making, the use of involvement as the sole dimension for classifying buying behavior is simplistic and ignores other relevant considerations. We next turn to a discussion of multidimensional frameworks, which consider several dimensions simultaneously.
purchase behavior and a proposal for a new typology
11
Multidimensional Typologies of Purchase Behavior Multidimensional typologies cross-classify purchase behaviors based on several underlying dimensions. Usually, these typologies combine two of the dimensions discussed in the previous sections, such as functional vs. psycho-social bases of decision making and degree of involvement. Four specific multidimensional models will be discussed in this section: Assael’s model, the FCB grid and its modification by Rossiter and his associates, and Bagozzi’s classification of choice processes. Assael’s Model of Consumer Decision Making Assael classifies consumer purchase processes on the basis of two dimensions (e.g., Assael 1995): extent of decision making (ranging from decision making to habit) and degree of involvement in the purchase (low to high). Decision making is characterized by information search and the consideration of multiple brand alternatives, while habit entails little or no information search and the repeated purchase of the same product based on previous experience. Involvement is a function of the importance of the purchase to the consumer and depends on the financial, social, and psychological risks associated with it. Combining the two dimensions yields four types of consumer purchase processes: complex decision making (high-involvement decision making), limited decision making (low-involvement decision making), brand loyalty (high-involvement habit), and inertia (low-involvement habit). Complex decision making occurs when a purchase is important to the consumer but little previous experience is available to guide the choice process. The consumer has to gather information and carefully evaluates alternatives. Assael includes symbolic purchasing behavior in the category of complex decision making and argues that it differs from high-involvement utilitarian purchase decisions only in the nature of information processing and brand evaluation. Instead of judging brands in terms of utilitarian performance, quality, and value, consumers view brands as symbolic entities and evaluate their consistency with desired self-images. When important purchases are made repetitively, decision making becomes a habit, but as long as involvement stays high and the consumer is satisfied with the product, there will be strong commitment to a particular brand. Assael calls this brand loyalty. Even if consumers are not involved in a purchase task, they may have to engage in some information search and alternative evaluation if they lack previous experience with the product. This is referred to as limited decision making. Assael regards variety seeking and impulse buying as instances of limited decision making because the consumer is not committed to a single brand and involvement is low. He also argues that such decisions are not preplanned and are usually made in the store. Low-involvement purchase decisions that are made habitually are termed inertia by Assael. Although consumers tend to buy the same brand, this does not signal commitment to the brand purchased, but simply reflects the convenience of not having to consider different alternatives. Assael’s model captures several important characteristics of consumer purchase processes. First, it clearly differentiates between low- and high-involvement purchases, which entail very different information-processing demands. Second, it considers amount of prior experience as an important influence on the kind of decision making that takes place and accords habit an important role in the purchase process, which is congruent with its practical importance for many product categories. Third, it distinguishes between two forms of repeat purchase behavior—brand loyalty and inertia—that differ in their psychological mechanisms and have very different implications for marketing practice.
12
Hans Baumgartner
The model also has several weaknesses. First, the two dimensions of high/low involvement and decision making/habit are probably not independent. For example, Assael himself argues that when consumers are highly involved, they will find it worthwhile to consider product alternatives carefully. However, information search and consideration of brand alternatives are seen as characteristics of decision making. Thus, extent of decision making should be positively correlated with degree of involvement. Second, with brand loyalty the consumer might be committed to the product, but the purchase decision per se is probably not highly involving because it is made habitually. Third, the phenomenology of high-involvement utilitarian and symbolic purchase decisions, or of mundane, limited decision tasks, variety seeking behavior, and impulse purchases is quite different, yet they are classified in the same quadrant in Assael’s model. The simple 2 by 2 classification thus probably does not do justice to the richness of actual purchase decisions. The FCB Grid The FCB (Foote, Cone, & Belding) grid was developed as an advertising planning model and classifies products and purchase decisions in terms of two dimensions: (1) thinking versus feeling, and (2) low versus high involvement (Berger 1986; Ratchford 1987; Ratchford and Vaughn 1989; Vaughn 1980, 1986). The thinking/feeling dimension is based on whether the purchase of a product is triggered by cognitive or affective motives and whether the mode of information processing is primarily cognitive or affective in nature. The major cognitive motive is a utilitarian need for a product that performs well on functional attributes, and it leads to left-brained, analytical, and semantic processing of relatively objective product features. Affective motives include consumers’ needs for ego gratification, social acceptance, and sensory pleasure, and purchase decisions in this case are characterized by right-brained, holistic, and sensory processing of more subjective product information. Involvement is conceptualized as the importance of a purchase to the consumer and the resulting attention and concern that important purchases engender. When the two dimensions are dichotomized and cross-classified, four quadrants emerge: highinvolvement thinking, high-involvement feeling, low-involvement thinking, and low-involvement feeling. Vaughn (1980, 1986) relates the FCB grid to the hierarchy-of-effects model and argues that the high-involvement thinking quadrant is characterized by a learn-feel-do sequence, the high-involvement feeling quadrant by a feel-learn-do sequence, the low-involvement thinking quadrant by a do-learn-feel sequence, and the low-involvement feeling quadrant by a do-feel-learn sequence. The advertising implications derived from the grid are as follows: high-involvement thinking products require informational advertising; high-involvement feeling products, emotional advertising; low-involvement thinking products, advertising that creates and reinforces habits; and low-involvement feeling products, advertising that emphasizes personal satisfaction. In order for the grid to be used in advertising planning, products have to be assigned to the appropriate quadrant of the grid. Five studies were conducted to construct scales for measuring think/ feel and involvement (see Ratchford 1987 for a detailed description). Although thinking/feeling is conceptualized as a bipolar dimension in the grid, the two poles are assessed with unipolar scales measuring the extent to which the purchase decision is logical, objective, and based on functional facts (for thinking) and the degree to which it expresses the consumer’s personality, is based on feelings, and involves sensory considerations (for feeling). Placement on the think/feel dimension depends on the relative amount of thinking or feeling associated with the purchase. Involvement is assessed with items measuring how important the decision is, how much thought the decision requires, and how much consumers stand to lose if they make the wrong decision. Grid studies conducted all over the world show that products tend to be classified in the same quadrant across
purchase behavior and a proposal for a new typology
13
countries, although some interesting situational variations exist. For example, Berger (1986) reports that life insurance is only a moderately involving decision in Sweden, possibly because of the extensive social support available in that country. The FCB grid has a great deal of heuristic value for analyzing purchase decisions and developing advertising that matches consumers’ natural way of buying products in different product categories. The grid was one of the earliest attempts to come up with a simple, multidimensional classification of purchase decisions, and it incorporated such emerging issues as experiential/ hedonic buying behavior and low-involvement decision making. The grid stimulated empirical (although mostly nonacademic) research and yielded a validated set of measurement scales for placing products along the think/feel and involvement axes. However, the FCB grid has also been criticized on several fronts. One problem is that even though thinking and feeling are measured on unipolar scales, they are eventually combined into a single measure by subtracting the average think score from the average feel score and placing products on a bipolar dimension ranging from thinking to feeling. This means that a product in the middle of the grid might be either relatively high or relatively low on both thinking and feeling. Another problem is that the two axes of the grid may not be independent. For example, Ratchford (1987) reports that in one study the involvement and thinking items tended to load on the same factor, which is not surprising since one of the involvement items (decision requires a lot/little thought) looks very much like a thinking item. In an attempt to correct these problems, Kim and Lord (1991) and Putrevu and Lord (1994) proposed a modification of the FCB grid in which high/ low cognitive involvement and high/low affective involvement with products are used as dimensions. This avoids treating feel and think as opposite ends of a bipolar scale and allows cases in which someone is in both a thinking and a feeling mode or does not experience either. However, while the modified model incorporates the high-involvement thinking and feeling quadrants of the original FCB grid and allows mixtures of think/feel for both low- and high-involvement products (that is, low cognitive and affective involvement and high cognitive and affective involvement), it does not clearly accommodate low-involvement thinking purchases (such as habitual buying behavior) or low-involvement feeling purchases (such as certain impulse purchases). Rossiter, Percy, and Donovan (1991) also criticized the FCB grid. First, they argue that the FCB conceptualization of involvement confuses product and brand involvement, does not take into account the familiarity of the target audience with the brand to be chosen, and does not specify the demarcation between low and high involvement. Second, the FCB conceptualization of think/ feel considers only one cognitive purchase motive, and the affective motives are not clearly captured by the measurement scale that was developed to measure the feel end of the dimension (for example, the social approval motive was dropped because it could not be successfully assessed with quantitative self-rating scales). Rossiter and his coauthors developed their own advertising planning grid to correct these weaknesses; this model is discussed next. The Rossiter-Percy-Donovan Grid The Rossiter, Percy, and Donovan model (Rossiter, Percy, and Donovan 1991; Rossiter and Percy 1997) consists of a brand-awareness component and a brand-attitude component. Rossiter and his colleagues argue that brand-awareness is an important communication objective, because before a consumer can form an attitude, he or she must be aware of the brand. However, for our purposes we are more interested in the brand-attitude component. The authors consider two dimensions, type of decision (low or high involvement) and type of motivation (informational or transformational), to arrive at a 2 by 2 grid that is superficially similar to the original FCB grid.
14
Hans Baumgartner
Involvement is defined as the risk perceived by a typical target audience member in choosing a particular brand on a particular purchase occasion. Low involvement means that the consumer is willing to try out the brand without information search, whereas high involvement means that search and conviction are necessary before the brand is purchased. With respect to type of motivation, Rossiter, Percy, and Donovan distinguish between informational and transformational motives. Informational motives are negatively reinforcing, in the sense that an aversive state is relieved, and they include problem removal, problem avoidance, incomplete satisfaction, mixed approach avoidance, and normal depletion. Transformational motives are positively reinforcing because they promise a rewarding state, such as sensory gratification, intellectual stimulation, or social approval (see Fennell 1978). Rossiter and his coauthors develop specific advertising tactics to be used in the four quadrants of the grid and argue that their model yields much richer insights into the advertising planning process than the FCB grid. The major advantage of the Rossiter-Percy-Donovan grid seems to be that the potential motivations underlying purchases are considered more explicitly than in the FCB grid (or, more generally, in discussions of the functional and psycho-social bases of decision making). However, the model was developed as an advertising planning tool; as a framework for classifying different purchase behaviors, it does not go much beyond the original FCB grid, and it cannot account for the differences among purchase behaviors such as variety seeking, impulse buying, or simple hedonic purchases. Bagozzi’s Classification of Consumer Choice Processes Bagozzi (1983) proposed that degree of involvement (low, medium, high) and extent of prior learning (no, yes) could be used to classify consumer purchase behavior. Involvement is defined as the saliency of a product for the consumer, or the importance of the product in a cognitive or affective sense, whereas learning refers to previous experience with the product. When involvement is low and no prior learning has occurred, consumer behavior is based on impulse. Bagozzi includes exploratory or variety-seeking behaviors, instinctual behaviors, and actions prompted by time or social pressures in this category. When involvement is low but the consumer has had previous experience with a product, behavior is driven by habit. Bagozzi argues that the consumer is “programmed to act in a certain way” (1983, p. 151) in this case. When involvement is medium to high and no or little prior learning has occurred, purchase behaviors follow what he calls the cognitive response model. In this case, cognitive information processing precedes affective responses, preference and intention formation, and overt choice behavior. Information processing is effortful because the consumer has little prior experience with purchasing the product, and the purchase is important so the consumer is willing to expend the necessary effort. When involvement is high and the consumer has had previous experience with a product, Bagozzi suggests that the affective response model best describes choices. Here, affective processes are primary and precede the remaining steps in the purchase sequence. He specifically mentions several cases in which affect plays a major role in purchases. These include situations of high-pressure communication or other stimuli that induce intense emotions and self-attributions of likes or dislikes based on past behavior (the so-called dissonance-attribution hierarchy). Finally, when involvement is medium and prior learning has led to stored thoughts and feelings, the parallel response model is said to represent consumer choice. Cognitive and affective processes work in tandem and can have independent effects on preferences and purchase decisions. Bagozzi also warns that consumers may not adhere to the choice process specified for a given cell if environmental stimuli (particularly stimuli in the social environment) are strong enough to dominate choice. For example, he also discusses a
purchase behavior and a proposal for a new typology
15
social response model that introduces social stimuli as an additional influence on cognitive and affective processes and preference. Bagozzi’s model is a novel attempt to classify choice processes, but it has several weaknesses. First, it is highly speculative, and no empirical research has investigated its usefulness. Second, although Bagozzi considers both cognitive and affective choice processes, this important dimension is not used explicitly in developing the classification. Third, the notion of prior learning is not conceptualized very clearly; the demarcation between low, medium, and high involvement is likely to create difficulties in empirical applications; and the relationship between involvement and prior learning is somewhat ambiguous. Categorical Typologies of Purchase Behavior Categorical purchase typologies provide a discrete classification of purchase behaviors without reference to underlying dimensions. This gives them greater flexibility, but it also makes it more difficult to see similarities and differences between the different forms of buying behavior and to assess the exhaustiveness of the typology. Four specific categorical typologies will be discussed in this section: Sheth and Raju’s choice mechanisms, Belk’s extension of Sheth and Raju’s model, the consumption-values approach of Sheth and his associates, and Sproles and Kendall’s research on decision-making styles. Sheth and Raju’s Choice Mechanisms Sheth and Raju (1974) proposed that choice behavior was a linear function of four different choice mechanisms: a situation-controlled choice mechanism (SCCM), a belief-controlled choice mechanism (BCCM), a habit-controlled choice mechanism (HCCM), and a curiosity-controlled choice mechanism (CCCM). In a given situation, an individual’s choice is dominated by one of the mechanisms, but in order to account for individual differences in consumers’ choices, a linear additive formulation is proposed. In situational choice, choice is under environmental control, and personal stimuli (hunger, thirst, and so forth), social stimuli (other people), significative stimuli (physical aspects of the focal object), and symbolic stimuli (symbolic aspects of the focal object) are the major influences on the consumer’s buy/no buy decision. One common example of situational choice is unplanned buying (including impulsive purchase behavior). Belief-controlled choice corresponds to rational decision making under high-involvement conditions, where the consumer carefully compares alternatives based on how well they satisfy his or her needs and wants. Habit-controlled choice involves a binary decision between buying the usual alternative or some other alternative, and it is based on a strong affective or conative tendency rooted in past experiences. It is especially common for frequently purchased products. Finally, curiosity-controlled choice is a manifestation of exploratory behavior in the consumer context in cases where the consumer is mainly motivated by the desire for stimulating consumption experiences. Sheth and Raju argue that there are cyclical sequential linkages among the four choice mechanisms, such that consumers move either from habit-controlled choice to situation-controlled choice via curiositycontrolled choice when decision making has become too simplistic, or from situation-controlled choice to habit-controlled choice via belief-controlled choice when decision making has become too complicated. Furthermore, they attempt to relate the four choice mechanisms to different choice rules such as the compensatory, conjunctive, disjunctive, and lexicographic heuristics. However, the discussion is highly speculative, and nobody seems to have investigated the authors’ proposals empirically.
16
Hans Baumgartner
Belk’s Choice Mechanisms Belk (1985) extended Sheth and Raju’s (1974) theorizing and, in addition to the four choice mechanisms they discussed, considered an affect-controlled choice mechanism. In Belk’s scheme, situation-controlled choice refers to direct influences of the acquisition situation or anticipated consumption situation on acquisition behavior. It does not include influences of the anticipated acquisition or consumption situation on choices that are mediated by behavioral intentions. Three situational effects are considered: contextual effects, precipitating effects, and cuing effects. Contextual effects are due to situational factors that enable or constrain behavior by influencing consumers’ perceptions of the appropriateness or feasibility of an action. Precipitating effects refer to environmental stimuli that trigger action (such as emergencies, strong atmospheric influences). Cuing effects are more subtle environmental influences that orient consumers toward some action (for example, point-of-purchase displays, shelf location). Situational influences need not be recognized by the consumer as affecting behavior. Belief-controlled choice refers most directly to acquisition behavior that is controlled by behavioral intentions. This includes habitual and situational influences, as long as their impact on behavior is mediated by behavioral intentions. It is characteristic of highly involving decisions. Another form of belief-controlled choice occurs in low-involvement situations when choice is due to weak object beliefs (often created by repetitive advertising) in the absence of well-formed intentions. Habit-controlled choice refers to acquisition behavior that is directly controlled by previous choice behavior. Belk suggests that habits will influence intentions most strongly for frequently purchased, low-involvement products, but he argues that habit-controlled choice can occur in both low- and high-involvement situations. Belk treats curiosity-controlled choice as the opposite of habit-controlled choice, and he refers to cases in which the positive effect of habits on acquisition behavior is disrupted or reversed. He suggests that it is most common for moderately involving choices. Finally, affect-controlled choice refers to a direct influence of object affect on acquisition behavior unmediated by behavioral intentions. Belk suggests that this choice mechanism is “especially likely for high-involvement objects with affective overtones and limited time for choice” (p. 26) (such as at an auction for art objects). Sheth and Associates’ Consumption Values Sheth, Newman, and Gross (1991) proposed a theory of market choice behavior according to which market choice is a function of five different kinds of values: functional value, social value, emotional value, epistemic value, and conditional value. A given choice may be dominated by one of these values, but usually is influenced by multiple values. Functional value denotes a product’s “ability to perform its functional, utilitarian, or physical purposes. Alternatives acquire functional value through the possession of salient functional, utilitarian, or physical attributes” (p. 18). Attributes commonly used to judge functional value include intrinsic quality cues such as performance, reliability, and durability. When products lack differentiation on functional attributes, price may be used to assess functional value. Social value pertains to the perceived utility of a product because of its congruence with the norms of relevant social groups or its ability to project a desired social image. Products whose consumption is socially visible are particularly likely to possess social value. Sheth, Newman, and Gross summarize research in the areas of social class, consumption symbolism, reference groups, conspicuous and compensatory consumption, the normative component of attitudes, and opinion leadership and diffusion of innovation that is consistent with the role of social value in product choices. Emotional value refers to a product’s ability to generate affective states. Impulse purchases, entertainment choices, decisions about products affecting one’s
purchase behavior and a proposal for a new typology
17
self-image, and the selection of food products are often based on emotional value. Under emotional value, the authors also subsume motivation research, personality research, work on nonverbal processing and hemispheral brain lateralization, and studies of subliminal perception. Epistemic value deals with a product’s “ability to arouse curiosity, provide novelty, and/or satisfy a desire for knowledge. Alternatives acquire epistemic value through their capacity to provide something new or different” (p. 21). Risky product choices, innovativeness, variety seeking, and exploratory information seeking are commonly attributed, at least in part, to a desire for epistemic value. Finally, conditional value refers to temporary functional or social value due to particular situational circumstances. Frequently, it involves deviations from planned purchase behavior because of contextual contingencies. Examples include purchases made in an emergency situation, seasonal demand for certain products (such as Christmas cards), and products bought to take advantage of a sale. Consumers’ Decision-Making Styles Sproles and Kendall (1986) suggest that consumers can be characterized in terms of how they approach purchase tasks and that these decision-making styles are like personality traits in the consumer domain (see also Sproles and Sproles 1990). Based on a review of the literature, they identified eight fundamental consumer decision-making characteristics and constructed an instrument (called the Consumer Styles Inventory) to measure these characteristics. Brief descriptions of the eight styles follow. Perfectionistic, high-quality conscious consumers make careful purchases and try to buy the best quality products. Brand conscious, “price equals quality” consumers have a preference for well-known national brands, use price as a signal of quality, and like shopping in nice department and specialty stores. Novelty-fashion–conscious consumers are interested in style, variety, and exciting new products. Recreational, hedonistic consumers enjoy going shopping. Price-conscious, “value for money” consumers buy as much as possible on sale and are concerned with getting the best value for their money. Impulsive, careless consumers often buy on impulse and sometimes make careless purchases. Confused by overchoice consumers find it hard to choose between brands and stores and often experience information overload. Habitual, brand-loyal consumers have their favorite brands and stores and stick with them. Summary of Categorical Typologies It seems fair to say that the categorical purchase typologies have not had a strong influence on thinking about buying behavior. They are not well known in the marketing field, and empirical research has been lacking. The various typologies focus on the factors that can control choice in a given situation (product-related beliefs, habits, affect, and so forth), or the types of value that consumers derive from products, and they incorporate some elements of the unidimensional approaches discussed earlier, but supplement them with particular buying motives that are deemed relevant (such as curiosity). However, because they are not derived from underlying dimensions, it is not clear how exhaustive the typologies are and how the purchase behaviors distinguished in the typologies are related to each other. Finally, it remains to be seen whether decision-making styles can be treated like personality traits, which are supposed to be stable across situations and time. A New Typology of Purchase Decisions None of the previously suggested dimensional typologies of purchase decisions is comprehensive enough to unambiguously accommodate the wide variety of purchase behaviors that are encoun-
18
Hans Baumgartner
tered in practice. Furthermore, there are inconsistencies in how certain purchase behaviors are classified, which suggests that one or two dimensions of variation may not be sufficient to accommodate all purchase decisions. Although several of the categorical approaches fare better in terms of comprehensiveness, they have the disadvantage that it is not obvious how the different categories are related. Finally, a common shortcoming of all typologies is that they seem to be based on conceptual discussions of how different purchase decisions are related to each other and that few empirical tests of the proposed categorizations are available. In the remainder of this chapter we will report three studies that were designed to develop and validate a new typology of buying behaviors based on the wide variety of motives that may stimulate purchases. Our assumption was that even though the reasons for making a purchase are numerous, it should be possible to come up with a new typology of purchase decisions based on a limited number of underlying dimensions that parsimoniously describes the buying behaviors commonly encountered in practice. As shown below, this assumption proved correct, and we will propose a new typology called the purchase cube, which distinguishes eight purchase behaviors based on three underlying dimensions. Study 1: Development of the New Typology Method Participants. A total of 306 undergraduate marketing students participated in the study to fulfill a course requirement. The data were collected in twenty-three sessions, with 3 to 24 people participating at any one time. Approximately 51 percent of the participants were female, and 298 respondents provided complete data. Materials and Procedures. Participants were given a short questionnaire that contained the instructions and the stimuli to be classified. The instructions stated: On the next page you will find a table listing reasons that consumers mentioned for why they bought certain things. These reasons reflect different motivations for making a purchase, and we will refer to them as purchase behaviors for short. We would like you to read through this list and then classify the purchase behaviors into groups according to how you think they belong together. You can form as many or as few groups as you think appropriate. There are no right or wrong answers. Classify all purchase behaviors into groups, and put each behavior into only one group. An example dealing with the categorization of four different vegetables was provided to clarify the task. Participants were reminded to read through the entire list before starting the classification, and to take their time and classify all purchase behaviors carefully. On the second page there was a list of forty-four different purchase behaviors. The list was assembled based on an extensive review of the academic literature dealing with different types of purchase decisions and various consumer behavior textbooks. Table 1.1 on page 20 lists the purchase behaviors alphabetically by their acronyms, which will be used later. In the questionnaire the behaviors were arranged in random order. Participants were told to classify the purchase behaviors by putting the same number in front of each behavior that they thought belonged in the same group. Analysis. The similarity between each pair of purchase behaviors was computed as the proportion of respondents who placed the two behaviors into the same category. The resulting 44 by 44 matrix
purchase behavior and a proposal for a new typology
19
of similarities was then submitted to a nonmetric multidimensional scaling (MDS) analysis and average linkage clustering. The data were also analyzed with metric MDS and a cluster analysis based on Ward’s minimum variance method, which yielded similar results. MDS solutions in 1 to 4 dimensions produced the following badness-of-fit measures (Kruskal’s stress formula 1): 0.39, 0.18, 0.08, and 0.05. The three-dimensional solution fit the data well and led to an interpretable solution, so it was chosen as the most appropriate representation of the data. Results Table 1.2 on page 21 reports the coordinates of the forty-four purchase behaviors on the three dimensions emerging from the MDS analysis. For ease of interpretation, the purchase behaviors are ranked separately from low to high on each dimension. Interpreting the three dimensions is relatively straightforward. Dimension 1 separates functional purchase behaviors in the negative domain (for example., making a replacement purchase, buying something because the purchase solves a problem, making a purchase based on functional or utilitarian reasons, making a purchase based on objective, logical criteria) from psycho-social purchase behaviors in the positive domain (for example, buying something because it’s currently fashionable, making a purchase for emotional reasons, buying something because it meets with social approval, buying something because of the status it confers upon you). Dimension 2 may be interpreted as a low- versus high-involvement continuum. At the low-involvement end are buying something because of habit, making a routine purchase, buying something that you usually buy, and buying something because you’ve bought it in the past; and at the high-involvement end are making a purchase based on a careful comparison of the available products, buying something because the price is good, making a purchase based on the quality of the available products, and buying something because it provides the best value. Dimension 3 distinguishes purchase behaviors that occur spontaneously (such as buying something out of curiosity, making an unplanned purchase, making purchases more or less randomly, buying something to try it out) from purchase behaviors that are undertaken deliberately (for example, buying a particular product because of the brand name, buying a brand to project a certain image, buying a particular product because of the reputation of the brand, buying a brand because of its style). The first two dimensions are similar to those used in the FCB or Rossiter, Percy, and Donovan grids. The third dimension is new, although it bears some semblance to Howard and Sheth’s (1969) dimension of familiarity with the product class or the brands within the product class, and to Bagozzi’s (1983) dimension of extent of prior learning (which he sees as distinct from degree of involvement). The third dimension can be seen as introducing an additional dimension of variation into each quadrant of the FCB grid. To simplify the presentation, this dimension has been dichotomized in Figure 1.1 (on page 23), and the left-hand grid shows the spontaneous purchase behaviors and the right-hand grid the deliberate purchase behaviors. Although this obscures differences in the extent to which purchase behaviors are spontaneous or deliberate, it simplifies the graphical presentation of the findings and enables an explicit comparison with the well-known FCB grid. In Figure 1.1, ellipses have been drawn around purchase behaviors that are grouped together in the cluster analysis. Figure 1.1 suggests a distinction among eight different purchase behaviors, corresponding to the eight cells of a 2 × 2 × 2 design obtained by dichotomizing the three dimensions of the MDS analysis into functional versus psycho-social purchases, low versus high purchase involvement, and spontaneous versus deliberate purchases. We will refer to this scheme as the purchase cube (see Figure 1.2, page 24 and Baumgartner 2002).
20
Hans Baumgartner
Table 1.1
Alphabetical Listing of Forty-four Purchase Behaviors by Acronym Brand name Change Comparison Convenience Curiosity Emotion Familiarity Fashion Feel good Fun Habit Image Impulse Liking Logic Loyalty Mindless Past purchase Performance Personality Preference Price Problem solving Quality Random Replacement Reputation Routine Sale Satisfaction Self-esteem Senses Social approval Status Style Thoughtless Time pressure Trial Unplanned Usual Utilitarian Value Variety Want
Buying a particular product because of the brand name Buying something for a change of pace Making a purchase based on a careful comparison of the available products Buying something out of convenience Buying something out of curiosity Making a purchase for emotional reasons Buying a brand because you’re familiar with it Buying something because it’s currently fashionable Buying something because it makes you feel good Buying something just for the fun of it Buying something because of habit Buying a brand to project a certain image Buying something on impulse Buying something because you just like it Making a purchase based on objective, logical criteria Buying a brand because you’re loyal to it Buying something mindlessly Buying something because you’ve bought it in the past Buying something because of its superior performance Buying something because the product expresses your personality Buying a brand because of a strong preference for it Buying something because the price is good Buying something because the purchase solves a problem Making a purchase based on the quality of the available products Making a purchase more or less randomly Making a replacement purchase Buying a particular product because of the reputation of the brand Making a routine purchase Buying something because it’s on sale Buying a brand because you’ve been satisfied with it in the past Buying something because it makes you feel good about yourself Buying something because of the way it looks, sounds, feels, tastes, or smells Buying something because it meets with social approval Buying something because of the status it confers upon you Buying a brand because of its style Buying something without thinking much about it Making a purchase under time pressure Buying something to try it out Making an unplanned purchase Buying something that you usually buy Making a purchase based on functional or utilitarian reasons Buying something because it provides the best value Buying something because of a desire for variety Buying something because it is what you want
purchase behavior and a proposal for a new typology
21
Table 1.2
Placement of the 44 Purchase Behaviors Along the 3 MDS Dimensions (Study 1) Dimension 1 Replacement Problem solving Utilitarian Logic Comparison Quality Value Routine Past purchase Usual Habit Price Convenience Performance Satisfaction Familiarity Time pressure Sale Loyalty Preference Thoughtless Mindless Reputation Random Unplanned Liking Want Brand name Impulse Curiosity Senses Trial Variety Feel good Change Fun Image Self-esteem Style Personality Status Social approval Emotion Fashion
Dimension 2 –1.60 –1.60 –1.48 –1.47 –1.36 –1.11 –1.06 –0.98 –0.91 –0.87 –0.86 –0.81 –0.79 –0.78 –0.77 –0.69 –0.62 –0.59 –0.58 –0.48 –0.19 0.08 0.11 0.22 0.30 0.41 0.42 0.47 0.55 0.74 0.79 0.84 0.91 0.91 1.05 1.12 1.14 1.15 1.19 1.27 1.35 1.51 1.52 1.55
Habit Routine Usual Past purchase Thoughtless Loyalty Familiarity Satisfaction Mindless Liking Preference Convenience Want Random Replacement Impulse Brand name Unplanned Self-esteem Feel good Personality Fun Time pressure Emotion Reputation Trial Curiosity Style Change Image Status Performance Variety Social approval Fashion Senses Utilitarian Problem solving Logic Sale Value Quality Price Comparison
–1.39 –1.32 –1.27 –1.23 –1.19 –1.19 –1.08 –1.03 –1.02 –0.84 –0.82 –0.70 –0.59 –0.50 –0.41 –0.41 –0.28 –0.20 –0.20 –0.17 –0.05 –0.05 –0.04 0.04 0.23 0.26 0.27 0.33 0.38 0.40 0.49 0.57 0.62 0.62 0.63 0.73 0.88 1.11 1.27 1.28 1.35 1.39 1.51 1.61
Dimension 3 Unplanned Curiosity Random Trial Time pressure Impulse Variety Change Fun Mindless Sale Convenience Thoughtless Price Replacement Problem solving Emotion Value Liking Want Habit Comparison Routine Utilitarian Quality Logic Usual Past purchase Senses Self-esteem Personality Familiarity Feel good Satisfaction Loyalty Fashion Preference Performance Social approval Status Style Reputation Image Brand name
–1.93 –1.93 –1.89 –1.88 –1.80 –1.79 –1.59 –1.58 –1.56 –1.40 –1.26 –1.17 –0.80 –0.68 –0.45 –0.24 0.15 0.16 0.20 0.28 0.33 0.34 0.38 0.51 0.54 0.55 0.55 0.59 0.62 0.85 0.85 0.86 0.86 0.89 0.90 1.04 1.11 1.17 1.25 1.30 1.31 1.44 1.46 1.47
mindless
thoughtless
convenience
time pressure
sale
price
change trial fun
impulse random
unplanned
curiosity
variety
A. Spontaneous purchase behaviors
preference satisfaction familiarity past purchase loyalty usual routine habit
replacement
performance
utilitarian
quality value logic problem solving
comparison
want liking
brand name
feel good
personality self-esteem
senses fashion social approval status image style reputation emotion
B. Deliberate purchase behaviors
Figure 1.1 Multidimensional Scaling Solution of 44 Purchase Behaviors (Study 1)
22
purchase behavior and a proposal for a new typology
23
Figure 1.2 The Purchase Cube: A Three-Dimensional Typology of Purchase Behaviors
Deliberate purchases High purchase involvement
Spontaneous purchases
Low purchase involvement
Functional Psycho-social purchases purchases Deliberate purchases
Spontaneous purchases Promotional purchases
Exploratory purchases
Casual purchases
Impulsive purchases
Extended purchase decision making Repetitive purchases
Symbolic purchases
Hedonic purchases
Note: Based on Baumgartner (2002).
Starting with the spontaneous half of the purchase cube, purchases that are based mostly on functional considerations and demand a fair amount of purchase involvement are called promotional purchases. Buying something because the price is good and buying something because it’s on sale represent this type of purchase behavior. Purchases that are psycho-social in nature and above average in purchase involvement are called exploratory purchases. Variety-seeking behavior, purchases motivated by curiosity, and trial purchases are classified in this category. Buying something just for the fun of it is also included in this group in the cluster analysis, although it
24
Hans Baumgartner
is also close to impulse purchases in Figure 1.1 and may be considered an example of hedonic purchases (to be discussed later). Functional, uninvolving purchases are called casual purchases. They include mindless purchases, purchases motivated by convenience, and purchases made under time pressure, although the latter are more involving than purchases that are given little or no thought. Finally, psycho-social purchases that are below average on purchase involvement will be referred to as impulsive purchases. Besides unplanned purchases, this category also includes purchases that are made more or less randomly, which is consistent with the often capricious nature of impulsive buying behavior. Turning to the deliberate half of the purchase cube, high-involvement purchase behavior based on functional criteria is called extended purchase decision making. In addition to purchases for which utilitarian and logical considerations are important, this group includes purchases that entail a careful comparison of alternatives, purchases that necessitate problem-solving activities, and purchases that involve an evaluation of the performance, quality, and value of the available alternatives. Functional purchases that are low in involvement are called repetitive purchases. Based on the cluster analysis results, two subgroups can be distinguished. One group includes purchases of products that the consumer knows and likes, and thus buys with some degree of loyalty. The other group includes habitual or routine purchases. Purchases that are made for psycho-social reasons and are somewhat higher in involvement are called symbolic purchases. Again, two categories of purchase behaviors can be distinguished. One deals with social motivations for buying things. This includes buying a brand to project a certain image, making a purchase because of the status it confers or because it meets with social approval, buying products that are fashionable or in style, and purchasing brands because of their name or reputation. The other taps more psychological purchase motives, such as buying a product because it expresses your personality, making a purchase for emotional reasons, or buying something because it makes you feel good about yourself. Sensory purchases are also included in this latter category. Finally, psycho-social purchases that are relatively low in involvement are called hedonic purchases. This includes buying something because you like it and buying something because it is what you want. Purchase behaviors such as buying something because it makes you feel good and buying something because of the way it looks, sounds, feels, tastes, or smells would also seem to belong in the hedonic group, and in fact in the cluster analysis solution based on Ward’s method, the psychologically motivated symbolic purchase behaviors (feel good, self-esteem, personality, emotion, and senses) are grouped together with the two hedonic purchase behaviors. However, in the MDS solution this more inclusive category of hedonic purchases is not well separated from the socially motivated symbolic purchase behaviors. Some additional comments are in order to illuminate the meaning of the purchase behaviors distinguished in the purchase cube. First, one strength of the proposed typology seems to be that it clarifies the differences between symbolic purchases, exploratory purchases, hedonic purchases, and impulsive purchases, although additional empirical research is called for to substantiate these differences. Symbolic purchases and exploratory purchases are similar in that psycho-social motives are the primary motivators and both are relatively high in involvement. However, they differ in whether the purchase is made deliberately or spontaneously. The same is true for hedonic and impulsive purchases, which are stimulated by psycho-social motives and are relatively low in involvement, but differ in whether they are made deliberately or spontaneously. It should be noted that hedonic is used in a specific sense in the present context, referring not simply to purchases made for psycho-social reasons (which is one meaning of hedonic), but more specifically to purchases that are also deliberate and relatively low in involvement. Second, promotional purchases are similar to purchases made through extended purchase
purchase behavior and a proposal for a new typology
25
decision making, but a good price or a sale can turn an otherwise deliberate purchase into a more spontaneous one, which seems consistent with actual practice. Third, exploratory purchases (including purchases motivated by a desire for variety) are not made randomly or thoughtlessly and with a minimum of involvement (and if they are made that way, they are not exploratory purchases in the sense used here). Instead, they involve some effort and care because the goal is to complicate an otherwise routine buying process. Finally, at first sight it might seem odd that purchases can be high in involvement and spontaneous, or low in involvement and deliberate. However, a somewhat similar distinction is also made by Bagozzi (1983), who uses degree of involvement and extent of prior learning in his classification of consumer choice processes. A buying decision may be low in involvement (because it requires little effort and care), yet it is made deliberately based on prior experience with the product or brand. On the other hand, a purchase may be highly involving although the buying decision is made relatively spontaneously. More detail concerning the distinction between the two dimensions will be provided in the third study. Study 2: Validating the New Typology In the second study we attempted to validate the findings of Study 1 by using a different task. Participants were given a specific product category and were asked to rate how well each of the forty-four purchase behaviors (reasons for purchasing things) described the purchase of the product in question. Method Participants. The respondents who participated in Study 1 completed a second task, and these data were used to validate the new purchase typology. As in the first study, 298 participants provided complete data. Materials and Procedures. Respondents were asked to rate to what extent each of forty-four different reasons for purchasing things accurately described the purchase of a certain product. Participants were told that, if possible, it might be helpful to think of a time or times when they had purchased the product in question. A five-point rating scale ranging from “does not describe it at all” to “describes it very well” was used to collect the data. Forty different products were used in the study. They were selected from Figure 1.1 in Ratchford (1987), such that ten products represented each of the four quadrants of the FCB grid: think versus feel, and low versus high involvement. An attempt was made to choose products that spanned both dimensions of the FCB grid and that were meaningful to undergraduate students. The products are listed in Table 1.3. Participants responded to only one product, so that each product was rated by seven or eight people. Analysis. Two sets of analyses were performed on the data. First, multidimensional scaling (MDS) was used to investigate whether the classification of purchase behaviors proposed in Study 1 would emerge again. To this end, the data of respondents rating the same product were averaged and correlations were computed between each of the forty-four purchase behaviors across the forty products. These correlations were then converted to Euclidean distances using the transformation − U , and the resulting matrix of dissimilarities was submitted to nonmetric MDS. The badness of fit measures in one to four dimensions were .43, .13, .07, and .05, and the corresponding cor-
26
Hans Baumgartner
Table 1.3
Products Used in Study 2 High-involvement think products
High-involvement feel products
Low-involvement think products
Low-involvement feel products
Auto insurance Console TV Stereo component Portable TV 35 mm camera Instamatic camera Motor oil Headache remedy Washer/dryer Economy car
Sports car Expensive watch Wallpaper Hair coloring Perfume Wine for dinner party Complexion/face soap Coffee Eyeglasses Jeans
Insecticide Suntan lotion Salad oil Dry bleach Insect repellent Regular shampoo Liquid bleach Non-disposable razor Disposable razor Paper towel
Inexpensive watch Greeting card Pizza Deodorant soap Peanut butter Magazine Diet soft drink Regular soft drink Salty snack Regular beer
relations between the actual and fitted distances were .66, .95, .98, and .99. The three-dimensional solution was again judged to be the best representation of the data. Second, a confirmatory factor analysis was conducted on respondents’ ratings of each product across the forty-four purchase behaviors. Eight factors were specified a priori, and each purchase behavior was initially hypothesized to load on only one factor based on Figure 1.1 (that is, the circles are the eight factors, and the purchase behaviors within each circle are assumed to load on the factor). Based on goodness-of-fit considerations, the model was subsequently respecified slightly as described in more detail below. Results MDS Results. Using oblique Procrustes rotation, the initial three-dimensional MDS solution was rotated to make it maximally congruent with the three-dimensional MDS solution from Study 1. After rotation, the corresponding dimension coefficients from the two solutions correlated 0.91, 0.77, and 0.86 for functional vs. psycho-social purchases, low vs. high purchase involvement, and spontaneous vs. deliberate purchase behaviors, respectively. The highest correlation between noncorresponding dimensions across the two studies was 0.23, and the correlations between the dimension coefficients in Study 2 were -.02, -.09, and .27 (the highest correlation being between purchase involvement and deliberateness of the purchase). These findings indicate that, overall, the MDS solutions from the two studies are quite similar and that the three-dimensional representation of purchase behaviors uncovered in Study 1 could be replicated using a different task in which respondents were not explicitly asked to classify reasons for purchase. Figure 1.3 displays the results of Study 2 graphically. As in Figure 1.1, the findings are presented separately for spontaneous and deliberate purchase behaviors. The four types of spontaneous purchase behavior are generally situated in the hypothesized quadrant, except that exploratory purchases (particularly trial purchases and curiosity-motivated purchases) are too low in involvement so that they are not well separated from impulsive purchase behavior. Furthermore, problemsolving purchases and routine purchases are not deliberate enough and are thus classified with the spontaneous purchase behaviors. The four types of deliberate purchase behavior also occupy the hypothesized quadrant. The only discrepancies with Study 1 are that brand name purchases
time pressure
impulse
variety
change
unplanned
random
curiosity mindless routine trial
thoughtless
convenience
sale
price
problem solving
A. Spontaneous purchase behaviors
comparison
logic
familiarity satisfaction
habit
loyalty
usual
past purchase
brand name
performance
reputation
value
quality replacement
utilitarian
want preference
senses liking
fun
personality style emotion image status fashion self-esteem feel good social approval
B. Deliberate purchase behaviors
Figure 1.3 Multidimensional Scaling Solution of 44 Purchase Behaviors (Study 2)
27
28
Hans Baumgartner
and purchases based on the reputation of the brand are situated in the functional half of the grid, and that purchases based on a strong preference for the brand, sensory purchases, and purchases motivated by fun are classified as hedonic purchases. The latter result is actually more in line with a priori expectations than with the findings of the first study. Overall, the results of Study 1 are replicated well. Confirmatory Factor Analysis Results. The initial confirmatory factor model did not fit the data well (χ2874 = 2830.53, p < .0001, RMSEA = .087, CFI = .73, TLI = .70). The model was respecified by splitting the repetitive purchase factor into two subfactors—loyalty purchases, consisting of preference, satisfaction, familiarity, past purchase, loyalty, reputation, and brand name; and habitual purchases, consisting of usual, routine, and habit—and relaxing thirty-seven non-target loadings. Of these, eighteen were below .3. The resulting solution achieved an acceptable fit (χ2829 = 1611.83, p < .0001, RMSEA = .056, CFI = .89, TLI = .88), and the matrix of factor loadings is shown in Table 1.4. The loadings of items defining a particular factor are shown in boldface. The results are very similar to the MDS solution and for the most part consistent with the results of Study 1. The following discrepancies are apparent. First, buying something because it provides the best value loads most highly on the promotional purchase factor. This is consistent with one of the meanings of value identified by Zeithaml (1988)—value as the best price available. Buying something under time pressure has the highest loading on the impulse purchase factor, which also includes unplanned and random purchases. Making a replacement purchase loads about equally strongly on extended purchase decision making and habitual purchase behavior, which is consistent with its placement between these two purchase types in Figure 1.1. It is also consistent with the ambiguous nature of replacement purchases, because for infrequently bought, high-involvement products, a replacement purchase may entail extended decision making, while for frequently purchased, low-involvement products, replacement purchases are probably more habitual. Buying something because of the reputation of the brand and purchasing a brand because of its name load most strongly on the loyalty purchase factor, and sensory purchases have the highest loading on hedonic purchase behavior. Both findings are consistent with the MDS results. Although repetitive purchase behaviors are split into two subfactors, they are fairly highly correlated at .70. Several other interfactor correlations are substantial as well; impulsive and casual purchase behaviors are correlated .81, impulsive and exploratory purchases .66, symbolic purchases and exploratory purchases .66, and loyal purchases and extended decision making .61. Study 3: Validating the Interpretation of the Three Dimensions of the New Typology The three dimensions of the new typology of purchase behaviors were interpreted as functional vs. psycho-social purchases, low- vs. high-involvement purchases, and spontaneous vs. deliberate purchases. This interpretation was largely based on intuitive grounds and on what is known about the different types of purchase behaviors. In the case of functional vs. psycho-social purchases, the name chosen for this dimension is probably least controversial. Previous authors have frequently distinguished purchases that are motivated either by utilitarian considerations such as quality, price, familiarity, and convenience, or by nonfunctional concerns about how a product makes one feel or how it makes others feel about oneself. Since many of the purchase behaviors that respondents classified directly reflect these underlying motivations, it was relatively straightforward to interpret the first dimension. The interpretation of the remaining two dimensions is not nearly as clear cut. First, what does involvement in a purchase really mean? Does it refer to the personal relevance or importance of
purchase behavior and a proposal for a new typology
29
Table 1.4
Confirmatory Factor Analysis Results F1 Price Sale Value Convenience Mindless Thoughtless Variety Change Curious Trial Fun Unplanned Impulse Random Time pressure Comparison Quality Logic Problem solving Utilitarian Performance Preference Satisfaction Familiarity Past purchase Loyalty Reputation Name Usual Routine Habit Replacement Fashion Social approval Status Image Style Emotion Feel good Personality Self-esteem Want Liking Senses
0.75 0.77 0.54 0.26
F2
F3
F4
F5
F6
F7
F8
F9
0.21 0.54 0.79 0.92
–0.20 –0.32 0.64 0.78 0.91 1.04 0.74
–0.39 –0.54 0.69 0.88 0.73 0.65
0.18 0.14 0.17
0.34 0.38
0.20 0.67 0.62 0.93 0.45 0.96 0.68
–0.33 –0.24
0.25
–0.47 –0.45 –0.32
–0.53 0.55 0.72 0.81 0.78 0.64 0.68 0.59
–0.17
.22 –0.29 0.19
–0.40 0.48
–0.44
0.14 0.21 –0.13
0.19
0.35
0.20 0.32 0.27 0.63 1.12 0.77 0.53
0.15
0.82 0.73 0.80 0.77 0.78 0.58 0.77 0.81 0.86
0.19 0.27
0.78 0.86 0.48
Note: F1—promotional purchase behavior; F2—casual purchase behavior; F3—exploratory purchase behavior; F4—impulsive purchase behavior; F5—extended purchase decision making; F6—loyalty purchase behavior; F7—habitual purchase behavior; F8—symbolic purchase behavior; F9—hedonic purchase behavior.
30
Hans Baumgartner
a purchase in general, the riskiness of a purchase more specifically (as suggested by Rossiter, Percy, and Donovan 1991), or the amount of care required by a purchase and the degree of effort expended on it? Second, is the interpretation of the third dimension as spontaneous vs. deliberate purchases appropriate, and what other aspects of purchase behavior (such as previous experience with purchasing the product, prior planning of the purchase) can be used to characterize such purchases? Third, how do high-involvement purchases differ from deliberate purchases? In the first validation study, high-involvement products selected from Ratchford’s work on the FCB grid were also products that were purchased deliberately, so the issue of discrimination arises. To answer these questions and, more generally, to validate the interpretation of the three MDS dimensions uncovered in Studies 1 and 2, we conducted another study in which respondents were asked to rate the forty-four purchase behaviors on a number of specific scales. The scales were selected to represent different meanings of the three presumed MDS dimensions in terms of functional vs. psycho-social purchases, purchase involvement, and spontaneous vs. deliberate purchases. Method Participants. A total of seventy-seven undergraduate marketing students participated in the study to fulfill a course requirement. Seventy-five respondents, 56 percent of whom were female, provided complete data. Materials and Procedures. Participants were asked to rate the forty-four purchase behaviors from the previous two studies on seventeen scales. Since collecting seventeen ratings for each of forty-four purchase behaviors was thought to be too tiresome, the purchase behaviors were split in half and each respondent was randomly assigned to rate one of the halves. For one of the groups, “buying something compulsively” was added to the list of purchase behaviors. This purchase type was omitted in the earlier studies but was thought to be relevant to the classification of purchase behaviors. The seventeen scales used to characterize the purchase behaviors were: rational-emotional, thinkingfeeling, unexciting-exciting, not fun-fun, not pleasurable-pleasurable, boring-interesting, says nothingsomething about me, does not require-requires care, effortless-effortful, unimportant-important, not personally-personally relevant, not risky-risky, spontaneous-deliberate, does not follow-follows a script, unintended-intended, little-lot of previous experience, unplanned-planned. Respondents were asked to rate whether purchases of a given type tended to be rational or emotional, involved thinking or feeling, and so forth. A seven-point scale was used to collect the ratings. Analysis. Property fitting was employed to determine how well each of the seventeen scales described the dimensions identified in the previous two studies. Specifically, for each purchase behavior, we computed the average value of a purchase type (for example, buying a particular product because of the brand name) on a given scale (for example, rational-emotional) across all respondents who rated that purchase behavior. These average scores were then regressed on the dimension coefficients from Studies 1 and 2. The results can be interpreted like a loading matrix in factor analysis to interpret the meaning of each dimension in an MDS analysis. Results The property fitting results are shown in Table 1.5. High loadings of a dimension on a given scale have been highlighted to aid in the interpretation of the solution. The findings are very similar
Dim. 2 –0.23 –0.32 0.21 0.12 0.06 0.19 0.01 0.53 0.56 0.26 –0.07 0.10 0.10 0.09 0.08 –0.23 0.05
Dim. 1 0.62 0.57 0.61 0.58 0.45 0.54 0.26 0.02 0.03 –0.16 0.25 0.47 –0.34 –0.27 –0.26 –0.26 –0.28
Scale
Rational-emotional Thinking-feeling Unexciting-exciting Not fun-fun Not pleasurable-pleasurable Boring-interesting Says nothing-something about me Does not require-requires care Effortless-effortful Unimportant-important Not personally-personally relevant Not risky-risky Spontaneous-deliberate Does not follow-follows a script Unintended-intended Little-lot of previous experience Unplanned-planned
0.02 0.01 –0.17 –0.12 0.05 –0.14 0.22 0.29 0.12 0.31 0.22 –0.42 0.50 0.41 0.48 0.48 0.47
Dim. 3
Study 1
Multiple Regression of Bipolar Scale Ratings on Dimension Coefficients
Table 1.5
0.86 0.86 0.77 0.69 0.50 0.73 0.55 0.80 0.69 0.54 0.46 0.80 0.82 0.81 0.75 0.83 0.77
R2 0.52 0.48 0.50 0.46 0.35 0.45 0.21 0.05 0.06 –0.09 0.23 0.37 –0.26 –0.23 –0.20 –0.23 –0.20
Dim. 1
Study 2 –0.11 –0.21 0.18 0.05 –0.06 0.13 –0.06 0.45 0.51 0.15 –0.12 0.18 0.02 0.01 0.00 –0.28 0.01
Dim. 2
0 0 –0.12 –0.03 0.20 –0.06 0.32 0.32 0.11 0.42 0.32 –0.52 0.56 0.45 0.54 0.55 0.53
Dim. 3
0.79 0.78 0.62 0.54 0.49 0.60 0.61 0.76 0.64 0.59 0.55 0.70 0.71 0.66 0.62 0.65 0.64
R2
31
32
Hans Baumgartner
for the dimension coefficients from the two studies, so we will focus on the findings from the first study. The adjective pairs rational-emotional and thinking-feeling are most closely related to the first dimension, which confirms the interpretation of this dimension as contrasting functional (thinking) psycho-social (feeling) purchases. The adjective pairs unexciting-exciting, not fun-fun, not pleasurable-pleasurable, and boring-interesting are also substantially related to the first dimension. This indicates that functional purchases are rated as relatively unexciting and boring, and as low in fun and pleasure. Contrary to expectations, says nothing-something about me does not describe the first dimension and in fact is not strongly related to any dimension. Dimension 2 is mostly associated with the two items does not require-requires care and effortlesseffortful. It does not have a strong relationship with unimportant-important and not personallypersonally relevant. Interestingly, the regression coefficients for the item not risky-risky indicate that both psycho-social purchases and spontaneous purchases are thought to be risky. Riskiness is not related to involvement, so the meaning of involvement in the present case is somewhat different from the way it is usually conceptualized. Involvement in the present usage refers specifically to the degree of care required by a purchase or the amount of effort expended on the purchase. It does not refer to importance or personal relevance in general or the riskiness of a purchase. In other words, purchase involvement is not systematically related to the importance or relevance of the product per se (that is, even a high-involvement product may require little purchase effort if strong prior attitudes guide the purchase). In a similar way, riskiness of the purchase is not a systematic correlate of purchase involvement, because even though a high-risk product may require a more careful purchase process, not devoting much effort to the buying decision also makes the purchase risky. Dimension 3 is closely related to spontaneous-deliberate, does not follow-follows a script, unintended-intended, little-lot of previous experience, and unplanned-planned. Thus, as hypothesized, this dimension captures differences in deliberateness, prior planning, and previous experience. The slight negative correlations of these items with the first dimension indicate that there is some tendency for deliberate purchases to be functional or, conversely, for spontaneous purchases to involve feeling. Discussion and Conclusion A classification of forty-four different purchase behaviors reflecting various purchase motives yielded a typology of eight distinct types of purchase behavior based on three underlying dimensions (functional vs. psycho-social purchases, low vs. high purchase involvement, and spontaneous vs. deliberate purchases). The typology was replicated using a task in which respondents rated how well the forty-four different purchase behaviors described the purchase of forty different products, and the interpretation of the three underlying dimensions was validated by assessing the correspondence between the placement of the forty-four purchase behaviors along the three dimensions and direct ratings of the purchase behaviors on hypothesized descriptors of the three dimensions. It is instructive to contrast the purchase types on opposite poles of the diagonals through the purchase cube. There are four different contrast pairs: extended purchase decision making vs. impulsive purchase behavior, symbolic purchase behavior vs. casual purchase behavior, repetitive purchase behavior vs. exploratory purchase behavior, and hedonic purchase behavior vs. promotional purchase behavior. All four pairs differ on all three underlying dimensions. For example, while extended purchase decision making is functional, high in purchase involvement, and deliberate, impulsive purchase behavior is psycho-social, low in purchase involvement, and spontaneous. Prior
purchase behavior and a proposal for a new typology
33
research has emphasized the contrast between repetitive purchase behavior (habit) and exploratory purchase behavior (variety seeking), particularly in the modeling literature (e.g., Kahn, Kalwani, and Morrison 1986). The contrast between extended purchase decision making and impulsive purchase behavior is at least implicit in some of the literature on impulsive buying behavior (Puri 1996). The contrast pairs symbolic purchase behavior vs. casual purchase behavior and hedonic purchase behavior vs. promotional purchase behavior are less well established, and more detailed study of these pairs might shed light on the categories of casual and promotional purchase behaviors, which are probably the least clearly defined purchase types in the purchase cube. The next step in further developing the proposed typology of purchase behaviors would be to delineate each type in greater detail, review what we already know about each kind of buying behavior, and identify opportunities for future research. At present, discussions of the purchase process in consumer behavior textbooks are focused too much on extended purchase decision making, which characterizes only a small portion of the purchases that consumers actually make. Development of mini-theories for all eight types of purchase behaviors would seem to be a promising avenue for the construction of a more comprehensive theory of buyer behavior. Acknowledgments The author acknowledges helpful comments on previous versions of this chapter by Lisa Bolton, Kunter Gunasti, Meg Meloy, Bill Ross, and participants of the marketing proseminar at Penn State. References Andrews, J. Craig, Srinivas Durvasula, and Syed H. Akhter. 1990. “A Framework for Conceptualizing and Measuring the Involvement Construct in Advertising Research.” Journal of Advertising 19 (4), 27–40. Ahtola, Olli T. 1985. “Hedonic and Utilitarian Aspects of Consumer Behavior: An Attitudinal Perspective.” In Advances in Consumer Research, vol. 12, ed. Elizabeth C. Hirschman and Morris B. Holbrook, 7–10. Provo, UT: Association for Consumer Research. Assael, Henry. 1995. Consumer Behavior and Marketing Action, 5th ed. Cincinnati, OH: South-Western College Publishing. Babin, J. Barry, William R. Darden, and Mitch Griffin. 1994. “Work and/or Fun: Measuring Hedonic and Utilitarian Shopping Value.” Journal of Consumer Research 20 (March), 644–656. Bagozzi, Richard P. 1983. “A Holistic Methodology for Modeling Consumer Response to Innovation.” Operations Research 31, 128–176. Batra, Rajeev, and Olli T. Ahtola. 1991. “Measuring the Hedonic and Utilitarian Sources of Consumer Attitudes.” Marketing Letters 2 (2), 159–170. Baumgartner, Hans. 2002. “Toward a Personology of the Consumer.” Journal of Consumer Research 29 (September), 286–292. Beatty, Sharon E., and Scott M. Smith. 1987. “External Search Effort: An Investigation Across Several Product Categories.” Journal of Consumer Research 14 (June), 83–95. Belk, Russell W. 1985. “Issues in the Intention-Behavior Discrepancy.” In Research in Consumer Behavior, vol. 1, ed. Jagdish N. Sheth, 1–34. Greenwich, CT: Jai Press. Berger, David. 1986. “Theory into Practice: The FCB Grid.” European Research 14 (1), 35–46. Bettman, James R. 1979. An Information Processing Theory of Consumer Choice. Reading, MA: AddisonWesley. Celsi, Richard L., and Jerry C. Olson. 1988. “The Role of Involvement in Attention and Comprehension Processes.” Journal of Consumer Research 15 (September), 210–224. Chitturi, Ravindra, Rajagopal Raghunathan, and Vijay Mahajan. 2008. “Delight by Design: The Role of Hedonic Versus Utilitarian Benefits.” Journal of Marketing 72 (May), 48–63. Claeys, C., A. Swinnen, and P. Vanden Abeele. 1995. “Consumers’ Means-End Chains for ‘Think’ and ‘Feel’
34
Hans Baumgartner
Products.” International Journal of Research in Marketing 12 (October), 193–208. Clarke, Keith, and Russell W. Belk. 1979. “The Effects of Product Involvement and Task Definitions on Anticipated Consumer Effort.” In Advances in Consumer Research, vol. 6, ed. William L. Willkie, 313–318. Ann Arbor, MI: Association for Consumer Research. Copeland, Melvin T. 1924. Principles of Merchandising. Chicago and New York: A.W. Shaw. Crites, Stephen L., Jr., Leandre R. Fabrigar, and Richard E. Petty. 1994. “Measuring the Affective and Cognitive Properties of Attitudes: Conceptual and Methodological Issues.” Personality and Social Psychology Bulletin 20, 619–634. Crowley, Ayn E., Eric R. Spangenberg, and Kevin R. Hughes. 1992. “Measuring the Hedonic and Utilitarian Dimensions of Attitudes Toward Product Categories.” Marketing Letters 3 (3), 239–249. Dhar, Ravi, and Klaus Wertenbroch. 2000. “Consumer Choice Between Hedonic and Utilitarian Goods.” Journal of Marketing Research 37 (February), 60–71. Dichter, Ernest. 1960. The Strategy of Desire. Garden City, NY: Doubleday. Engel, James F., David T. Kollat, and Roger D. Blackwell. 1968. Consumer Behavior. New York: Holt, Rinehart, and Winston. Fennell, Geraldine. 1978. “Consumers’ Perceptions of the Product-Use Situation.” Journal of Marketing 42 (April), 38–47. Gardner, B.B., and Sidney J. Levy. 1955. “The Product and the Brand.” Harvard Business Review 33 (MarchApril), 33–39. Gensch, Dennis H., and Rakshekhar G. Javalgi. 1987. “The Influence of Involvement on Disaggregate Attribute Choice Models.” Journal of Consumer Research 14 (June), 71–82. Greenwald, Anthony G., and Clark Leavitt. 1984. “Audience Involvement in Advertising: Four Levels.” Journal of Consumer Research 11 (June), 581–592. Hirschman, Elizabeth, and Morris B. Holbrook. 1982. “Hedonic Consumption: Emerging Concepts, Methods, and Propositions,” Journal of Marketing 46 (Summer), 92–101. Holbrook, Morris B., and Elizabeth C. Hirschman. 1982. “The Experiential Aspects of Consumption: Consumer Fantasies, Feelings, and Fun.” Journal of Consumer Research 9 (September), 132–140. Houston, Michael J., and Michael L. Rothschild. 1978. “Conceptual and Methodological Perspectives on Involvement.” In Research Frontiers in Marketing: Dialogues and Directions, ed. S. C. Jain, 184–187. Chicago, IL: American Marketing Association. Howard, John A. 1977. Consumer Behavior: Application of Theory. New York: McGraw-Hill. ———. 1989. Consumer Behavior in Marketing Strategy. Englewood Cliffs, NJ: Prentice Hall. ———. 1994. Buyer Behavior in Marketing Strategy, 2nd ed. Englewood Cliffs, NJ: Prentice Hall. Howard, John A., and Jagdish N. Sheth. 1969. The Theory of Buyer Behavior. New York: John Wiley & Sons. Hupfer, Nancy T., and David Gardner. 1971. “Differential Involvement with Products and Issues: An Exploratory Study.” In Proceedings of the 2nd Annual Conference of the Association for Consumer Research, ed. David Gardner, 262–270. College Park, MD: Association for Consumer Research. Johar, J.S., and M. Joseph Sirgy. 1991. “Value-Expressive Versus Utilitarian Advertising Appeals: When and Why to Use Which Appeal.” Journal of Advertising 20 (September), 23–33. Kahn, Barbara, Manohar U. Kalwani, and Donald G. Morrison. 1986. “Measuring Variety-Seeking and Reinforcement Behavior Using Panel Data.” Journal of Marketing Research 23 (May), 89–100. Kassarjian, Harold H. 1978. “Presidential Address, 1977: Anthropomorphism and Parsimony.” In Advances in Consumer Research, vol. 5, ed. H. Keith Hunt, xiii-xiv. Ann Arbor, MI: Association for Consumer Research. ———. 1981. “Low Involvement—A Second Look.” In Advances in Consumer Research, vol. 8, ed. Kent B. Monroe, 31–34. Ann Arbor, MI: Association for Consumer Research. ———. 1994. “Scholarly Traditions and European Roots of American Consumer Research.” In Research Traditions in Marketing, ed. Gilles Laurent, Gary L. Lilien, and Bernard Pras, 265–287. Boston, MA: Kluwer Academic Publishers. Katz, Daniel. 1960. “The Functional Approach to the Study of Attitudes.” Public Opinion Quarterly 24, 163–204. Kim, Chung K., and Kenneth R. Lord. 1991. “A New FCB Grid and Its Strategic Implications for Advertising.” In Proceedings of the Annual Conference of the Administrative Sciences Association of Canada, ed. Tony Schellinck, 51–60. Niagara Falls, Ontario: Administrative Sciences Association of Canada. Krugman, Herbert E. 1965. “The Impact of Television Advertising: Learning Without Involvement.” Public Opinion Quarterly 29 (Fall), 349–356.
purchase behavior and a proposal for a new typology
35
Lastovicka, John L., and David M. Gardner. 1979. “Components of Involvement.” In Attitude Research Plays for High Stakes, ed. John C. Maloney and Bernard Silverman, 53–73. Chicago: American Marketing Association. Laurent, Gilles, and Jean-Noel Kapferer. 1985. “Measuring Consumer Involvement Profiles.” Journal of Marketing Research 22 (February), 41–53. Levy, Sidney J. 1959. “Symbols for Sale.” Harvard Business Review 37 (July-August), 117–124. Lutz, Richard J. 1979. “A Functional Theory Framework for Designing and Pretesting Advertising Themes.” In Attitude Research Plays for High Stakes, ed. John C. Maloney and Bernard Silverman, 37–49. Chicago American Marketing Association. Millar, Murray C. and Abraham Tesser. 1986. “Thought-induced Attitude Change: The Effects of Schema and Commitment.” Journal of Personality and Social Psychology, 51 (August), 259–269. Mittal, Banwari. 1988. “The Role of Affective Choice Mode in the Consumer Purchase of Expressive Products.” Journal of Economic Psychology 9, 499–524. ———. 1989. “A Theoretical Analysis of Two Recent Measures of Involvement.” In Advances in Consumer Research, vol. 16, ed. Thomas K. Srull, 697–702. Provo, UT: Association for Consumer Research . Mittal, Banwari, Brian Ratchford, and Paul Prabhakar. 1990. “Functional and Expressive Attributes as Determinants of Brand-Attitude.” In Research in Marketing vol. 10, ed. Jagdish N. Sheth, 135–155. Greenwich, CT: JAI Press. Nicosia, Francesco M. 1966. Consumer Decision Processes: Marketing and Advertising Implications. Englewood Cliffs, NJ: Prentice-Hall. Okada, Erica. 2005. “Justification Effects on Consumer Choice of Hedonic and Utilitarian Goods.” Journal of Marketing Research 42 (February), 43–53. Olshavsky, Richard W., and Donald H. Granbois. 1979. “Consumer Decision Making—Fact or Fiction?” Journal of Consumer Research 6 (September), 93–100. Park, C. Whan, Bernie J. Jaworski, and Deborah J. MacInnis. 1986. “Strategic Brand Concept-Image Management.” Journal of Marketing 50 (October), 135–145. Park, C. Whan, and Banwari Mittal. 1985. “A Theory of Involvement in Consumer Behavior: Problems and Issues.” In Research in Consumer Behavior, vol.1, ed. Jagdish N. Sheth, 201–231. Greenwich, CT: JAI Press. Park, C. Whan, and S. Mark Young. 1986. “Consumer Response to Television Commercials: The Impact of Involvement and Background Music on Brand Attitude Formation.” Journal of Marketing Research 23 (February), 11–24. Petty, Richard E., John T. Cacioppo, and David Schumann. 1983. “Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement.” Journal of Consumer Research 10 (September), 135–146. Puri, Radhika. 1996. “Measuring and Modifying Consumer Impulsiveness: A Cost-Benefit Accessibility Framework.” Journal of Consumer Psychology 5 (2), 87–114. Putrevu, S., and Kenneth R. Lord. 1994. “Comparative and Noncomparative Advertising: Attitudinal Effects Under Cognitive and Affective Involvement Conditions.” Journal of Advertising 23 (2), 77–90. Ratchford, Brian T. 1987. “New Insights about the FCB Grid.” Journal of Advertising Research, 27 (August/ September), 24–38. Ratchford, Brian T., and Richard Vaughn. 1989. “On the Relationship Between Motives and Purchase Decisions: Some Empirical Approaches.” In Advances in Consumer Research, vol. 16, ed. Thomas K. Srull, 293–299. Provo, UT: Association for Consumer Research. Ray, Michael L., Alan G. Sawyer, Michael L. Rothschild, Roger M. Heeler, Edward C. Strong, and Jerome R. Reed. 1973. “Marketing Communication and the Hierarchy of Effects.” In New Models for Communication Research, ed. P. Clarke, 147–176. Beverly Hills, CA: Sage. Rossiter, John R., and Larry Percy. 1997. Advertising Communications and Promotion Management, 2nd ed. New York: McGraw Hill. Rossiter, John R., Larry Percy, and Robert J. Donovan. 1991. “A Better Advertising Planning Grid.” Journal of Advertising Research 31 (October/November), 11–21. Shavitt, Sharon. 1989. “Products, Personalities and Situations in Attitude Functions: Implications for Consumer Behavior.” In Advances in Consumer Research, vol. 16, ed. Thomas K. Srull, 300–305. Provo, UT: Association for Consumer Research. ———. 1990. “The Role of Attitude Objects in Attitude Functions.” Journal of Experimental Social Psychology 26 (March), 124–148.
36
Hans Baumgartner
Shavitt, Sharon, and Tina M. Lowrey. 1992. “Attitude Functions in Advertising Effectiveness: The Interactive Role of Product Type and Personality Type.” In Advances in Consumer Research, vol. 19, ed. John F. Sherry Jr. and Brian Sternthal, 323–328. Provo, UT: Association for Consumer Research. Shavitt, Sharon, Suzanne Swan, Tina M. Lowrey, and Michaela Wänke. 1994. “The Interaction of Endorser Attractiveness and Involvement in Persuasion Depends on the Goal That Guides Message Processing.” Journal of Consumer Psychology 3 (2), 137–162. Sheth, Jagdish N., and P.S. Raju. 1974. “Sequential and Cyclical Nature of Information Processing Models in Repetitive Choice Behavior.” In Advances in Consumer Research, vol. 1, ed. Scott Ward and Peter Wright, 348–358. Urbana, IL: Association for Consumer Research. Sheth, Jagdish N., Bruce I. Newman, and Barbara L. Gross. 1991. Consumption Values and Market Choices: Theory and Applications. Cincinnati, OH: South-Western Publishing. Shiv, Baba, and Alexander Fedorikhin. 1999. “Heart and Mind in Conflict: The Interplay of Affect and Cognition in Consumer Decision Making.” Journal of Consumer Research 26 (December), 278–292. Slama, Mark E., and Armen Tashchian. 1985. “Selected Socioeconomic and Demographic Characteristics Associated with Purchasing Involvement.” Journal of Marketing (Winter), 72–82. Smith, M.B., J.S. Bruner, and R.W. White. 1956. Opinions and Personality. New York: Wiley. Sproles, George B., and Elizabeth L. Kendall. 1986. “A Methodology for Profiling Consumers’ DecisionMaking Styles.” Journal of Consumer Affairs 20 (Winter), 267–279. Sproles, Elizabeth K., and George B. Sproles. 1990. “Consumer Decision-Making Styles as a Function of Individual Learning Styles.” Journal of Consumer Affairs, 24 (Summer), 134–147. Taylor, J.W. 1974. “The Role of Risk in Consumer Behavior.” Journal of Marketing 38 (April), 54–60. Vaughn, Richard. 1980. “How Advertising Works: A Planning Model.” Journal of Advertising Research 20 (October), 27–33. ———. 1986. “How Advertising Works: A Planning Model Revisited.” Journal of Advertising Research 26 (February/March), 57–66. Voss, Kevin E., Eric R. Spangenberg, and Bianca Grohmann. 2003. “Measuring the Hedonic and Utilitarian Dimensions of Consumer Attitude.” Journal of Marketing Research 40 (August), 310–320. Woods, Walter A. 1960. “Psychological Dimensions of Consumer Decision.” Journal of Marketing 24 (January), 15–19. Zaichkowsky, Judith L. 1985. “Measuring the Involvement Construct.” Journal of Consumer Research 12 (December), 341–352. Zeithaml, Valeria A. 1988. “Consumer Perceptions of Price, Quality, and Value: A Means-End Model.” Journal of Marketing 52 (3), 2–22.
Chapter 2
MEASURING CUSTOMER LIFETIME VALUE Models and Analysis Siddharth S. Singh and Dipak C. Jain
Abstract This chapter focuses on the literature related to customer lifetime value (CLV) measurement. It first discusses the issues related to the context of CLV measurement and presents a contextual framework for understanding and categorizing models of CLV. The chapter then reviews some prominent models for measuring customer lifetime value in different contexts and discusses the strengths and weaknesses of each. Finally, the chapter discusses the key issues that need to be addressed to advance the literature. Introduction The last few years have seen an explosion of research into customer lifetime value (CLV).1 This has followed the increased focus of firms on customer relationship management (CRM), where firms consider their interactions with customers over the entire duration of the customer-firm relationship, also called customer lifetime, to improve profitability. To evaluate strategies in CRM, firms need to measure their impact. This is where CLV enters the picture. It is used as a metric to evaluate actions of the firm (Borle, Singh, and Jain 2008; Gupta and Zeithaml 2006).2 Definition of Customer Lifetime Value (CLV) The literature has generally defined CLV in ways that differ subtly. For example, Dwyer (1997) defines lifetime value as the present value of the expected benefits (for example, gross margin) less the burdens (for example, direct costs of servicing and communicating) from customers. Kumar, Ramani, and Bohling (2004) define CLV as the sum of cumulated cash flows—discounted using the weighted average cost of capital—of a customer over his or her entire lifetime with the firm. Berger and Nasr (1998) quote Kotler and Armstrong (1996) to define a profitable customer as “a person, household, or company whose revenues over time exceed, by an acceptable amount, the company costs of attracting, selling, and servicing that customer.” They refer to this excess as customer lifetime value. Although they appear to be similar at first glance, it is important to note the differences in the definitions. In the first, by Dwyer (1997), expected benefits from a customer can be interpreted 37
38
Siddharth S. Singh and Dipak C. Jain
broadly to include both direct benefits through direct revenues from the customer and indirect benefits such as through word-of-mouth effect. Also, since this definition considers only the expected benefits and costs, it ignores the past (such as cost of customer acquisition and realized revenues from a customer). While the definition by Kumar, Ramani, and Bohling (2004) focuses on the cash flows from a customer and does not consider costs, that of Berger and Nasr (1998) does consider costs, including the cost of customer acquisition. As far as benefits from a customer are concerned, the CLV models developed so far generally focus on the revenue stream from a customer and do not account for the network benefits. The next step in the development of these models has to consider important factors such as word-ofmouth and other network effects (such as participation in customer communities) in valuing a customer. It is well recognized in the literature and the popular press that such network effects due to a customer add to the value of a customer and are part of CLV (e.g., Lee, Lee, and Feick 2006). While some researchers have attempted to deal with this issue by separating CLV from the network effects such as the word-of-mouth effect (Kumar, Petersen, and Leone 2007), this only underscores the issue that we raise here. Since word-of-mouth and other network effects impact the value of a customer to the firm, these should be part of CLV. The differences in the definitions of CLV—the basic foundation of this literature—highlight the need for developing clear definitions of key terms. We believe that the definition by Dwyer (1997) is broad enough to include important factors such as the word-of-mouth effect. Therefore, in this paper we will consider the CLV definition according to Dwyer. In case of prospects (that is, potential customers), we recommend using the term prospect lifetime value (PLV) instead of customer lifetime value (CLV), where PLV includes all the factors considered in CLV in addition to the cost of customer acquisition. In this chapter however, for simplicity we will henceforth use the term CLV to include both CLV and PLV. Measurement of CLV For CLV to be effectively used as a metric, firms need to measure it. Therefore, as a logical next step in the development of this literature, many researchers have focused on developing models to measure CLV. These models provide a way to estimate CLV given the nature of the customer-firm relationship and the data available. The models incorporate factors that affect CLV to the extent possible for an accurate measure. However, this is a challenging task. Figure 2.1 shows the main factors that influence CLV, underscoring the complexity underlying CLV and the challenges in measuring it accurately. To aid the users of CLV models, there is a need to categorize these models (or methods for estimating CLV) more clearly based on some reasonable criteria that can be used to easily choose between the alternative methods for any specific application. Understanding the strengths and limitations of each method within a proper context allows a user to apply them intelligently. We will discuss the contexts where some of the key proposed models can be used, the strengths and limitations of each model, and the challenges that lie ahead in further developing this literature. The chapter is organized as follows. The second section discusses various contexts of the customerfirm relationship and why an understanding of these contexts is important for the measurement of CLV. It also presents a contextual framework to organize the literature in a meaningful way. The following section discusses some prominent models proposed for measuring CLV in different contexts and the strengths and limitations of each. The fourth section discusses some important issues that need to be addressed in order to take the literature to the next level, and the final section summarizes the chapter.
MEASURING CUSTOMER LIFETIME VALUE
39
Figure 2.1 Factors Impacting CLV
Customer Lifetime Interpurchase Time Purchases Returns Marketing Activities of Firm Network Effects (e.g., Word-of-Mouth) Discount Rate
Customer Lifetime Value (CLV)
Cost of Customer Acquisition Acquisition Cost of Customer Retention Retention Cost of Returns Cost of Marketing Activities Activities
Context of Customer Lifetime Value Measurement A review of the customer lifetime value literature (Borle, Singh, and Jain 2008; Fader, Hardie, and Lee 2006a; Venkatesan and Kumar 2004; Reinartz and Kumar 2000 and 2003; Bolton 1998; Bhattacharya 1998) soon reveals that the context of CLV measurement plays a key role in the methods proposed for measuring CLV and the issues that become important both from a modeling point of view and the managerial point of view. By context, we mean the context of the customer-firm relationship that generated the data to be used for estimating CLV. From a modeling perspective, the context defines the data available to estimate a CLV model, and from a managerial perspective, the context defines the issues that become important in managing customer profitability. The importance of the context of customer-firm relationship for modeling CLV, however, is not well understood by either managers or educators, as is evident commonly both in the use and teaching of CLV (Fader, Hardie, and Lee 2006a). We feel that this lack of understanding is a significant roadblock to providing prescriptions for use of different models in different managerial situations. To understand the importance of context, consider the following examples: (a) grocery purchases from a retail outlet, (b) purchases from a direct marketing catalog, (c) child-care services, and (d) subscription services (for example, magazine and cable television). In most contexts such as (a) and (b), the firm does not know when a customer defects. After purchasing, if a customer does not purchase for a long time, what does that mean? Is the customer still “active,” or has the customer defected? One key implication of this uncertainty is that the firm has to manage both customer retention and purchases to enhance the lifetime value of a customer. The lack of knowledge of customer defections (that is, customer lifetime duration) also makes modeling CLV in this context more challenging. In contexts such as (c) (child-care services) the firm would know with certainty when a customer defects. Also, the firm knows that all customers will defect at some point in a few years time no matter what it does to retain them, because children grow up and do not require child-care services. Therefore, the focus of the firm is more on up-selling/cross-selling (for example, selling additional services). Similarly in contexts such as (d) (subscription services) a firm would know when a customer defects. Here longer customer lifetime implies higher lifetime value. Therefore, the primary focus of firms is on managing customer retention. In both (c) and (d), knowledge of customer lifetimes makes it relatively easier to estimate CLV. Note that customers have to be acquired before a firm can sell to them and retain them. Therefore, customer acquisition is always an important issue.
40
Siddharth S. Singh and Dipak C. Jain
Let us now consider the revenue stream from customers in these example contexts. The revenue stream from a customer in the grocery purchases context (a) or direct marketing catalog purchase (b) is likely to be stochastic from the firm’s point of view, and therefore customer lifetimes and profitability relationship are not known to the firm. In the child-care (c) and subscription context (d), the revenue stream is more predictable. Thus, longer customer lifetime implies higher CLV.3 Clearly, the context dictates the type of data available to the researcher/firm (for example, whether completed lifetimes are known or not) and the issues that become important. Despite the importance of context in CLV research, there has been little systematic study of the contexts. Various researchers have attempted to divide the context into categories that conveniently suited their purpose. The three most common context classifications are: lost-for-good and always-a-share, membership and nonmembership, and contractual and noncontractual. Lost-for-Good and Always-a-Share Conceived by Jackson (1985) (see also Dwyer 1997), this categorization of industrial buyers proposes that in a lost-for-good case customers remain in business with the firm until they defect, and once they defect, they never return—they are lost for good (such as in telecommunication system purchases where buyers typically select one vendor). In the always-a-share case, customers might do business with multiple vendors. Therefore, even if a customer temporarily does not do business with a firm, he or she has some probability of doing it in future (for example, purchase of office supplies). Dwyer (1997) explains the essence of this categorization and applies it to other businesses, suggesting that in an always-a-share case, customers can generally evaluate products, adjust their purchase volume, try new vendors relatively easily, and buy heavily even after periods of inactivity. An example of this case is catalog buying. In the lost-for-good case, on the other hand, more complex products are involved and buyers are looking for solutions to complex problems through the purchase. These customers also depend upon high quality and level of services that come with the purchase. It suggests that settings involving significant resource commitment on the part of the customer or contracts such as financial services and magazine subscriptions are examples of lost-for-good cases. Customers who defect return to the pool of prospects for possible reacquisition in future. Although this categorization has been used by researchers (e.g., Rust, Lemon, and Zeithaml 2004; Pfeifer and Carraway 2000; Venkatesan and Kumar 2004), it has limitations and has not been developed beyond what was presented by Dwyer (1997). For example, consider product categories such as music, books, and packaged foods that can be purchased commonly through numerous retail outlets as well as through several music clubs, book clubs, and purchase clubs such as Sam’s Club. In the latter case, the nature of these products still implies an always-a-share situation, but there is a contract involved. Thus while the product characteristics suggest putting them in an always-a-share context, the contract and service characteristics suggest that they belong to a lost-for-good context, underscoring the limitations of this categorization. From the perspective of modeling CLV, when a firm has data only on its interactions with its own customers, which is a very common scenario, the usefulness of this categorization is not clear. When data concerning customer purchases from multiple vendors is available (such as survey data, panel data), an always-a-share case can be modeled using something akin to the brand-switching matrix commonly used in marketing. An example of such a model for estimating customer equity is Rust, Lemon, and Zeithaml (2004). Hazard rate models can be used to model lifetimes and subsequently CLV in a lost-for-good case. The important highlight in this categorization is the consideration of competition that is ignored in other classification schemes.
MEASURING CUSTOMER LIFETIME VALUE
41
Membership and Nonmembership This classification puts each context into one of the two categories—membership and non-membership. In the membership category, customers have to join a firm as a member before purchasing from the firm or making use of its services. The other contexts fall under the non-membership category. Examples are club memberships such as purchase clubs (such as Sam’s Club and Costco), health clubs (for example, LA Fitness), and direct marketing clubs (such as book clubs or music clubs). Examples of studies in marketing focusing on the membership context are Borle, Singh, and Jain (2008), Bhattacharya (1998), and Bhattacharya, Rao, and Glynn (1995). Although sometimes useful, this categorization has issues that make it less useful than others because it is based on neither clear customer behavioral differences that have significant implications for measuring CLV nor information that a firm has about its customer that could be used to estimate CLV. For example, membership is a form of contract; however, all contracts are not memberships. Does a customer-firm relationship in a membership context lead to different customer purchase behavior than similar nonmembership contexts where a purchase contract between a customer and firm exists? Consider a health club that requires membership (a membership context) and a newspaper subscription (a non-membership context). These are similar in many respects as far as measuring CLV is concerned. In both cases, the firm knows when a customer defects and it is relatively easier to make an accurate forecast of the revenue stream from customers. A context categorization that ignores such similarity is less useful. Contractual and Noncontractual This is the most popular categorization of contexts in the CLV literature (e.g., Fader, Hardie, and Lee 2005a and 2005b; Venkatesan and Kumar, 2004; Reinartz and Kumar, 2000 and 2003). Reinartz and Kumar (2000) define a contractual context as one in which the expected revenues can be forecast fairly accurately and, given a constant usage of service, increasing cumulative profits over the customer’s lifetime would be expected. Noncontractual contexts are those in which the firm must ensure that a relationship stays alive because customers typically split their category expenses among several firms, for example, department store purchases or mail-order purchases in the catalog and direct marketing industry. The article further says that customer lifetime analyses have been conducted typically in contractual settings such as magazine subscriptions and cellular telephone services. The characteristics that define these categories are not clear. For example, the sharing of a customer’s business by multiple vendors is considered a typical characteristic of a noncontractual context, but nothing is said about this sharing in a contractual context. It seems that the categorization as proposed in its early development is loosely modeled after always-a-share and lost-for-good context categories, where the former is similar to the noncontractual category and the later is similar to contractual. One can easily find situations (such as a purchase club, for example, Costco or Sam’s Club) in which the customer-firm relationship possesses characteristics of both categories. A customer could be a member of multiple purchase clubs that sell similar items. Other studies that have used this classification have attempted to refine it further. Reinartz and Kumar (2003) say that “ areas in need of research are noncontractual relationships—relationships between buyers and sellers that are not governed by a contract or membership. . . . Given that switching costs are low and customers choose to interact with the firms at their own volition, this is a nontrivial question for noncontractual relationships.” Thus this article considers a noncontractual relationship as one where there is (a) no contract or membership involved, that is, firms do not
42
Siddharth S. Singh and Dipak C. Jain
observe customer defection; (b) switching costs are low; and (c) customers choose to interact with the firms as and when they desire. Again we can see that a noncontractual context is considered similar to an always-a-share context. Fader, Hardie, and Lee (2005b) say “ . . . noncontractual setting (i.e., where the opportunities for transactions are continuous and the time at which customers become inactive is unobserved),” and Venkatesan and Kumar (2004) consider the issues that become important for firms in different contexts and say that in contractual contexts managers are interested in predicting customer retention, while in noncontractual contexts the focus is more on predicting future customer activity and the predicted contribution margin from each customer because there is always a chance that the customer will purchase in the future. Therefore, implicitly Venkatesh and Kumar consider a contractual context where longer customer lifetime implies higher customer value. Borle, Singh, and Jain (2008) point out this lack of clarity in context categorization in the literature and attempt to summarize the definition of contractual and noncontractual contexts based on articles in the literature. However, they dwell on this issue just enough to highlight the unique characteristics of the data and do not attempt to provide a better classification scheme. As we can see, researchers have not defined contractual and noncontractual contexts very clearly, and this classification needs further refinement. This is perhaps because no article has focused primarily on the context issue itself except Fader, Hardie, and Lee (2006a), which we discuss later. The definitions of contexts in each paper have been provided on an ad hoc basis just to define the parameters of the particular study, leading to the lack of clarity that we have found. Based on several articles in the CLV literature, we can summarize the definition of contractual and noncontractual contexts as follows. In a contractual context, the firm knows when a customer defects; thus the lifetime of each customer is known with certainty once that customer defects. The buyer-seller relationship is governed by a contract or membership. The duration of the customer-firm relationship is closely tied to the revenue stream from the customer; thus customer lifetime is related to the CLV such that longer customer lifetime means higher CLV for the firm. Therefore, customer retention becomes the firm’s primary focus in managing customer relationships. Examples of such contexts would be utilities, insurance services, cable television, magazine and newspaper subscription services, some cellular services, and some club memberships. Studies in contractual contexts include Borle, Singh, and Jain (2008), Thomas (2001), Bolton (1998), Bhattacharya (1998), and Bhattacharya, Rao, and Glynn (1995). A noncontractual context is one in which a firm does not know when a customer defects, and the relationship between customer lifetime and purchase behavior is uncertain (for example, retailing, catalog purchases, and charity fund drives). Therefore, any model that measures CLV or investigates the factors driving CLV has to make some assumption about customer lifetime with the firm. A common approach is to assume exponential customer lifetime distribution (e.g., Schmittlein, Morrison, and Colombo 1987; Schmittlein and Petersen 1994; Fader, Hardie, and Lee 2005a), which may be restrictive in many situations. Another alternative is to assume a fixed time horizon for CLV prediction. However, such a judgment is likely to vary from one situation to another, leading to biases in estimation of CLV. Are the contractual and noncontractual categories as defined in the literature so far adequate? We find this not to be the case. Note that there is heterogeneity in contexts within each category that has significant implications for estimating CLV. For example, within the contractual context, one can have situations such as a newspaper subscription or a cable TV subscription (where longer customer lifetime implies higher customer lifetime value) as well as membership contracts (such as purchase clubs) where customer spending has an unknown relationship with customer
MEASURING CUSTOMER LIFETIME VALUE
43
Table 2.1
Context Categorization Proposed by Fader, Hardie, and Lee (2006a) Opportunities for Transactions
Noncontractual
Contractual
Continuous
Grocery purchases Doctor visits Hotel stays Event attendance
Credit card Student meal plan Cell phone usage Magazine subscriptions
Discrete
Prescription refills Charity fund drives
Insurance policy Health club
Source: As proposed by Fader, Hardic, and Lee 2006.
lifetime duration. Therefore, there is a need for further refinement of the categories to address these issues. Fader, Hardie, and Lee (2006a) attempt to refine this categorization by considering contexts along two dimensions: “continuous time and discrete time” and “contractual and noncontractual.” They define the notion of “continuous time” where the customer-firm transaction can occur at any time, and “discrete time” where transactions can occur only at discrete points in time. And they define a contractual case as one in which the time when a customer becomes inactive is observed, and a noncontractual case as one in which the time when a customer becomes inactive is not observed by the firm.4 Based on these two dimensions (that is, types of relationship with customers), they classify the customer base as shown in Table 2.1. Although this categorization takes a significant step forward by using two dimensions, it can be refined further. It considers opportunities of transactions and customer defections because variations in customer behavior and the data available with the firm along these dimensions have significant implications for measuring CLV and the issues that become important for managers. However, spending by customers during a purchase occasion that is equally important has not been considered. In Table 2.1 let us focus on the discrete-noncontractual cell, in which are both prescription refills and charity fund drives. However, for the firm involved in each case, the issues that are meaningful are not the same. In a charity fund drive, an organization soliciting funds has to make an attempt to get funds, and get more of them. The amount donated can vary significantly from one occasion to another and across donors. In the prescription refill case, generally the amount remains the same or more predictable for any customer and similar to other customers with same illness. From a modeling point of view as well, a model of CLV in the case of charity fund drive is likely to include more complexity to deal with the uncertainty in the amount raised from a donor. The prescription refill case can do with a much simpler approach for modeling the spending component of the CLV model. We believe that an adequate context categorization can result from a “bottom-up” approach that considers all the main drivers of CLV, namely, customer lifetime, purchase time, and spending. While the classification in Table 2.1 considers lifetime via the contractual-noncontractual dimension, and purchase time via the continuous-discrete dimension, it ignores the third critical dimension, which is customer spending. Let us consider the money spent by a customer during each purchase occasion as falling into one of the following two categories: Fixed and Variable. In the Fixed category, the amount spent on each occasion is predictable easily, that is, it is either the same or attains one of a few possible discrete values and thus can be predicted relatively easily.
44
Siddharth S. Singh and Dipak C. Jain
Table 2.2
Proposed Context Categorization Customer Lifetime Dimension
Purchase Time Dimension
Noncontractual
Contractual
Customer Spending Dimension
Customer Spending Dimension
Fixed
Variable
Fixed
Variable
Continuous
(A) Offline movie rental
(B) Grocery purchases Hotel stays Air travel
(C) Student meal plan
(D) Credit card Purchase clubs
Discrete
(E) Event attendance Prescription refills
(F) Charity fund drives
(G) Magazine subs Insurance policy Health club
(H) Cell phone/ PDA payments
Note: The examples in each cell are for illustration only and represent only the cases that are appropriate for the cell. Within each example, such as student meal plan, there could be examples that fall in another cell.
For example, a magazine subscriber pays the same amount per time period for the magazine. In the Variable spending category, a customer can spend any amount. For the firm in a context with Fixed spending, the issues of focus are the timing of purchase (purchase time) and/or customer defection (customer lifetime), whereas a firm in the Variable spending category needs to focus on customer spending as well. From the point of view of estimating CLV, the Variable spending category poses more challenges. Examples of studies where spending falls in the Variable category are Borle, Singh, and Jain (2008) (where spending per transaction for each customer is modeled as Lognormal) and Fader, Hardie, and Lee (2005a) (where spending per transaction is modeled as gamma-gamma). Table 2.2 represents our refinement of the categorization in Table 2.1. Our aim is to propose a customer-firm relationship framework that adds to the existing framework in a meaningful manner and is reasonably comprehensive. It is noteworthy that a customer-firm transaction might not be monetary. For example, a cell phone is used all the time, which is a nonmonetary transaction; however, payments for the phone service happen at fixed intervals. In developing the framework, we have focused on the monetary transactions as the key transactions of interest. This is reasonable because the timings of customer spending, as opposed to the timings of product usage,5 have a bearing on CLV. The context categorization presented in Table 2.2 is useful in understanding and prescribing models for use in a specific situation. Generally, it is most challenging to develop an accurate CLV model for context (B) and (D), where both the spending and purchase times are variable. Between these two, (B) presents a higher degree of difficulty due to lack of information about customer defection. It is noteworthy that the two broad categories of the customer lifetime dimension (contractual and noncontractual) are the most interesting for modeling CLV. The other two components of a CLV model (spending and purchase time) can be suitably modeled relatively easily given the significant work in the general marketing literature on revenues from a customer and interpurchase
MEASURING CUSTOMER LIFETIME VALUE
45
Table 2.3
Models of CLV Discussed Contractual contexts 1 2 3
Basic Structural Model of CLV (Jain and Singh 2002; Berger and Nasr 1998) Regression/Recency, Frequency, and Monetary (RFM) Models (e.g., Donkers, Verhoef, and de Jong 2007) Hazard Rate Models (Borle, Singh, and Jain 2008) Noncontractual contexts
1 2 3 4
Pareto/NBD (Schmittlein, Morrison, and Colombo 1987) Beta-Geometric/NBD or BG/NBD (Fader, Hardie, and Lee 2005b) Markov Chain Models (Pfeifer and Carraway 2000; Rust, Lemon, and Zeithaml 2004) Markov Chain Monte Carlo (MCMC) Data Augmentation Based Estimation Framework (Singh, Borle, and Jain 2008)
time. Therefore, in presenting the models for CLV, we focus on this lifetime dimension only. Any model proposed for a contractual or a noncontractual context can be modified for use in any subcategory within these respective categories. Models for Measuring Customer Lifetime Value In the general marketing literature, significant work has been done to investigate and forecast drivers of CLV such as customer lifetime, customer spending, interpurchase time, promotions, and so on (e.g., Bolton 1998; Chintagunta 1993; Allenby, Leone, and Jen 1999; Bhattacharya 1998; Jain and Vilcassim 1991). As a result, numerous possibilities exist for combining models for different drivers of CLV to get its estimate. In this section, we do not consider the individual models for a specific driver of CLV. Our focus here is on the prominent models proposed for estimating CLV. The CLV models that we discuss contain component models (or submodels) for the drivers of CLV considered. These submodels in turn are based on the extant research in the general marketing literature as mentioned earlier. Therefore, the CLV models discussed here have already considered the numerous possibilities for the component submodels, which justifies our focus on these models while ignoring the numerous other ad hoc possibilities. We use the framework in Table 2.2 as a guide. To facilitate the descriptions, we categorize the models into two broad categories based on the context discussed earlier, namely, models for contractual and noncontractual contexts. The models for each context can be modified suitably for use in any subcategory within the context. Table 2.3 presents the models that we discuss in this section, beginning with models for contractual contexts and followed by models for noncontractual contexts. Models for Contractual Contexts In this context, firms know the time of customer defections. While a customer is in business with a firm, it also has information on the realized purchase behavior of the customer. Donkers, Verhoef, and de Jong (2007) provide an excellent comparison of CLV models in a contractual context. Our discussion adds to their work. We first discuss the basic structural model of CLV.
46
Siddharth S. Singh and Dipak C. Jain
Basic Structural Model of CLV The basic model for CLV of a customer is (see Jain and Singh 2002):
&/9 = ∑L = Q
(5L − &L ) ,
+ G L −
(1)
where i = the period of cash flow from customer transaction, Ri = revenue from the customer in period i, Ci = total cost of generating the revenue Ri in period i, and n = total number of periods of projected life of the customer. Several variations of this basic model and their details can be found in Berger and Nasr (1998). Using the basic structural model, one simple way to estimate CLV is to take the expected customer lifetime, the average interpurchase time, and dollar purchase amounts observed in the past and use them to predict the present value of future customer revenues assuming these average values for the inputs in the CLV model. Such a heuristic model would be a quick and useful method to calculate CLV in the absence of any better model (Borle, Singh, and Jain 2008). Although the basic model (and its variations) is useful to understand the essence of CLV, there are several issues with this formulation that relate to the assumptions it makes about customer purchase behavior. For example, it assumes fixed and known customer lifetime duration, revenues (both times and amounts), and costs of generating revenues. These restrictive assumptions make the model unattractive in most cases. Examples of situations where this model can be useful are newspaper or magazine subscriptions where the revenues from a customer are known and occur at fixed known times. The length of the subscription, that is, customer lifetime, is the only unknown variable. The firm could use some rule of thumb for this length—such as a fixed time period of, say, three years—or take the average subscription period for its past customers to get this time. Such simplistic measures will be biased, but the low cost of applying this model might justify its use in the right situation. To develop more accurate models, one has to consider modeling customer lifetime to better reflect its variation across customers. A good method would be to use hazard rate models (Helsen and Schmittlein 1993), which can provide more accurate information about the expected lifetime of each existing customer. This lifetime in turn can be used in the CLV models along with the revenues and cost information to get an estimate of the lifetime value of each customer. The models discussed so far in this section are most suitable for cases that fall in cell (G) in Table 2.2. In other situations within the contractual context, the assumptions of fixed revenues and costs cannot be justified. Therefore, researchers have developed ways to make the basic model more flexible or to find alternatives to it. Regression/RFM Models The regression/RFM methods are commonly used techniques to score customers for a variety of purposes (such as targeting customers for a direct mail campaign). An RFM framework uses information on a customer’s past purchase behavior along three dimensions (its recency, frequency, and monetary value) to “score” customers. Regression methods can use these and other variables to score customers as well as estimate CLV of each customer. The scores are related to the expected potential customer purchase behavior and hence can be considered as another measure of customer value to the firm. Scores generally serve a similar purpose as the CLV measure. These methods have been used in noncontractual contexts as well; see, for example, Borle, Singh, and
MEASURING CUSTOMER LIFETIME VALUE
47
Jain (2008), Donkers, Verhoef, and de Jong (2007), Malthouse and Blattberg (2005), Reinartz and Kumar (2003), and Colombo and Jiang (1999). Hazard Rate Models Jain and Vilcassim (1991) and Helsen and Schmittlein (1993) provide a good introduction of hazard rate models in marketing. Hazard of an event means the risk of its occurring. Here, the event is customer defection or purchase. These models are well suited to model the risk of customer defection and the risk of purchase happening when data about these events is available. The estimates from these models in turn provide the expected duration of customer lifetime and interpurchase time. Depending upon the characteristics of customer spending, a suitable model can be used for its estimation. The estimates from these three submodels (lifetime, interpurchase time, and spending) can provide estimates of CLV. A prominent example of this approach in a contractual context is provided in Borle, Singh, and Jain (2008); it is henceforth referred to as the BSJ 2008 model. The BSJ 2008 Model. This hierarchical Bayesian model proposed by Borle, Singh, and Jain jointly predicts a customer’s risk of defection, risk of purchase, and spending at each purchase occasion. This information is then used to estimate the lifetime value of each of the firm’s customers at every purchase occasion. The firm can use this customer lifetime value information to segment customers and target them. Borle, Singh, and Jain develop and apply the model to data from a direct marketing company. They test the model and show that it performs significantly better in both predicting CLV and targeting valuable customers than a simple heuristic model, an advanced RFM model proposed by Reinartz and Kumar (2003), and two other models nested within the proposed BSJ 2008 model. The BSJ 2008 model is a joint model of the three dependent drivers of CLV, namely the interpurchase time, the purchase amount, and the probability of leaving, given that a customer has survived a particular purchase occasion (that is, the hazard rate or the risk of defection). The model for each of these three quantities is specified along with a correlation structure across these three submodels, leading to a joint model of these three quantities. In the model, the interpurchase time is assumed to follow a NBD process, the amount expended by a customer on a purchase occasion follows a lognormal process, and the hazard of lifetime—that is, the risk of customer defection in a particular interpurchase time period—is modeled using a discrete hazard model. The model also incorporates time-varying effects to improve the predictive performance. CLV Models for Noncontractual Contexts In this context, firms do not know whether a customer has defected or intends to remain in business with the firm. Since knowledge of customer lifetime is essential to estimate CLV, this becomes a challenging context for CLV measurement. Some users deal with this challenge by considering a future time period of a fixed duration, say three years, and estimating the net present value of the profits from a customer during this period. Regression models can be used to estimate future timing and amount of spending by the customer during this period. Past purchase behavior and customer characteristics are some of the variables that can be included as explanatory variables in the regression model. Such methods can be applied in many ways (e.g., Malthouse and Blattberg 2005). It is noteworthy that such ad hoc approaches are best suited for forecasting an outcome during the next time period. As the forecasting horizon increases, these methods produce more error.
48
Siddharth S. Singh and Dipak C. Jain
The following discussion focuses primarily on stochastic models of CLV that assume some underlying customer purchase behavior characteristics. These models assume that the observed customer purchase behavior is generated due to these latent customer characteristics. Such models have advantages over other approaches, as discussed in detail in Fader, Hardie, and Lee (2006a). The Pareto/NBD Model The Pareto/NBD model, developed by Schmittlein, Morrison, and Colombo (1987), uses observed past purchase behavior of customers to forecast their likely future purchase behavior. This outcome is then used as input to estimate CLV (e.g., Reinartz and Kumar 2000 and 2003; Fader, Hardie, and Lee 2005a; Singh, Borle, and Jain 2008). When customer defections are not observed, the Pareto/NBD model is a very elegant way of getting the probability of a customer’s activity in a relationship with a firm. This model can be used to estimate (a) the number of customers currently active, (b) how this number changes over time, (c) the likelihood of a specific customer’s being active, (e) how long a customer is likely to remain active, and (f) the expected number of purchases from a customer during a future time interval. The model assumptions are as follows: 1. While the customer is active in a business relationship with the firm, he/she makes transactions with the firm that are randomly distributed in time with customer–specific rate λ. Therefore, the number of transactions X made by a customer in a time period of length t is a Poisson random variable. 2. A customer has an unobserved lifetime duration represented by τ, which is an exponential random variable. Customers drop out or defect from the firm randomly according to a rate µ. 3. The transaction rate λ follows a Gamma distribution across customers with parameters r, α > 0. The mean purchase rate across customers is E[λ] = r/α and the variance is r/α2. 4. The dropout rate µ follows a Gamma distribution across customers with parameters s/β > 0. The mean dropout rate is E[µ] = s/β and the variance is s/β2. 5. The purchase rate λ and dropout rate µ vary independently across customers. For a customer, the expected number of transactions made in T units of time following an initial purchase is ( > ; U α V β 7 @ =
V − Uβ β − α (V − ) β + 7
(2)
.
For α > β, the probability that a customer is still active, given an observed purchase history of X purchases in time (0,T) since the initial purchase with the latest purchase happening at time t, is
α + 7 U + [ β + 7 V ) (D E F ] (W )) V α + W α + W 3($OLYH U V α β ; = [ W 7 ) = + V U + [ + V β +7 − ) (D E F ] (7 )) + 7 α
−
, (3)
MEASURING CUSTOMER LIFETIME VALUE
49
where a1 = r + x + s; b1 = s + 1; c1 = r + x + s + 1; z1(y) = (α-β)/(α+y). For cases when α = β and α < β, see Schmittlein, Morrison, and Colombo (1987). Extended Pareto/NBD Model. Schmittlein and Petersen (1994) empirically validate the Pareto/NBD model and propose an extension to it by incorporating dollar volume of purchase. The key assumption here is that the distribution of average amount spent across customers is independent of the distribution of the transaction rate λ and the dropout rate µ. The expected future dollar volume per reorder is provided in Schmittlein and Peterson. Note that for the estimation of CLV, the output of the Pareto/ NBD model is used along with the output from a suitable model for customer spending. Despite the significant merits of the Pareto/NBD model, it imposes strong assumptions on the underlying customer purchase behavior that might not be suitable for many situations (Borle, Singh, and Jain 2008; Singh, Borle, and Jain 2008). In addition, it is somewhat difficult to estimate as recommended, thus restricting its usage in practice. Several researchers have attempted to generalize the Pareto/NBD model and/or suggest alternatives that are easier to estimate. For example, Abe (2008) extends the Pareto/NBD model using a hierarchical Bayesian (HB) framework by relaxing the independence assumption of the purchase and dropout rates. The BG/NBD Model Proposed by Fader, Hardie, and Lee (2005b) as an “easy to estimate” alternative to the Pareto/ NBD model, the Beta-Geometric-NBD (BG/NBD) model makes slightly different assumptions about the underlying customer purchase behavior that make it significantly easier to implement and yet allow the user to obtain similar benefits. If ease of estimation is desired, the BG/NBD model is an excellent alternative to the well-known Pareto/NBD model. The assumptions underlying the BG/NBD model are as follows: 1. While active, the number of transactions made by a customer follows a Poisson process with transaction rate λ. Therefore, the time between transactions follows exponential distribution. 2. λ follows Gamma distribution across customers with parameters r and α. 3. After any transaction, a customer defects (becomes inactive) with probability p. Therefore the customer dropout is distributed across transactions as a (shifted) geometric distribution. 4. p follows a beta distribution with parameters a and b. 5. λ and p vary independently across customers. Note that assumptions (1) and (2) are identical to the Pareto/NBD model. While the Pareto/ NBD model assumes that customers can defect at any time, the BG/NBD model assumes that defections occur immediately after a purchase. Using these assumptions, the expression for the expected number of transactions in a future time period of length t for an individual with observed past purchase behavior is obtained as
( (< (W ) ; = [ W [ 7 U α D E ) =
D + E + [ − α + 7 − D − α + 7 + W
W ) U + [ E + [ D + E + [ − α 7 W + + , (4) U+[ α +7 D E + [ − α + W [
U+[
+ δ [ >
50
Siddharth S. Singh and Dipak C. Jain ) (D E F ] ) =
W E− ( − W )
( − ]W )−DGW F > E
where is the Euler’s integral for the Gaussian %(E F − E ) ∫ hypergeometric function, x is the number of transactions observed in time period (0,T), and tx (0 < tx ≤ T) is the time of the last transaction. See Fader, Hardie, and Lee (2005b) for more details of the BG/NBD model. The article also compares the forecast of future purchasing using the BG/NBD model with that from the Pareto/NBD model and finds that both the models are accurate (at both the individual customer level and the aggregate level). Just as in the case of Pareto/NBD model, the outcome from the BG/NBD model in combination with the outcome from a suitable customer spending model can be used to estimate CLV of each customer. F −E −
MCMC-Based Data Augmentation Algorithm by Singh, Borle, and Jain (2008) The Pareto/NBD and the BG/NBD models discussed so far in this section have several advantages over the earlier-used RFM scoring methods. The key advantages accrue as a result of the behavioral story underlying these models, where the observed transactions are considered a manifestation of underlying latent customer characteristics (see Fader, Hardie, and Lee 2006a). The problem is that the “story” underlying the modeling approach so far has been limited to strict assumptions imposed on customer characteristics such as the purchase behavior. For example, the Pareto/NBD model assumes that individual customer lifetimes and interpurchase times follow different exponential distributions. This is a restrictive assumption that might not be supported in many situations. In addition, the model assumes that the outcomes of lifetime, interpurchase time, and spending are independent of each other. Since these outcomes belong to each customer, the assumption of independence between these outcomes is also restrictive. Finally, the estimation of these models (except the BG/NBD model) is challenging. The Markov Chain Monte Carlo (MCMC)-based data augmentation framework proposed by Singh, Borle, and Jain (2008) (henceforth referred to as the SBJ framework for simplicity) addresses these issues successfully. The SBJ framework has at its core a data augmentation algorithm for estimating CLV models in noncontractual contexts and is not a specific model. Therefore it is fundamentally different from the models proposed so far in the literature. The SBJ framework can be used to estimate all the models discussed so far in addition to many different models with varying distributional assumptions for the underlying customer behavior. In fact, any standard statistical distribution can be used for the underlying customer characteristics generating the data. Therefore the user is not restricted to assuming that the interpurchase times and lifetimes follow exponential distributions. The algorithm then allows the forecast of future purchase transactions and the CLV using these. Both the Pareto/NBD and the BG/NBD models assume that the underlying drivers of CLV (customer lifetime and interpurchase time) are independent. Also, when these models are used for estimating CLV by adding a model for customer spending, spending also is assumed to be independent of these other behaviors. In reality, this independence is very hard to justify, as these outcomes are for the same customer. The SBJ framework allows for the estimation of correlation between the defection, purchase, and spending outcomes. See Singh, Borle, and Jain (2008) for applications of the framework including the estimation of the extended Pareto/NBD model, the BG/NBD model, and a number of other models that use different underlying distributions for customer lifetime and purchase behavior. The authors find that the assumptions of more flexible distributions for the underlying customer behavior yield more accurate forecasts. Some models compared yield significantly better performance than the Pareto/NBD and the BG/NBD, underscoring the value of the framework. It is noteworthy that such flexibility in modeling of CLV in noncontractual contexts was not available to the users earlier. The advantages of the SBJ frame-
MEASURING CUSTOMER LIFETIME VALUE
51
Table 2.4
Notation Used in the Framework (using Pareto/NBD assumptions) Notation
Explanation
h Th th
A particular customer. The total time for a customer h from the initial purchase occasion until the current time. The time from the initial purchase occasion until the last observed purchase occasion for customer h (therefore th ≤ Th). The total number of purchases made by customer h since the initial purchase in time period Th, with the latest purchase being at time th. The interpurchase time between the i th and the (i–1)th purchase for customer h. Total lifetime of customer h from the initial purchase occasion until the customer defects. This is a latent variable, that is, unobserved by the firm.
Xh IPThi totLIFEh totLIFEh(old) totLIFEh(new)
Intermediate variables used in the simulation of the latent variable totLIFEh.
work over existing alternatives are very significant, and we recommend its usage for forecasting customer purchase behavior in this context. We now describe the key algorithm in the SBJ framework using the example of the Pareto/NBD model. If we assume that customer purchase behavior follows the assumptions underlying the Pareto/NBD model and use the SBJ framework to estimate the Pareto/NBD model, the assumptions and the simulation/data augmentation steps involved are as shown in Tables 2.4 and 2.5. To estimate CLV, a model for customer spending is required as well. See Singh, Borle, and Jain (2008) for more details regarding estimating CLV using the framework, and modeling of covariates and correlations between the outcomes. The authors also show many other applications of the framework using different distributional assumptions for the drivers of CLV. Markov Chain Models Markov chain models have been used in marketing for years to model brand switching (e.g., Kalwani and Morrison 1977). Their flexibility and ability to model competition allows their use in modeling CLV as well. Suitable modifications of these models can be used both in a contractual and noncontractual context. An early use of this class of models for modeling CLV can be found in Pfeifer and Carraway (2000), who illustrate the relationship between Markov chain models and the commonly used recency, frequency, and monetary (RFM) framework. Rust, Lemon, and Zeithaml (2004) present a relatively sophisticated version of a Markov chain model to model CLV. Customer Equity Model of Rust, Lemon, and Zeithaml (2004) When data on customer purchase behavior from multiple vendors is available, this model is a very nice option for modeling the lifetime value of customers for specific brand/vendor. The flexibility of the model allows estimation of the impact of several drivers of customer lifetime value on the total lifetime value of a firm’s customers as well as that of its competitors. Another advantage of the model is that it can be estimated using cross-sectional survey data as well as longitudinal panel data.
52
Siddharth S. Singh and Dipak C. Jain
Table 2.5
Simulation/Data Augmentation Steps Involved in Estimating the Pareto/NBD Model Using the SBJ Algorithm Simulation Step
Description
Step 0 Initialization Initialize: totLIFEh = Th (or any number ≥ th). Initialize: totLIFEh(old) = totLIFEh(new) = totLIFEh. Initialize: The parameters (r,α) and (s,β). Step 1
Draw totLIFEh(new) (such that totLIFEh(new) ≥ th) from the Pareto distribution with the current parameters (s,β).
Step 2
Calculate two likelihoods for each customer—likeh(old) is the likelihood of the interpurchase time data for customer h conditional on the lifetime of h being totLIFEh(old), and likeh(new) is the likelihood of the interpurchase time data for customer h conditional on the lifetime of h being totLIFEh(new) drawn in the previous step.
Step 3
Assign totLIFEh = totLIFEh(new) with probability
SUREK =
OLNHK QHZ , or assign totLIFEh = totLIFEh(old) + OLNHK ROG OLNH QHZ K
Step 4
Estimate the Pareto distribution parameters (s,β) and the NBD distribution parameters (r,α) conditional on the total lifetime of the customer h being totLIFEh. That is, update NBD parameters (r,α) and Pareto parameters (s,β).
Step 5
Set totLIFEh(old) = totLIFEh.
Step 6
Go to Step 1.
Source: Singh, Borle, and Jain (2008).
Using individual-level data from a cross-sectional sample of customers and combining it with purchase or purchase intentions data, each customer’s switching matrix is modeled. The individual choice probabilities in the switching matrix, that is, the probability that individual i chooses brand k given that brand j was most recently chosen, is modeled using a multinomial logit model. Therefore, each customer i has a JXJ switching matrix, where J is the number of brands. The lifetime value of customer i to brand j is then calculated using these probabilities, the average purchase rate per unit time, the average purchase volume for brand j, and the per unit contribution margin for brand j. The other details can be found in Rust, Lemon, and Zeithaml (2004). Key Issues in Modeling CLV The literature on CLV models has exploded in the past few years. So far, the models proposed to estimate CLV have generally considered the revenue stream from customers and some obvious costs involved. Underlying this revenue stream are many complex and important factors that have generally not been considered in depth while modeling CLV. In addition, there are other factors related to the customer that are still not well understood and that can benefit the firm (such as network effects).
MEASURING CUSTOMER LIFETIME VALUE
53
Consider revenues from a customer. These revenues might come from purchases in multiple categories, and customer purchase behavior with respect to each category might vary. They might come in response to promotions with varying costs and effectiveness. Examples of such factors that can impact CLV are cross-selling, word-of-mouth effect, returns, and marketing actions and their impact on the revenues from different sources. To estimate CLV, one could potentially forecast the revenues directly using some sort of model for total revenues or one could model the revenues due to each underlying factor, such as purchases in each category, and build up the forecast for the total revenues. Some other areas in marketing have focused on many of these underlying customer-level factors to understand and forecast them better (e.g., Hess and Mayhew 1997). However, within the CLV measurement literature, models proposed so far have generally explicitly or implicitly made assumptions about many of these factors to deal with them. These assumptions take various forms, such as a factor having no impact (that is, when it is ignored) or a factor having the average value for all customers (e.g., Kumar, Ramani, and Bohling 2004). To take the CLV literature forward, bridges have to be built between the research on the drivers of CLV (and the factors that impact these drivers) and the literature on CLV measurement models such that the CLV measurement models can be refined further to provide better understanding of the effect of various factors on CLV and to forecast CLV more accurately. We now discuss some of the key issues that need to be considered in models of CLV to advance the CLV measurement literature. Cost of Customer Acquisition In analyzing the return on investment in customer relationships, how much a firm spends on acquiring a new customer is an important consideration. Acquiring customers at a price more than their lifetime value (that is, acquiring customers with negative PLV) would obviously result in a loss. It is common to see the popular press talk about the average cost of acquiring a customer in a specific industry (e.g., Schmid 2001); however, such average figures could be misleading. The experience of many direct marketing companies suggests that while some customers are quick to start business with the firm, others are reluctant and need many solicitations before they do. Clearly, then, the cost of acquiring different customers can be different. Once acquired, how does the purchase behavior (and subsequently the lifetime value) of these customers differ? Research suggests that the value of customers acquired through different channels can vary (Villanueva, Yoo, and Hanssens 2008; Lewis 2006). For example, Villanueva, Yoo, and Hanssens, using Web hosting company data, find that customers acquired through costly but fast-acting marketing investments add more short-term value, and customers acquired via word of mouth add nearly twice as much long-term value to the firm. Some other studies have linked acquisition and retention (e.g., Blattberg and Deighton 1996; Thomas 2001; Reinartz, Thomas, and Kumar 2005); however, to our knowledge no one has considered the total cost of individual customer acquisition (for example, cost attributable to the channel of acquisition and promotions for acquisition) with the subsequent customer purchase behavior and CLV. The heterogeneity in the cost of acquiring different customers needs more research and should be accounted for in evaluating prospects. Cost of Customer Relationship Management Once a prospect has been acquired as a customer, firms spend resources for managing their relationship with the customer and selling to him/her. These resources involve costs that can be complex and hard to pinpoint. For example, while the cost of manufacturing a product or the costs
54
Siddharth S. Singh and Dipak C. Jain
associated with promotions are relatively easy to obtain, the cost of employee time on an activity or the cost of managing returns is difficult to estimate. Since information about costs is generally not available to a researcher easily, most models of CLV make assumptions about these costs. For example, Venkatesan and Kumar (2004), while discussing cost, say that “the discounting of cost allocations is straightforward if we assume that there is a yearly allocation of resources (as is the case in most organizations) and that the cost allocation occurs at the beginning of the year (the present period).” The issue is that many organizations have aggregate cost figures, and allocating them in any model for individual customers without considering the variation in costs across them is likely to provide biased CLV estimates. Research suggests that such average cost allocation might not be appropriate, as the cost of serving a customer might be different for different customers (e.g., Reichheld and Teal 1996). Some studies focus on net revenues from customers and ignore the cost part altogether due to lack of availability of cost data. For example, Borle, Singh, and Jain (2008) propose a sophisticated model for measuring CLV in a contractual context. Due to unavailability of cost data, they use spending by customers to apply their model. Although their approach can be modified easily to include cost data, application of their model without such cost data can reduce the accuracy of the model forecasts. The cost of managing customer relationships and selling to them has several different sources. Some of the key sources that need consideration include: Cost of Customer Retention Firms take various initiatives such as loyalty programs to retain customers. The cost of such programs has yet to be adequately considered in models for CLV. This cost includes the cost of designing the program, launching it, and managing it in an ongoing fashion. While the first two components are part of the fixed cost of such programs and can be ignored, the cost of managing the program could be variable and should be considered if appropriate in evaluating customers. Within these cost components is the cost of physical and employee resources expended. Given that CLV is a customer-level measure, how does one allocate these costs to an individual customer? Should it be allocated equally, or should some other factor such as the frequency of product usage, the time spent with a customer service representative, or a combination of these form the basis of cost allocation? To our knowledge, these questions have not been considered in the literature. Cost of Marketing Activities Marketing activities are commonly used as explanatory variables in models for measuring CLV. The cost of such activities, however, is generally ignored. This cost might vary across customers, and a model has to account for this variation to accurately forecast CLV (Kumar, Shah, and Venkatesan 2006). For example, while some customers would be sent many promotions before they make a purchase, others might require just one. Cost of Returns Returns by customers has generally been ignored in developing CLV models. Given its importance, we discuss it in greater detail here. Return of purchased product by a customer is an undesirable activity for both the customer and the firm. Customers return an estimated $100 billion worth of
MEASURING CUSTOMER LIFETIME VALUE
55
products each year. The reasons for this trend include the rise of electronic retailing, the increase in catalog purchases, and a lower tolerance among buyers for imperfection (Stock, Speh, and Shear, 2002). Hess and Mayhew (1997) suggest that in direct marketing, the fact that customers do not physically evaluate a product before purchasing increases the risk of returns. Further, direct marketers should expect 4–25 percent of their sales returned. It is clear that returns are important, but this importance is not reflected in the research output related to product returns. In the CLV literature, some researchers have used returns as an explanatory variable in a model for a specific driver of CLV (e.g., Venkatesan and Kumar 2004; Reinartz and Kumar 2003), however, returns have not been incorporated in CLV measurement models directly despite being one of the key drivers of CLV. Demonstrating the value of returns for measuring CLV, Singh and Jain (2008a) use data from a direct marketing company to segment customers on the basis of purchases and returns. They then estimate segment level CLV first while ignoring returns and then while accounting for returns by individual customers. They find that consideration of returns significantly alters CLV. One segment of customers that is very attractive when returns are ignored becomes highly unattractive when returns are accounted for. The impact of returns on CLV varies significantly across segments, underscoring the need for considering returns at the individual customer level in modeling CLV. Cross-selling Many businesses such as retailers (for example, Amazon.com), consumer products companies (for example, Johnson & Johnson), banks (such as Citibank), and insurance companies (Allstate, for example) attempt to sell different types of products and services to their existing customers. Consider a woman who purchases a digital camera from Amazon.com. As soon as she purchases the camera, she gets personalized promotions of accessories such as a camera bag from Amazon. com. These cross-selling opportunities to a customer are an important reason for firms to invest in customer retention. If a company acquires a customer who buys from one division, the firm has information about the customer and her purchase behavior that can be used to its advantage by more effectively selling other products/services to her. Further, cross-selling enhances customer retention due to higher switching costs for customers who purchase multiple products and services from a firm (Kumar, George, and Pancras 2008). This implies that cross-selling to customers is likely to increase revenues and reduce the cost of customer retention. Therefore, for the firm, part of the customer value lies in the potential for cross-selling. The importance of cross-selling to customer value is well recognized by researchers (Kumar, George, and Pancras 2008; Verhoef and Donkers 2005; Ngobo 2004; Venkatesan and Kumar 2004; Reinartz and Kumar 2003; Blattberg, Getz, and Thomas 2001; Verhoef, Philip, and Hoekstra 2001; Kamakura, Ramaswami, and Srivastava 1991). However, how to measure and incorporate the cross-selling potential of a customer (which is likely to vary across customers) into CLV models remains a research question. Word-of-Mouth Effect Word of mouth represents informal communication between customers/consumers about a firm and/or its product and services (Tax, Chandrashekaran, and Christiansen 1993). Such communication can be both positive and negative. The value of customer word of mouth for a firm is well recognized (Zeithaml 2000; Harrison-Walker 2001; Helm 2003; Rogers 1995; Danaher and Rust 1996; Herr, Kardes, and Kim 1991; Walker 1995; Murray 1991; Buttle 1998). The importance
56
Siddharth S. Singh and Dipak C. Jain
of customer referrals for new customer acquisition is also recognized by researchers (Mangold, Miller, and Brockway 1999; Rogers 1995). The popularity of referral reward programs and viral marketing programs in practice reflects the value that firms place on word-of-mouth effects. While some researchers have attempted to estimate word-of-mouth effects, the research is still in its early stages due to the complexity involved (e.g., Helm, 2003; Hogan, Lemon, and Libai 2004 ). Kumar, Petersen, and Leone (2007) use data from a telecommunications company and financial services firm to estimate the value of customer referrals and conclude that a firm’s best marketers may be worth far more to the firm than the most enthusiastic customers, thus highlighting the importance of word of mouth in evaluating customers. Despite the importance of word-of-mouth effects, it is only recently that these and other network effects (such as participation in customer communities) have started getting the attention of researchers in CLV (e.g., Lee, Lee, and Feick 2006; Kumar, Petersen, and Leone 2007; Hogan, Lemon, and Libai 2004). It is obvious that these factors have an impact on customer value. What is not obvious is how to incorporate these factors in the model for measuring CLV, given the significant challenges in measuring them—related to the availability of appropriate data and the model for estimating them. So far, to our knowledge, these effects have not been incorporated adequately in the mainstream models for CLV. But this is an obvious direction for the literature to develop. Competition Most of the models proposed for measuring CLV use a firm’s database of past transactions with its customers for estimation. Since information on customer purchases from competitors is not available in such databases, competition has been ignored in modeling CLV. If CLV is to be used as a metric for making strategic decisions, then a firm needs to understand the impact of its actions on both its own CLV and that of its competitors. Lemon (2006) suggests that for CLV models to be really useful, they need to incorporate competition through share of wallet or the probability of purchasing from the competition in the future. Although desirable, competition cannot be incorporated into CLV models until relevant data is available. What should a firm do in the absence of such data? Should it not use the CLV metric? It is important to remember that CLV models are just a tool toward an end. CLV measures provide just one input into decision making by managers. Just as in conducting traditional marketing research we find the most useful tool in a given situation, considering given the time and other resource constraints, we have to consider CLV models in light of the availability of data and other constraints. If data is available, then incorporation of such information into CLV models is likely to provide very useful insights. A good example is Rust, Lemon, and Zeithaml (2004) discussed earlier. Forecasting and Planning Jain and Singh (2002) present the research on CLV as focusing on three main areas: (1) development of models to estimate CLV, (2) customer base analysis, and (3) analysis of CLV and its implications using analytical models. If we focus on (1) and (2), we can see that firms might require CLV models for at least two objectives. The first objective concerns understanding the implications of different marketing actions on CLV. The second concerns estimation of the lifetime value of each existing customer and prospect to take some action based on the estimated CLV (such as targeting high-value customers). In the former case, the firm has to incorporate marketing activities as covariates in the model for CLV. Most models allow this to be done in some manner. Data on past customer transactions can be used to estimate these
MEASURING CUSTOMER LIFETIME VALUE
57
models to understand the impact of these factors on CLV. Many research articles demonstrate such applications (e.g., Dreze and Bonfrer 2008). The second objective of forecasting CLV, however, raises some interesting questions. Malthouse and Blattberg (2005) examine how accurately future profitability of customers can be estimated. They use four data sets from different industries to study this issue. Based on the results, they propose two rules: of the highest 20 percent lifetime value customers, 55 percent will be misclassified, and of the bottom 80 percent lifetime value customers, 15 percent will be misclassified. As this study suggests, CLV forecasting is a risky business, as there is likely to be high error in the forecasts. Does this mean that it is not useful? We do not think so. CLV is a forward-looking dynamic concept, since it includes the value of the entire duration of the customer-firm relationship to the firm. Also, we have seen so far that it depends upon the firm’s marketing activities, its cost of maintaining the relationship with the customer, and the customer’s purchase and return behavior. Therefore, in order to accurately forecast CLV, one has to be able to forecast all the factors that drive CLV accurately in order to include these factors in a CLV forecasting model. This is an extremely difficult challenge. For example, Kumar, Shah, and Venkatesan (2006) assume the direct marketing cost for each customer to be the same for the next three years. The errors in the forecast of each driver of CLV are likely to add to the forecast error in the CLV. An alternative in models for estimating CLV is to ignore the factors affecting CLV that cannot be forecast accurately and consider only those that can be. Another alternative is to consider just the past purchase behavior and other time invariant factors (such as gender) to estimate CLV. These considerations will, however, affect the accuracy of CLV estimates obtained. All these issues raise questions about the use of CLV as a forward-looking metric. In order to decide which alternative is better, one has to consider the main objective of using CLV metric in the first place along with the constraints a firm faces. Adding complexity to a model properly is likely to improve estimates. The key question is whether adding this complexity, if possible, is worth the improvement in results obtained. Fader, Hardie, and Lee (2006b) contend that covariates and competition should be ignored in computing CLV because adding them only adds noise and complexity. Thus they advocate estimating CLV by considering only the past customer purchases and some time invariant covariates. What one finally chooses to do will depend upon the judgment of the user, as opinions differ on this issue. Here we want to point out that CLV is important for its use as a metric. This metric is useful only if it can be measured reasonably well and its estimation can be done with the data that firms generally possess or can collect. The literature on CLV measurement so far indicates that the models proposed, such as the Pareto/NBD model, provide an accurate enough forecast of CLV to be useful in practice. Applications of these models now available in the literature talk about how these models help a firm achieve better results. Clearly, these applications are proof of their utility. Fader, Hardie, and Lee (2006a and b) provide a strong defense of the use of probability models such as the Pareto/NBD and the BG/NBD model based on their performance in rigorous tests in many studies. Given the strong evidence that these models are useful to firms, we recommend using a suitable model for estimating CLV even if data on competition is not available and a forecast of key drivers of CLV cannot be done accurately. Endogeneity of CLV Drivers The modeling issues related to forecasting the value of customers have been discussed in the previous section. Here we focus on the modeling issues that relate to the analysis of CLV and the study of various factors on CLV.
58
Siddharth S. Singh and Dipak C. Jain
In the extant literature, researchers have segmented customers on the basis of customer lifetimes to analyze segment-level CLV while treating customer lifetime as exogenous (Reinartz and Kumar 2000). Recent research suggests that customer lifetime changes with each customer-firm interaction and therefore should not be treated as exogenous in analyzing lifetime value (Borle, Singh, and Jain 2008; Singh and Jain 2008b). It is common to build up a model for CLV using the key drivers of CLV (e.g., Venkatesan and Kumar 2004; Fader, Hardie, and Lee 2005a and b). Since these drivers, such as customer spending, interpurchase time, and customer lifetime, relate to each customer, they should be modeled jointly to account for the relationship between them (Seetharaman 2006; Borle, Singh, and Jain 2008). Also, values of these drivers that appear as explanatory variables are likely to be endogenous. The modeling should account for this potential endogeneity. However, these issues are either ignored or not adequately addressed. Not accounting for the endogeneity and simultaneity of various factors in the model is difficult to justify and can lead to misleading conclusions. While these issues are commonly taken care of in other research streams within marketing, the CLV literature is only now moving in the direction of addressing them (e.g., Singh and Jain 2008; Borle, Singh, and Jain 2008). Summary and Discussion The literature on customer lifetime value (CLV) has reached a junction where numerous models have been proposed to estimate CLV in different contexts. There is, however, little research on presenting this literature in a proper contextual framework, which is necessary to provide a user with easy recommendations for the models to use in any particular situation. Besides the issue of the context, there are key factors that impact CLV that have not been considered adequately, if at all, in the models for measuring and analyzing CLV. This chapter reviews the literature on models/methods proposed to measure CLV to do the following: 1. Discuss the relevance of the context of CLV measurement and propose a contextual framework that can be used to understand and classify CLV models better. 2. Describe some prominent models/methods proposed for measuring CLV in different contexts and discuss the strengths and weaknesses of each. 3. Discuss factors that impact CLV but have not been adequately considered in modeling CLV. These factors include network effects (such as word-of-mouth effects), returns, cost of managing customer relationships, and so forth. In short, this paper takes a pause to review what has been done in the CLV measurement literature so far and considers what needs to be done now to advance the literature. In addition, it attempts to sort out some contextual issues to present the literature in a proper contextual framework. The literature on CLV is expanding fast. We hope that this work will help readers understand it better and provide them with directions for future research. Notes 1. The numerous articles related to customer lifetime value address many issues within the field such as measurement of customer lifetime value (e.g., Borle, Singh, and Jain 2008; Lewis 2005; Fader, Hardie, and Lee 2005a; Gupta, Lehmann, and Stuart 2004; Libai, Narayandas, and Humby 2002; Reinartz and Kumar
MEASURING CUSTOMER LIFETIME VALUE
59
2000); study of the drivers of customer value (e.g., Rust, Lemon, and Zeithaml 2004; Magi 2003; Berger, Bolton, Bowman, and Briggs 2002); customer loyalty programs (e.g., Shugan 2005; Lewis 2004; Kim, Shi, Srinivasan 2001; Cigliano, Georgiadis, Pleasance, and Whally 2000; Dowling and Uncles 1997); and customer acquisition and retention (e.g., Capraro, Broniarczyk, and Srivastava 2003; Thomas 2001; Blattberg, Getz, and Thomas 2001). 2. Some researchers have also used the total value of a firm’s customers in measuring the value of the firm (Gupta, Lehmann, and Stuart 2004). 3. In the child care context, customer retention has to be managed until the child “grows up.” 4. Note that in Fader, Hardie, and Lee (2005b), the definition of a noncontractual context would be the same as a noncontractual–continuous-time context as defined here. 5. This is not to suggest that the timings of product usage do not affect CLV. Cost of managing customer relationships, returns, and so forth might occur at times different from the spending by a customer. These costs, if significant and when considered, would lead to further refinement of the categorization. Our purpose is to propose a categorization that achieves a balance between the most important criteria and parsimony of the framework.
References Abe, Makoto. 2008. “‘Counting Your Customers’ One by One: A Hierachical Bayes Extension to the Pareto/ NBD Model.” Marketing Science: Articles in Advance, published online before print June 23, 2008. Allenby, Greg M., Robert P. Leone, and Lichung Jen. 1999. “A Dynamic Model of Purchase Timing with Application to Direct Marketing.” Journal of American Statistical Association 94 (June), 365–374. Berger, P., R. Bolton, D. Bowman, and E. Briggs. 2002. “Marketing Actions and the Value of Customer Assets: A Framework for Customer Asset Management.” Journal of Service Research 5 (1), 39–54. Berger, Paul D., and Nada I. Nasr. 1998. “Customer Lifetime Value: Marketing Models and Applications.” Journal of Interactive Marketing 12 (1) (Winter), 17–30. Bhattacharya, C.B. 1998. “When Customers Are Members: Customer Retention in Paid Membership Contexts.” Journal of the Academy of Marketing Science 26 (1), 31–44. Bhattacharya, C.B., Hayagreeva Rao, and Mary Ann Glynn. 1995. “Understanding the Bond of Identification: An Investigation of Its Correlates Among Art Museum Members.” Journal of Marketing 59 (October), 46–57. Blattberg, Robert C., and John Deighton. 1996. “Manage Marketing by the Customer Equity Test.” Harvard Business Review 74 (4) (July/August), 136–145. Blattberg, Robert, Gary Getz, and Jacquelyn S. Thomas. 2001. Customer Equity: Building and Managing Relationship as Assets. Boston: Harvard Business School Press. Bolton, Ruth N. 1998. “A Dynamic Model of the Duration of the Customer’s Relationship with a Continuous Service Provider: The Role of Satisfaction.” Marketing Science 17 (1), 45–65. Borle, Sharad, Siddharth S. Singh, and Dipak C. Jain. 2008. “Customer Lifetime Value Measurement.” Management Science, 54 (1) (January), 100–112. Buttle, F.A. 1998. “Word-of-Mouth: Understanding and Managing Referral Marketing.” Journal of Strategic Marketing, 6 (3), 241–254. Capraro, Anthony J., Susan Broniarczyk, and Rajendra K. Srivastava. 2003. “Factors Influencing the Likelihood of Customer Defection: The Role of Consumer Knowledge.” Journal of the Academy of Marketing Science 31 (2), 164–175. Chintagunta, Pradeep K. 1993. “Investigating Purchase Incidence, Brand Choice and Purchase Quantity Decisions of Households.” Marketing Science 12 (2) (Spring), 184–208. Cigliano, J., M. Georgiadis, D. Pleasance, and S. Whally. 2000. “The Price of Loyalty.” McKinsey Quarterly no. 4, 69–77. Colombo, Richard, and Weina Jiang. 1999. “A Stochastic RFM Model.” Journal of Interactive Marketing 13 (3) (Summer), 2–12. Danaher, P., and Roland T. Rust. 1996. “Indirect Financial Benefits from Service Quality.” Quality Management Journal 3 (2), 63–75. Donkers, Bas, Peter C. Verhoef, and Martijn G. de Jong. 2007. “Modeling CLV: A Test of Competing Models in the Insurance Industry.” Quantitative Marketing and Economics 5, 163–190. Dowling, Grahame R., and Mark Uncles. 1997. “Do Customer Loyalty Programs Really Work?” Sloan Management Review 38 (4) (Summer), 71–82.
60
Siddharth S. Singh and Dipak C. Jain
Dreze, Xavier, and Andre Bonfrer. 2008. “An Empirical Investigation of the Impact of Communication Timing on Customer Equity.” Journal of Interactive Marketing 22 (1) Winter, 36–50. Dwyer, F. Robert. 1997. “Customer Lifetime Valuation to Support Marketing Decision Making.” Journal of Direct Marketing 11 (4), 6–13. Fader, Peter S., Bruce G.S. Hardie, and Ka Lok Lee. 2005a. “RFM and CLV: Using Iso-Value Curves for Customer Base Analysis.” Journal of Marketing Research (November), 415–430. ———. 2005b. “‘Counting Your Customers’ the Easy Way: An Alternative to the Pareto/NBD Model.” Marketing Science 24 (2) (Spring), 275–284. ———. 2006a. “More than Meets the Eye.” Marketing Research (Summer), 9–14. ———. 2006b. “The Authors Respond.” Marketing Research, Back Talk, 18 (3) (Fall), 56. Gupta, Sunil, Donald R. Lehmann, and Jennifer Ames Stuart. 2004. “Valuing Customers.” Journal of Marketing Research 41 (1) (February), 7. Gupta, Sunil, and Valarie Zeithaml. 2006. “Customer Metrics and Their Impact on Financial Performance.” Marketing Science 25 (6) (November-December), 718–739. Harrison-Walker, L.J. 2001. “The Measurement of Word-of-Mouth Communication and an Investigation of Service Quality and Customer Commitment as Potential Antecedents.” Journal of Service Research 4 (1), 60–75. Helm, Sabrina. 2003. “Calculating the Value of Customers’ Referrals.” Managing Service Quality 13 (2), 124–133. Helsen, Kristiaan, and David C. Schmittlein. 1993. “Analyzing Duration Times in Marketing: Evidence for the Effectiveness of Hazard Rate Models.” Marketing Science 12 (4) (Autumn), 395–414. Herr, Paul M., Frank R. Kardes, and John Kim. 1991. “Effects of Word of Mouth and Product Attributes Information on Persuasion: An Accessibility–Diagnosticity Perspective.” Journal of Consumer Research 17 (4), 454–462. Hess, J.D., and G.E. Mayhew. 1997. “Modeling Merchandise Returns in Direct Marketing.” Journal of Direct Marketing 11 (2), 20–35. Hogan, John E., Katherine N. Lemon, and Barak Libai. 2004. “Quantifying the Ripple: Word-of-Mouth and Advertising Effectiveness.” Journal of Advertising Research (September), 271–280. Jackson, Barbara B. 1985. Winning and Keeping Industrial Customers. Lexington, MA: D.C. Heath. Jain, Dipak C., and Siddhartha S. Singh. 2002. “Customer Lifetime Value Research in Marketing: A Review and Future Directions.” Journal of Interactive Marketing 16 (2) (Spring), 34–46. Jain, Dipak C., and Naufel J. Vilcassim. 1991. “Investigating Household Purchase Timing Decisions: A Conditional Hazard Function Approach.” Marketing Science 10 (1) (Winter), 1–23. Kalwani, Manohar U., and Donald Morrison. 1977. “A Parsimonious Description of the Hendry System.” Management Science 23 (January), 467–477. Kamakura, Wagner A., Sridhar N. Ramaswami, and Rajendra K. Srivastava. 1991. “Applying Latent Trait Analysis in the Evaluation of Prospects for Cross-Selling of Financial Services.” International Journal of Research in Marketing 8, 329–349. Kim, Byung-Do, Mengze Shi, Kannan Srinivasan. 2001. “Reward Programs and Tacit Collusion.” Marketing Science 20 (2), 99–120. Kotler, P., and G. Armstrong. 1996. Principles of Marketing, 7th ed. Englewood Cliffs, NJ: PrenticeHall. Kumar, V., Morris George, and Joseph Pancras. 2008. “Cross-Buying in Retailing: Drivers and Consequences.” Journal of Retailing 84 (1), 15–27. Kumar, V., Girish Ramani, and Timothy Bohling. 2004. “Customer Lifetime Value Approaches and Best Practice Applications.” Journal of Interactive Marketing 18 (3) (Summer), 60–72. Kumar, V., J. Andrew Petersen, and Robert P. Leone. 2007. “How Valuable Is Word of Mouth?” Harvard Business Review 85 (10) (October), 139. Kumar, V., Denish Shah, and Rajkumar Venkatesan. 2006. “Managing Retailer Profitability—One Customer at a Time!” Journal of Retailing 82 (4), 277–294. Lee, Jonathan, Janghyuk Lee, and Lawrence Feick. 2006. “Incorporating Word-of-Mouth Effects in Estimating Customer Lifetime Value.” Database Marketing & Customer Strategy Management 14 (1) (October), 29–39. Lemon, Katherine. 2006. “Still Some Weaknesses.” Marketing Research, Back Talk, 18 (3) (Fall), 53. Lewis, Michael. 2004. “The Influence of Loyalty Programs and Short-Term Promotions on Customer Retention.” Journal of Marketing Research 41 (August), 281–292.
MEASURING CUSTOMER LIFETIME VALUE
61
———. 2005. “Incorporating Strategic Consumer Behavior into Customer Valuation.” Journal of Marketing 69 (October), 230–238. ———. 2006. “Customer Acquisition Promotions and Customer Asset Value.” Journal of Marketing Research 43 (May), 195–203. Libai, Barak, Das Narayandas, and Clive Humby. 2002. “Toward an Individual Customer Profitability Model: A Segment Based Approach.” Journal of Service Research 5 (1), 69–76. Magi, Anne W. 2003. “Share of Wallet in Retailing: The Effects of Customer Satisfaction, Loyalty Cards and Shopper Characteristics.” Journal of Retailing 79 (Summer), 97–106. Malthouse, Edward C., and Robert C. Blattberg. 2005. “Can We Predict Customer Lifetime Value?” Journal of Interactive Marketing 19 (1) (Winter), 2–16. Mangold, W.G., F. Miller, and G.R. Brockway. 1999. “Word-of-Mouth Communication in the Service Marketplace.” The Journal of Services Marketing 13 (1), 873–889. Murray, Keith B. 1991. “A Test of Services Marketing Theory: Consumer Information Acquisition Activities.” Journal of Marketing 55 (1), 10–25. Ngobo, Paul Valentin. 2004. “Drivers of Customers’ Cross-Buying Intentions.” European Journal of Marketing 38 (9/10), 1129–1157. Pfeifer, Phillip E., and Robert L. Carraway. 2000. “Modeling Customer Relationships as Markov Chains.” Journal of Interactive Marketing, 14 (2) (Spring). Reichheld, Frederick, and Thomas Teal. 1996. The Loyalty Effect. Boston, MA: Harvard Business School Press. Reinartz, Werner J., and V. Kumar. 2000. “On the Profitability of Long Lifetime Customers: An Empirical Investigation and Implications for Marketing.” Journal of Marketing 64, 17–35. ———. 2003. “The Impact of Customer Relationship Characteristics on Profitable Life Duration.” Journal of Marketing 67 (January), 77–99. Reinartz, Werner, Jacquelyn S. Thomas, and V. Kumar. 2005. “Balancing Acquisition and Retention Resources to Maximize Customer Profitability.” Journal of Marketing 69 (January), 63–79. Rogers, Everett M. 1995. The Diffusion of Innovations. New York: The Free Press. Rust, Roland T., Katherine N. Lemon, and Valarie A. Zeithaml. 2004. “Return on Marketing: Using Customer Equity to Focus Marketing Strategy.” Journal of Marketing 68 (January), 109–127. Schmid, Jack. 2001. “The Cost of Acquiring a Customer.” Catalog Age, February2001, vol 18, no. 2 p. 65. Schmittlein, David C., Donald G. Morrison, and Richard Colombo. 1987. “Counting Your Customers: Who Are They and What Will They Do Next?” Management Science 33 (January), 1–24. Schmittlein, David C., and Robert A. Peterson. 1994. “Customer Base Analysis: An Industrial Purchase Process Application.” Marketing Science 13 (1), 41–68. Seetharaman, Seethu. 2006. “Don’t Oversimplify.” Marketing Research, Back Talk, 18 (3), 53. Shugan, Steve M. 2005. “Editorial: Brand Loyalty Programs: Are They Shams?” Marketing Science 24 (2) (Spring), 185–193. Singh, Siddharth S., and Dipak C. Jain. 2008a. “An Empirical Analysis of Product Returns in Direct Marketing.” Working paper, Rice University. ———. 2008b. “The Economics of Customer Relationship Management: Insights from Customer Lifetime Purchase Behavior.” Working paper, Rice University. Singh, Siddharth S., Sharad Borle, and Dipak C. Jain. 2008. “A Generalized Framework for Estimating Customer Lifetime Value When Customer Lifetimes Are Not Observed.” Working paper, Rice University. Stock, James, Thomas Speh, and Herbert Shear. 2002. “Many Happy (Product) Returns.” Harvard Business Review 80 (7) (July), 16–17. Tax, S.S., M. Chandrashekaran, and T. Christiansen. 1993. “Word-of-Mouth in Consumer Decision-Making: An Agenda for Research.” Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior 6, 75–80. Thomas, Jacquelyn S. 2001. “A Methodology for Linking Customer Acquisition to Customer Retention.” Journal of Marketing Research 38 (May), 262–268. Venkatesan, Rajkumar, and V. Kumar. 2004. “A Customer Lifetime Value Framework for Customer Selection and Resource Allocation Strategy.” Journal of Marketing 68 (October), 106–125. Verhoef, Peter C., and Bas Donkers. 2005. “The Effect of Acquisition Channels on Customer Loyalty and Cross-Buying.” Journal of Interactive Marketing 19 (Spring), 31–43.
62
Siddharth S. Singh and Dipak C. Jain
Verhoef, Peter C., Hans Franses Philip, and Janny C. Hoekstra. 2001. “The Impact of Satisfaction and Payment Equity on Cross-Buying: A Dynamic Model for a Multi-service Provider.” Journal of Retailing 77, 359–378. Villanueva, Julian, Shijin Yoo, and Dominique M. Hanssens. 2008. “The Impact of Marketing-Induced Versus Word-of-Mouth Customer Acquisition on Customer Equity Growth.” Journal of Marketing Research 45 (February), 48–59. Walker, C. 1995. “Word-of-Mouth.” American Demographics 17 (7), 38–44. Zeithaml, V.A. 2000. “Service Quality, Profitability, and the Economic Worth of Customers: What We Know and What We Need to Learn.” Journal of the Academy of Marketing Science 28 (1), 67–85.
Chapter 3
LEARNING MODELS S. Sriram and Pradeep K. Chintagunta
Abstract Choice models typically assume that agents know with certainty the utility they would derive from various alternatives. Such an assumption is likely to be violated in instances where (a) the agent is new to the context or (b) the choice set has new alternatives. Learning models specify a mechanism by which consumers resolve uncertainty regarding products or their characteristics in such “turbulent” contexts. In this chapter, we provide a critical review of the extant learning literature in marketing and economics. We also discuss some avenues for future research in this area. Introduction The extensive literature on choice models has been built on the premise that consumers derive utility from the various alternatives in their choice set and choose the alternative that gives them the greatest utility. Typically, these models assume that consumers know the various components of this utility with certainty. This would be the case for categories in which (a) consumers make purchases regularly and (b) there are no new product introductions and consumers have been participating in the category over an extended period of time. However, if consumers are either new to the category or if the category experiences several new product introductions, this assumption is likely to be violated. In such instances, consumers would perceive some uncertainty in the utility they would derive from the various alternatives. Consequently, this would play a role in their choice decision. Moreover, consumers may try to resolve this uncertainty by learning about the utility they would derive from the products through various information sources such as actual purchase and consumption of the product, advertising messages, and word-of-mouth interactions with other consumers. Learning models specify the mechanism by which consumers use these various sources of information to resolve their uncertainty. A choice model that formally incorporates learning behavior would then be able to assess the relative efficacy of the different information sources in aiding consumer learning. The literature that considers consumer learning can be broadly classified into two streams. The first consists of reduced-form models that allow for the evolution of a consumer’s brand preferences and price sensitivity parameter to be a function of her past experience in the category (Heilman, Bowman, and Wright 2000) or due to her exposure to advertising by these brands (see, for example, Jedidi, Mela, and Gupta 1999; Sriram, Chintagunta, and Neelamegham 2006; Sriram and Kalwani 2007). Therefore, these models do not provide a structural representation of the learning mechanism. One criticism of such reduced form models is that they may not be invariant to policy changes (also known as Lucas Critique). Therefore, they may not be useful if the focus of the researcher is to 63
64
S. Sriram and Pradeep K. Chintagunta
understand the implications of significant policy changes. The second stream of literature explicitly accounts for how consumers update their beliefs regarding components of their utility about which they are uncertain. The structural nature of such models provides some defense against the Lucas Critique. Most of these studies assume that consumers update their beliefs in a Bayesian fashion with the extent of updating being related to their perceived precision of the signals that aid in such learning. In this chapter, we present a review of this latter stream of the learning literature. While the above discussion was based on consumers’ learning about the utility they would derive from the various alternatives, the literature encompasses various other contexts, such as physicians’ learning about a drug and managers’ learning about the attractiveness of a market. The rest of the chapter is organized as follows. First, we begin by discussing the basic structure of learning models. We then discuss the differences between the various learning models in the marketing and economics literatures in terms of the agent who is learning, the entity that the agent is uncertain about, and the signals that help in learning. Subsequently, we talk about how the basic learning model has been extended in the literature. The following section sheds some light on potential avenues for future research, and the final section provides some concluding comments. The Basic Structure of Learning Models Learning models are typically characterized by the following four aspects: (a) an agent, (b) a component of the agent’s utility function that she is uncertain about (an unknown entity), (c) signals that the agent receives about the unknown entity, and (d) a mechanism by which information in the signal is used to resolve the uncertainty in (b). A common example in marketing involves consumers (the agents) who are uncertain about the quality of the product they are deciding to purchase (the unknown entity). Upon purchase and experience of the product, they receive some information regarding its true quality (the signal). The signal is usually assumed to be “noisy,” that is, a single purchase or experience usually does not resolve all the uncertainty regarding the product. In a standard Bayesian learning model (the mechanism), the agent has some prior belief about the quality of the product. The (noisy) signal the consumer receives from the purchase and use of the product allows the agent to combine the prior belief with the signal in a Bayesian fashion to update beliefs about the product’s quality. Usually, researchers assume that the noisy information that agents receive each period comes from a distribution whose mean equals the true value of the unknown entity, that is, that the signals are unbiased. Hence, if the agent receives signals over an extended period of time by making repeated purchases, the updated belief about product quality will converge to its “true” value. Below, we present a formal discussion of the model structure in the context of this example. Consider a market with only one new nondurable product that can be purchased every period. Consumers in such a market decide on whether to purchase the product or not during each period t, t = 1, 2, . . . , T. Further, we assume that consumers can make only one purchase during each period. Let Q be the true quality of the product. In our illustrative learning model, consumers do not know this true quality. At period 0, all consumers start with a prior belief that the quality of this product is normally distributed with mean Q0 and variance s02, that is,
Prior ~ 1 4 σ .
(1)
In period 1, consumers would make their purchase decisions based on this prior belief. If consumer i, i = 1, 2, . . . I, purchases the product, she can assess the quality of the product from her consumption experience, QEi1. If we assume that the consumer always derives the experience
LEARNING MODELS
65
of quality that is equal to the true quality QEi1 Q, i, t, then this one consumption experience is sufficient to assess the true quality of the product. However, in reality, this experienced quality might differ from the true quality, Q, because of (a) intrinsic product variability (Roberts and Urban 1988) and/or (b) idiosyncratic consumer perceptions (Erdem and Keane 1996). Hence, researchers typically assume that these experienced quality signals are draws from a normal distribution whose mean equals the true quality, that is, that these are unbiased signals. Thus, we have
4(LW a 1 4 σ 4 , W = 7 L = , ,
(2)
where σ 4 captures the extent to which the signals are noisy. Thus, for learning to extend beyond the initial purchase, we need σ 4 > . Subsequent to the first purchase (and consumption experience) the consumer has some more information than the prior she started with. Consumers use this new information along with the prior to update their beliefs about the true quality of the product in a Bayesian fashion. Specifically, since both the prior and the signal are normally distributed, conjugacy implies that the posterior belief at the end of period 1 would also follow a normal distribution (DeGroot 1970) with mean 4L and variance σ L such that 4L = θ L4 + ϖ L4(L and
where
σ L =
+ σ σ 4
,
(3a) (3b)
σ 4 σ and θ L = = + σ + σ 4 σ σ4
(3c)
σ4 σ ϖ L = = . + σ +σ4 σ σ4
(3d)
This posterior belief at the end of period 1 acts as the prior belief at the beginning of period 2. Thus, when the consumer makes a purchase decision in period 2, she would expect her quality experience a to come from this distribution, that is, 4L a 1 4L σ L . On the other hand, a consumer who does not make a purchase in period 1 will use the same prior in period 2 as she did in period 1. Hence, we can generalize equations 3a, 3b, 3c, and 3d for any time period t, t = 1, 2, . . . , T, as follows:
4LW = θ4L W − + ϖ4(LW ,
σ LW =
σ LW−
+ , LW σ4
=
W , Lτ +∑ σ τ = σ 4
(3a´) (3b´)
66
S. Sriram and Pradeep K. Chintagunta
σ 4 σ L W − θ LW = = + , LW , LW σ L W − + σ 4 σ L W − σ4 , LW
ϖ LW =
σ L W −
σ 4 , + LW
= σ 4
, LW σ LW − , LW σ LW − + σ 4
,
(3c´)
(3d´)
where Iit is an indicator variable that takes on the value 1 if consumer i makes a purchase in period t and 0 otherwise. Similarly, when the consumer makes a purchase in period t+1, she would a assume that the quality of the product, 4LW +, comes from this posterior distribution at the end of a period t, that is, 4LW + a 1 4LW σ LW . Equations 3a´, 3b´, 3c´, and 3d´ imply that as the number of consumption experiences increases, the consumer learns more and more about the true quality of the product. As a result, her posterior mean would shift away from her initial prior and move closer to the true mean quality. Similarly, as she receives more information, her posterior variance would decrease. In order to demonstrate how a consumer’s posterior belief would evolve as she receives these signals, we performed a simulation wherein the true quality of the product, Q, is set at 5. The consumer has a prior belief that the true quality of the product, Q0, is 0 with a variance of 5 (σ = ). The consumer receives unbiased signals around this true quality with a signal variance of 2 (σ 4 = ). In Figure 3.1, we plot the evolution of the posterior mean and variance as the number of purchase occasions increase. As discussed above, the figure reveals that the consumer’s posterior belief about the true quality of the product converges to its true value as she receives more signals. Furthermore, her uncertainty about this belief (posterior variance) falls with each additional signal and tends to zero asymptotically. This concludes our discussion of a basic mechanism by which consumers learn about the quality of a new (to them) product. Later we will discuss other learning mechanisms. Now we turn to a discussion of the utility function that drives purchases. Specification of the Utility Function For the sake of simplicity, we define the utility that the consumer derives from the product at time t as a function of her quality perception and the price of the product at that time. As discussed above, when a consumer makes a purchase decision at period t, she still perceives some uncertainty about the quality of the product she would receive. Hence, her utility will also be a random variable. Specifically,
a XaLW = I 4LW SW β + ε LW ,
(4)
where XaLW is the utility that consumer i derives from purchasing the product at time t, pt is the price of the product at time t, β is the price sensitivity parameter, and εit is a consumer and time-
LEARNING MODELS
67
Figure 3.1 Change in Posterior Mean and Variance with Number of Purchases
6 5 4 3 2 1 0 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
No. Purchases (Signals) True Quality
Posterior Mean
Posterior Variance
varying idiosyncratic term that is not observed by the researcher.1 Since the consumer does not know the true quality and, hence, the true utility, she makes her purchase decision based on the expected utility,
a a (>XaLW @ = (> I 4LW SW β + ε LW @ = (> I 4LW SW β @ + ε LW ,
(5)
where E[.] is the expectation operator. The expectation is taken over the prior distribution at the beginning of that period (or the posterior distribution at the end of the previous period). Since the error term is perfectly known to the consumer and unknown only to the researcher, it can be taken out of the expectation operator as in equation 5. Further, if we assume that the deterministic component of the utility from not purchasing is 0 and the error term εit follows a type I extreme value distribution, we can write out the probability that consumer i would make a purchase at time t, Prit as
3ULW =
a H[S (> I 4LW SW β @ . a + H[S (> I 4LW SW β @
(6)
The specification of the utility function in equation 4 will have implications for how the posterior mean and variance enter the expected utility in equation 5 and hence the probability of purchase a in equation 6. For example, if f(.) is a linear function of 4LW such that
a XaLW = 4LW + β SW + ε LW ,
(7a)
a (>XaLW @ = (>4LW @ + β SW + ε LW
= 4L W − + β SW + ε LW
.
(7b)
68
S. Sriram and Pradeep K. Chintagunta
The expression for the expected utility in equation 7b implies that it depends solely on the posterior mean from the previous period and is independent of the posterior variance. On the other hand, concave or convex specifications of f(.) would lead to expected utilities that depend on both the posterior mean and variance. More specifically, a concave utility function would imply that the consumers are risk averse. Consequently, the expected utility would be negatively influenced by the posterior variance. Likewise, a convex utility function would have the posterior variance affecting the expected utility positively. In Table 3.1, we present the utility function and the expected utility for three commonly used functional forms in the literature— linear, quadratic, and constant absolute risk aversion (CARA). Note that the term γ in the quadratic and CARA specifications corresponds to the level of risk aversion, with positive values corresponding to risk-averse consumers (concave utility) and negative values corresponding to risk-seeking consumers (convex utility). Learning Models in Marketing and Economics We now turn our attention to a discussion of the different applications of learning models in the marketing and economics literature. Specifically, we break down the discussion into how these studies differ in terms of the following three questions: (a) who is learning? (the agent), (b) what are they learning about? (the uncertain entity), and (c) how do they learn? (the signal). In Appendix 3.1, on pages 82–83, we provide a summary of selected studies on learning. The appendix also provides details on how these studies differ on these three dimensions. Who Is Learning? (The Agent) Broadly, there are three types of agents in the literature: (a) consumers who are making decisions regarding their own consumption or purchases, (b) physicians making decisions on behalf of their patients, and (c) managers making decisions on behalf of their firms. As regards the first group, consumers, the literature on learning models spans several industries including consumer packaged goods (see, for example, Ackerberg 2003; Erdem and Keane 1996; Mehta, Rajiv, and Srinivasan 2003, 2004), consumer durables such as automobiles (Roberts and Urban 1988) and computers (Erdem, Keane, Oncu, and Strebel 2005), and services such as local telephone (Narayanan, Chintagunta, and Miravete 2007) and wireless (Xiao, Chan, and Narasimhan 2007). The choice of the category has typically been dictated by the nature of data required to infer consumer learning. Since inference of learning requires us to observe consumer decisions over several time periods, most of the work has been in the context of frequently purchased consumer packaged goods or services such as telephone and wireless where consumers are not bound by a contract and therefore have the option of switching between services during each period. A casual perusal of the appendix would confirm this. The only exceptions are the studies by Roberts and Urban (1988) and Erdem and colleagues (2005), set in the context of consumer purchases of automobiles and computers, respectively. Clearly, if one needs to infer learning in these contexts based on how consumers modify their purchases over time, it would require data that track the purchases of these consumers over several purchase occasions. Given the lifetime of these categories (especially automobiles), we may have to track purchases over several decades to arrive at a reasonably large purchase history to infer learning. Therefore, in both cases, inference regarding learning is not based on repeat purchases by consumers. Rather, they use data from consumer surveys to infer how consumers learn over time. For example, Roberts and Urban investigate how car buyers would learn about a new car model through word of mouth from current customers. In order to infer this, they collect
(>XaLW @ = 4L W − + β SW + ε LW (>XaLW @ = 4L W − − γ 4L W − − γσ LW − + β SW + ε LW (>XaLW @ = − H[S−γ 4L W − − γσ LW − + β SW + ε LW
a XaLW = 4LW + β SW + ε LW
a a XaLW = 4LW − γ4LW + β SW + ε LW
a XaLW = − H[S−γ 4LW + β SW + ε LW
Quadratic
CARA
Linear
Expected Utility
Utility Function
Functional Form
Expected Utility for Alternative Utility Functions
Table 3.1
69
70
S. Sriram and Pradeep K. Chintagunta
information from potential buyers of the model regarding their willingness to recommend the model to other potential buyers. The second class of agents includes physicians who make decisions regarding the drugs to prescribe to their patients. Past studies have investigated decisions by these agents in a broad class of prescription drugs, including those for allergies (Narayanan, Manchanda, and Chintagunta 2005), anti-ulcer medication (see, for example, Coscelli and Shum 2004; Crawford and Shum 2005; Chintagunta, Jiang, and Jin 2008), erectile dysfunction (Narayanan and Manchanda 2009), and heart disease (Ching 2008). Since the agent (physician) and the end consumer (the patient) are not the same, this context is vulnerable to the critique that the two players may have different objectives. However, past research has argued that the fear of malpractice lawsuits would ensure that the physicians have the same objectives as their patients. The third category of agents comprises managers making decisions on behalf of their firms. Typically, the decision revolves around whether they should continue with a product or a service in a given market. For example, Hitsch (2006) investigates the decision of a manager of a brand of cereals on whether to continue with a new product or discontinue it. Similarly, Dixit and Chintagunta (2007) consider the decision of discount airlines on whether to exit a city-pair market. Other researchers such as Balvers and Cosimano (1990) deal with the situation in which rational, forward-looking firms are uncertain about their true demand and use pricing strategically to generate more informative quantity observations. Yet another category of research along these lines is the adoption of an innovation of uncertain profitability by a firm. Here the firm receives signals on the profitability of the innovation, thereby learning about profitability from them; its decision problem is in the form of an optimal stopping problem in which the stopping value is the expected return from adoption, and the value from optimal continuation is the discounted expected value of the next piece of information (see Jensen 1982). Of the three categories of agents, strategic decision-making by managers appears to be the least studied in the literature. What Are They Learning About? (The Uncertain Entity) The most common entity in the literature that agents learn about is the overall quality of a product. These studies typically start with the assumption that, at the beginning of the data, the agent is uncertain about the product’s true quality. As the agent makes a series of decisions (such as purchase and subsequent consumption of the product), she obtains unbiased information regarding the true value of the unknown entity, which helps in updating her beliefs regarding that value, as discussed above in the section on the basic structure of learning models. In the basic structure, the true quality does not vary across agents (see, for example, Erdem and Keane 1996). This basic structure has been extended along two dimensions. First, researchers have extended the notion of an overall true quality to accommodate the fact that the product match might vary across agents (Crawford and Shum 2005). For example, when physicians prescribe drugs to patients, they are likely to consider the fact that the efficacy of a drug would vary across patients. Therefore, apart from drawing inferences about the overall quality of a drug, they would also learn about how a drug matches a particular patient (see, for example, Chintagunta, Jiang, and Jin 2008). Similarly, agents can have different preferences for consumption that could make some options more attractive than others. For example, Narayanan, Chintagunta, and Miravete (2007) consider the choice of fixed-rate or metered calling plans for local telephone service. Given uncertainty about their true consumption rate or type (due to being only on the fixed-rate plan till the introduction of the metered plan), the agents (consumers) have to choose a calling plan that would suit their needs best. As consumers learn about their true types, they would converge on a plan that is the best match
LEARNING MODELS
71
for them. Hence, this bears similarity to the case where physicians infer the patient-drug match based on patient feedback. In a variant of this extension, Coscelli and Shum (2004) investigate how physicians learn about the efficacy of a drug in treating different diagnoses. In addition to extending the basic structure to accommodate differences in match across agents, researchers have added more dimensions of learning. For example, Mehta, Rajiv, and Srinivasan (2003) extend the basic learning model of consumers’ learning about quality to also accommodate consumer price search behavior. Consumers are not fully informed about the price levels of the products in the category and need to undertake a costly search to uncover these prices. Since they are assumed to know the distribution of prices for each of the brands, their choice of the size of the consideration set represents a trade-off between the cost of searching for as many prices as there are brands in the set with the expected benefits from those products. Thus authors assume that the consumers follow a fixed sample size search strategy. At the same time, they also assume that learning about the two entities–quality and price—are unrelated. Therefore, information regarding the price of a product does not help in updating beliefs about the product’s quality. Recently, Erdem, Keane, and Sun (2008) have allowed for both price and advertising to influence consumers’ learning of quality. Similarly, Crawford and Shum (2005) consider two unknown entities in the context of prescription drugs: symptomatic and curative effects. While the symptomatic effect corresponds to the efficacy of the drug in relieving symptoms, curative effects relate to the ability of the drug to cure the patient. As in Mehta, Rajiv, and Srinivasan (2003), they assume that learning about the efficacy of the drug on one effect does not inform the patient or the physician about its efficacy on the other effect. Ackerberg (2003) extends the notion of learning about multiple entities by allowing their signals to be correlated. In his model set in the context of packaged goods, consumers are uncertain about (a) the quality and (b) advertising expenditures of different brands in the yogurt category. The uncertainty about advertising expenditures arises because consumers may not watch all the advertisements of the different brands. Therefore, based on their viewing patterns, consumers update their beliefs about the extent to which different brands spend on advertising. Ackerberg further argues that consumers might view advertising expenditure by a brand as a signal of its true quality. For example, only brands that have a high enough quality can afford to invest heavily in advertising. He achieves this by allowing the prior beliefs of the consumers regarding quality and advertising levels to be correlated. Therefore, unlike in Mehta, Rajiv, and Srinivasan (2003) and Crawford and Shum (2005), learning about one unknown entity (advertising expenditure) also helps in learning about another unknown entity (quality). How Are They Learning? (Signals) Researchers have considered how agents learn from a variety of signals such as consumption, usage, advertising, detailing (in case of physicians), and market demand (in case of managers). As discussed above in the section on the structure of learning models, the standard assumption in the literature is that these signals are unbiased (that is, mean = true value of the unknown entity) but noisy indicators of the true value of the unknown entity. What is the source of the signal noise? In instances where the signal is in the form of consumption, the noise could be because of the inherent variability in the product (Roberts and Urban 1988). It could also be due to market conditions (Hitsch 2006; Dixit and Chintagunta 2007) as well as due to the inability of the agents to perfectly evaluate the true value of the signal. For example, it might take consumers several usage occasions to evaluate the stain-removal property of detergents (Erdem and Keane 1996). In the case of signals such as advertising and detailing, the noise might reflect the level of precision in the
72
S. Sriram and Pradeep K. Chintagunta
signal. If the agents view the signal as being less precise, it would take several signals to change their beliefs about the true value of the unknown entity. Such low-precision signals would have a relatively high variance. At the other end of the spectrum, one can envision scenarios wherein the agent realizes the true value of the unknown entity after receiving one signal. For example, it would typically take only one viewing of a movie to evaluate its true quality. In such instances, the signal would have no variability. Nevertheless, most applications in marketing and economics have considered only noisy signals. However, the case of one-shot learning is an outcome that is nested in such models (that is, precision approaches ∞ asymptotically). We can classify the signals into two broad groups based on whether or not they are an outcome of the agent’s decision. We call these intrinsic and extrinsic signals respectively. For example, consumption experience is an outcome of the consumer’s decision to purchase the product. Therefore, it can be classified as an intrinsic signal. Similarly, the realization of demand for a new product is a result of a manager’s decision to keep the product in the market. In the case of a physician, the feedback signals received from patients regarding the efficacy of a prescription drug are realized as a consequence of his decision to prescribe the drug. On the other hand, when a consumer receives advertising signals regarding a product, it is not a result of her actively seeking the information. Therefore, we can classify it as an extrinsic signal. Detailing efforts by pharmaceutical companies directed at physicians fall in the category of signals that are generated by firms and not by the direct actions of the physicians (although presumably a physician is detailed based on his prescription activity). Extensions of the Basic Learning Model In this section, we discuss some ways in which the basic learning model discussed in the previous sections has been extended in the literature to accommodate more realistic decision making by agents. First, we discuss the extension of the decision horizon for the agent from myopic to forward-looking wherein they consider the effect of learning in being able to make more informed decisions in the future. Next, we discuss an extension that accommodates instances where the agents might not fully retain the information they get from the signals. We conclude the section with a discussion of a learning model in which the agents learn from the action of other agents even when these agents do not directly communicate with each other. We discuss this under the subsection entitled “Silent Word of Mouth.” Decision Horizon The basic structure of learning models discussed earlier assumes that the agents maximize only the current utility, that is, they are myopic. The appendix suggests that most of the work in the literature of learning models falls in this category. However, one can envision scenarios when the agents would consider payoffs beyond the current period while making their decisions. To illustrate this point, consider equations 3a´ and 3b´, which provide the expressions for the posterior mean and variance. These equations imply the following. First, from equation 3a´ it is clear that the prior mean belief that an agent has about the uncertain entity for the next period is a function of the current mean and the information that the agent receives during this period. Furthermore, as shown in Figure 3.1, when the signals are unbiased, the prior mean of the unknown entity would asymptotically converge to its true value. Therefore, the agent realizes that as she gathers more information regarding the unknown entity, her beliefs regarding its mean are likely to be closer to the true value. Second, from equation 3b´ it is clear that the prior variance is strictly decreasing in
LEARNING MODELS
73
the amount of information gathered. Therefore, if the agent is unlikely to be risk neutral (that is, have a nonlinear utility function), then the expected utility that she would face at the time of making her decision will be a function of her prior variance (see Table 3.1). Therefore, information gathered would also have an effect on the expected utility of the agent in a future period through its effect on the prior variance in future periods. As a result, she is likely to face a trade-off between the following two choices: (a) possibly making a suboptimal decision in the current period (either keeping a product that has a negative expected current profitability in the market or choosing a product that has a lower expected utility in the current period) and (b) using the information thus gathered to make more-informed decisions in the future. If the benefit of making more-informed decisions in the future outweighs the cost of making a suboptimal decision in the current period, the agent might deviate from the myopically optimal choice. For example, Erdem and Keane (1996) argue that a consumer (the agent) might “experiment” by purchasing a product that has a lower expected utility if she could learn about the product’s attributes from the purchase (and eventual consumption). This “experimentation” will help her make a more informed decision in future periods. Hitsch (2006) illustrates this point using a simple example in the context of a manager (the agent) who has to decide whether or not to keep a new product in the market. The agent’s uncertainty pertains to the profits the new product would generate. Let her belief about the product’s per-period profitability be 1 with probability q (profitable outcome) and –1 with probability 1 – q (unprofitable outcome). Therefore, the expected net present value of profits the new product would generate over an infinite horizon would be
× T + − × − T T − , = − β − β
(8)
where β, 0 ≤ β ≤ 1 is the discount factor. Based on this expected profit, the agent would keep the product in the market only if the probability of the profitable outcome q > 0.5. However, this ignores the possibility that the agent might learn something about the product’s profitability by keeping it in the market. For example, consider that the agent would learn about the true profitability of the product if she keeps it in the market for one period, that is, the signal is completely informative. In such a scenario, the product would have an expected profit of 2q – 1 in the current period (same as above). However, armed with the information regarding the true profitability of the product, the agent can make the optimal decision in future periods. Therefore, the net present value of the profit stream when the agent considers the fact that she could learn about the true profitability of the product would be
T − + β
T . − β
(9)
In the above expression, the first part, 2q – 1, captures the expected profits in the current period, and the second part corresponds to the net present value of the expected profits that the agent would derive after having learned about its true profitability. Under this scenario, the agent − β − β < T> would keep the product in the market as long as . Note that if β > 0. −β −β Therefore, when the discount factor, β, is greater than 0, the agent could make different decisions regarding retaining the product in the market depending on whether or not she considers the role of learning in enabling her to make more informed decisions in future periods. Specifically, the decision criteria for keeping the product in the market would differ under the two scenarios for
74
S. Sriram and Pradeep K. Chintagunta − β
values of q that lie in the range − β < T < . More intuitively, when the agent considers the role of signals in helping her make more informed decisions in the future, the criteria for keeping the product alive are likely to be less stringent. Despite the strong argument in favor of formulating the agent as a forward-looking decision maker, most of the research in learning models has focused on myopic decision makers. In some contexts such as a physician prescribing drugs for a patient, it can be argued that the threat of malpractice lawsuits might prevent the agent from experimenting in order to learn about the product’s true attributes. In a specific empirical context, one can look to the data to get some preliminary evidence regarding experimentation by the agents. So, for example, when looking at new parents making decisions on the brand of diaper to purchase for their firstborn child, one can check to see whether we observe a lot of brand switching along with the purchase of small package sizes soon after birth, with the pattern settling down to infrequent switching and larger package sizes subsequently. Such a purchase pattern might be taken as some evidence consistent with rational forward-looking behavior on the part of agents (see also Crawford and Shum 2005). Nevertheless, the dynamic nature of learning models provides sufficient reason to model the agent as a forward-looking decision maker. The Role of Forgetting Learning models in the literature have typically assumed that the agents can perfectly recall the information accumulated in the past. Mehta, Rajiv, and Srinivasan (2004) invoke the literature on memory and recall (Anderson 1999; Cook and Flay 1978) to challenge this assumption. Specifically, they use evidence from the social psychology literature to formulate a model wherein an agent’s evaluation of an unknown entity converges to its prior value as the time between signals increases. The basic structure of the Mehta, Rajiv, and Srinivasan (2004) model hinges on the conjecture by Alba, Hutchinson, and Lynch (1991) that consumers (the agent) construct their quality evaluations of products using (a) the attribute information on the product’s package and (b) the information retrieved from their memory regarding their quality evaluations of the product from past consumption experiences. In line with this theory, Mehta and his coauthors divide the information that the consumers use to evaluate a brand’s quality into two groups: (a) the positioning set, which corresponds to the information that the consumer learns about the quality of the brand from its advertisements and/or brand name and (b) the consumption set, which contains the cumulative information regarding the consumer’s assessment of the brand’s true quality from her prior consumption experiences. The consumer’s total evaluation of the brand’s quality would be a convex combination of the information contained in these two sets. The authors further assume that while the information contained in the positioning set remains unaltered over time, the information in the consumption set changes as the consumer makes purchases in the category. Moreover, the authors assume that the consumer might not be able to perfectly recall her past evaluations based on the choice set. To accommodate this, they define the recalled evaluation based on the choice set as a sum of the past evaluation based on the choice set and a random term that has a zero mean. The random term implies that the recalled evaluation based on the choice set can be higher or lower than the actual evaluation. Furthermore, a higher variance of this random term would imply greater forgetting, while zero variance would imply no forgetting. In order to allow this level of forgetting to increase with time, they model the variance of the random forgetting term to be a function of the time elapsed since the previous purchase of the brand. When consumers receive new signals about the quality of the brand, they update the evaluation in their consumption set in the typical Bayesian manner with the recalled evaluation taking the role of the prior. The authors
LEARNING MODELS
75
show that the standard learning model without forgetting is a nested version of their model. From their calibration of the model using consumer purchases in the laundry detergent category, the authors show that not accounting for forgetting would lead to biased estimates.2 Silent Word of Mouth The learning models discussed thus far assume that the agents observe the value of the signals they receive. As discussed earlier, these signals could be in the form of their own experiences, advertising exposures, or other marketing activities (such as detailing by pharmaceutical companies). Alternatively, the agents could also learn through word-of-mouth communication regarding the signals received by other agents (Roberts and Urban 1988). In all these instances, the agent observes the real value of the signals even if the signals themselves may be noisy indicators of the true value of the unknown entity. Zhang (2006) extends this by allowing agents to observe only the discrete decisions that other agents make as a result of the signal they receive. However, they would not observe the signals that the other agents actually received (as would be the case in word of mouth). Zhang calls this “Silent Word of Mouth.” Under silent word of mouth, the agent has to infer the distribution of the signals that the other agents would have received based on their actions. Examples of instances where agents make decisions based on silent word of mouth include choice of restaurants based on the number of patrons waiting outside or watching movies based on box office collections. Zhang develops her model in the context of patients (agents) who are waiting in line for a suitable kidney for transplantation. At the time of making her decision, a patient receives her own private signal about the quality of the kidney (the unknown entity) as well as the information that all the other patients ahead of her in the line have rejected the kidney. Since she doesn’t observe the signals received by the patients ahead of her in the line, she has to infer their signals based on their decisions to reject the kidney. In doing so, she also needs to account for the fact that these patients would have rejected the kidney either because they consistently received bad signals or they employed a more stringent criterion for accepting one. Zhang shows that an implication of her model is that if two successive patients receive identical private signals, the decision of the first patient would have an effect on the second patient’s evaluation of the kidney’s quality. Specifically, while an acceptance by the first agent would increase the second agent’s expected evaluation, a rejection would decrease it. As a result, patients toward the end of the line are more likely to turn down the kidney offer. Zhang notes that, given the shortage of available kidneys for transplantation and the clinically acceptable quality of most rejected kidneys, this finding could have implications for policy makers. In particular, facilitation of word-of-mouth communication between patients and suppression of observational learning might be two mechanisms to break the sequential nature of the decision process. Avenues for Future Research In this section, we discuss several possible extensions of the learning literature. Broadly, we group them into three categories: (a) biased signals, (b) changing value of the unknown entity, and (c) integration of Bayesian and alternative learning mechanisms. Biased Signals Learning models typically assume that although the signals received by the agents are likely to be noisy, they are nevertheless unbiased. Therefore, if one were to take the average (across agents and time) of these signals, it would equal the true value of the unknown entity. This
76
S. Sriram and Pradeep K. Chintagunta
assumption is likely to be true when consumers learn about the true quality of a product via consumption experiences or when managers infer the true attractiveness of a market based on observed demand. However, this assumption is likely to be violated in instances where the signal comes from a non-neutral entity. For example, advertising or detailing signals by firms are likely to place a greater emphasis on the positive aspects of their products. Likewise, information reported in news outlets with political leanings are likely to be biased in favor of their respective ideologies (Gentzkow and Shapiro 2008). It would thus be worthwhile to extend the literature to accommodate learning in the presence of such biased signals. If agents are indeed aware of the biased nature of signals, they are likely to discount positive claims and/or place greater emphasis on the negative ones. Time-Varying Values of the Unknown Entity While most of the Bayesian learning literature in marketing has assumed that the value of the unknown entity remains unchanged over time, one can envision scenarios where this is not true. One such scenario is when the true value of the unknown entity fluctuates stochastically around a constant mean (see, for example, Lovett 2008). Under such a scenario, the agent knows that the true value of the unknown entity varies stochastically from time to time as follows:
4W = 4W − + ς W ς W a 1 σ ς , ,
(10)
where Qt is the true value of the unknown entity at time t and ζt is its stochastic variation at time t. Since the mean of this stochastic variation is zero, the true value of the unknown entity fluctuates around a constant mean. Note that the agent does not observe the true value of the unknown entity, Qt. However, the consumer knows the temporal fluctuation in the true value of the unknown entity, ς W a 1 σ ς . The consumer, therefore, observes a noisy measure of the unknown entity such that
4(LW a 1 4W σ 4 , W = 7 L = , .
(2´)
Thus, the only difference between equations 2 and 2´ is that in the latter, the process that generates the signals varies over time. Therefore, when the agent receives signals that vary over time, it could be either because of (a) noise in the signal generating process, σ 4 , or (b) the temporal variation in the signal generating process, ζt. Taken together, equations 10 and 2´ represent the system of equations in a standard Kalman filter (Kalman 1960) with equation 2´ playing the role of the observation (or measurement) equation and equation 10 acting as the system (or state) equation. Hence, the agent’s updating mechanism can be readily obtained based on derivation of the standard Kalman filter (see, for example, Hamilton 1994; Meinhold and Singpurwalla 1983; West and Harrison 1994 for simple exposition of the derivation). In what follows, we present an intuitive discussion of the implications of this stochastic fluctuation. Once again, we note that equation 2´ is similar to equation 2, with the exception that the mean of the process that generates the signals is time varying. An implication of this temporal fluctuation is that as the agent receives signals and updates her belief about the true value of the unknown entity, she also needs to consider this fluctuation. As a result, the agent’s posterior belief at the end of period t–1 and her prior belief at the beginning of period t will not coincide.3 Since the mean of the fluctuation is zero, the posterior mean at the end of period t–1 and the prior mean
LEARNING MODELS
77
at the beginning of period t would remain the same. On the other hand, this will not be the case for the posterior and prior variances. Specifically, if the posterior variance that the agent i perceives about the true value of the unknown entity at the end of period t is σ LW−_W −, then her prior variance at the beginning of period t is
σ LW_W − = σ LW−_W − + σ ς .
(11)
ς
Hence, the variance of the stochastic fluctuation, σ , is added to the posterior variance at the end of period t–1 in arriving at the uncertainty that the agent perceives about the true value of the unknown entity at the beginning of period t. The intuition behind the above equation is that the stochastic fluctuation increases the uncertainty that an agent would perceive about the true quality of the product. As discussed subsequently, it can be shown that when agents anticipate such stochastic fluctuations in the true value of the unknown entity, the rate of reduction of the perceived uncertainty (posterior variance) would be slower than it would be if the true value were time invariant. One can augment the temporal evolution discussed above to allow the mean of the true value of the unknown entity to evolve over time. In a simple extension of equation 10, the true value of the unknown entity can evolve such that
4W = γ 4W − + ς W , ς W a 1 σ ς .
(10’)
Thus, if 0 < γ < 1 and Qt > 0, then the true value of the unknown entity would decrease stochastically over time. An implication of the temporal variation in the mean is that both the prior mean and the prior variance at the beginning of period t would differ from their corresponding posteriors at the end of period t–1. More formally,
4LW_W − = γ4LW −_W − ,
(12)
σ LW_W − = γ σ LW−_W − + σ ς ,
(13)
where, 4LW −_W − and σ LW−_W− are the posterior mean and variance at the end of period t–1 and 4LW_W − and σ LW_W − is the prior mean and variance at the beginning of period t. Furthermore, equations 3a´, 3b´, 3c´, and 3d´ can be rewritten as follows:
4LW_W = θ LW 4LW_W − + ϖ LW 4(LW , σ LW_W =
(3a”)
σ LW_W −
+ , LW σ4
, (3b”)
σ 4 σ L W_W − , θ LW = = + , LW , LW σ L W_W − + σ 4 σ L W_W − σ4
(3c”)
78
S. Sriram and Pradeep K. Chintagunta
, LW
ϖ LW =
σ 4 , + LW
σ L W_W −
= σ 4
, LW σ LW_W − , LW σ LW_W − + σ 4
.
(3d”)
The above formulation can be extended to accommodate the evolution of the true value of the unknown entity, Qt, as a function of other covariates (see, for example, Akcura, Gonul, and Petrova 2004). Notwithstanding the mathematical formulation of the belief updating process, the following question arises: how is this evolution identified? For example, consider the case where the mean of the unknown entity does not vary over time. Under such a scenario, how can one separately identify fluctuations in the delivery mechanism of the unknown entity from the temporal fluctuations in its true value? There are two arguments that favor this identification. First, in the absence of the stochastic temporal variation, the weight that the agent places on the realized values of the unknown entity, ϖit , would steadily decrease over time and asymptotically tend to zero. On the other hand, in the presence of a stochastic fluctuation in the true value of the unknown entity, the weight would decrease at a slower rate and asymptote to a value greater than zero. As in Lovett (2008), we plot the evolution of this weight for different values of σ2ς/σ2Q in Figure 3.2. From this figure, it is evident that for the traditional Bayesian learning model with no stochastic fluctuation in the true value (σ2ς/σ2Q = 0), the weight, ϖit, approaches zero. However, as the stochastic fluctuation gets more pronounced and σ2ς/σ2Q increases, the weight asymptotes away from zero. Therefore, if one had a sufficiently large time series of observations, the extent to which the weight asymptotes away from zero can be used to infer the variance of the stochastic fluctuation, σ2ς. The second argument in favor of identification is based on the fact that if the true value of the unknown entity does not vary over time, the underlying distribution of an agent’s beliefs should remain unaltered over time if she does not receive any signals. On the other hand, if the agent expects temporal fluctuations in the true value of the unknown entity, her beliefs would exhibit changes even in the absence of signals. Hence, to the extent that one observes fluctuations in an agent’s prior beliefs (variance when there is only stochastic variation, and both mean and variance when there is both change in mean and stochastic variation) across several periods when she does not receive signals, the model parameters can be identified. Despite the possible richness of a model that allows for temporal variations in the true value of the unknown entity and the arguments behind identification of such a model, there has been very limited empirical work. It would hence be worthwhile to extend the literature along these lines. Moreover, it would be worthwhile to construct structural versions of reduced form models based on the Kalman filter. Alternatives and Augmentations to Bayesian Learning While most of the discussion in this chapter has revolved around Bayesian learning by agents, alternative learning mechanisms as well as augmentations to Bayesian learning are possible. In this regard, it might be worthwhile to integrate the rich theories in the psychology as well as behavioral marketing literatures with the extant state of the art in Bayesian learning. We believe that there are at least three possible avenues to doing this. First, an implication of Bayesian learning is that an agent’s belief about the entity she is uncertain about does not depend on the order in which
LEARNING MODELS
79
Figure 3.2 Weight Placed on Current Information Over Time 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1
2
3
4
5
6
7
8
9
10
No. Signals (ıȢ)sq/ (ı Q)sq=0
(ıȢ)sq/ (ıQ)sq=0.6
(ıȢ)sq/ (ı Q)sq=1
(ıȢ)sq/ (ıQ)sq=1.4
she receives information (or signals). However, experimental evidence suggests otherwise (see, for example, Hogarth and Einhorn 1992). It would thus be useful to develop alternative learning models that account for order effects in belief updating. Second, when agents receive signals about the entity they are uncertain about, these signals can exhibit either positive or negative deviations from what they expected based on their beliefs. In a standard model of Bayesian updating, both positive and negative deviations are given equal weight. However, according to prospect theory (Kahneman and Tversky 1979), agents tend to place a greater emphasis on negative deviations than they would on positive ones, that is, losses loom larger than gains. Integrating such asymmetric weighting into a model of Bayesian updating would be a worthwhile extension. Third, although Bayesian learning has been very popular in the marketing literature, consumers might use several other heuristics while making decisions as well as in updating their beliefs (Tversky and Kahneman 1974). Extensions that accommodate these heuristics might be worthwhile additions to the literature. Although these extensions will help us better understand the mechanism that agents use to update their beliefs about an unknown entity, the data requirements to facilitate the identification of such augmented models is likely to be stringent. Most of the alternative learning mechanisms as well as heuristics used in decision-making have been tested using experimental data. On the other hand, most of the empirical work on Bayesian learning has used secondary data sources. Hence, researchers need to be cognizant of the additional data requirements that might be needed to model alternative learning mechanisms. Conclusion In this chapter, we present a critical review of learning models in the marketing literature. We also discuss some avenues in which the literature can be extended to accommodate more realistic
80
S. Sriram and Pradeep K. Chintagunta
behavior by agents. Since learning models discussed in this literature are structural in nature, they provide a basis for examining the effect of structural changes in marketing variables on the behavior of agents. We believe that this stream of research provides sufficient grounds to conduct future research that would be interesting to academics and practitioners alike. We hope that this literature review would provide the basic building blocks and a consolidated overview for researchers who seek to push the frontiers of knowledge in this area. Notes 1. Note that in typical random utility models, the presence of the error term εit implies randomness from the perspective of the researcher and not the consumer. In other words, εit is fully known to the consumer. 2. Gowrisankaran, Ho, and Town (2006) consider the situation where doctors learn about complicated heart surgery procedures by performing them on many patients. In their model, they allow these doctors to forget what they have learned over time if they do not practice the procedure. However, their application does not use a Bayesian learning model. 3. Since the mean of the fluctuation is zero, the posterior mean at the end of period t–1 and the period mean at the beginning of period t would remain the same.
References Ackerberg, Daniel. 2003. “Advertising, Learning, and Consumer Choice in Experience Goods Markets: An Empirical Examination.” International Economic Review 44 (3), 1007–1040. Akcura, Tolga, Fusun Gonul, and Elina Petrova. 2004. “Consumer Learning and Brand Valuation: An Application on Over-the-Counter Drugs.” Marketing Science 23 (1), 156–169. Alba, Joseph, J.W. Hutchinson, and John Lynch, Jr. 1991. “Memory and Decision Making.” In Handbook of Consumer Behavior, ed. H.H. Kassarjian and Thomas Robertson, 1–48 NJ: Prentice Hall Professional. Anderson, J.R. 1999. Learning and Memory: An Integrated Approach. NY: John Wiley and Sons. Balvers, Ronald, and Thomas Cosimano. 1990. “Actively Learning the Demand Curve and the Dynamics of Price Adjustment.” Economic Journal 100 (3), 882–898. Ching, Andrew. 2008. “Consumer Learning and Heterogeneity: Dynamics of Demand for Prescription Drugs After Patent Expiration” Working Paper, Rotman School of Management, University of Toronto. Chintagunta, Pradeep, Renna Jiang, and Ginger Jin. 2008. “Information, Learning, and Drug Diffusion: The Case of Cox-2 Inhibitors.” Quantitative Marketing and Economics, Forthcoming. Cook, T.D., and B.R. Flay. 1978. “The Persistence of Experimentally Induced Attitude Change.” Advances in Experimental and Social Psychology 11, 1–57. Coscelli, Andrea, and Matthew Shum. 2004. “An Empirical Model of Learning and Patient Spillovers in New Drug Entry.” Journal of Econometrics 122, 213–246. Crawford, Gregory, and Matthew Shum. 2005. “Uncertainty and Learning in Pharmaceutical Demand.” Econometrica 73 (4), 1137–1173. DeGroot, Morris. 1970. Optimal Statistical Decisions. New York: McGraw-Hill. Dixit, Ashutosh, and Pradeep Chintagunta. 2007. “Learning and Exit Behavior of New Entrant Discount Airlines from City-Pair Markets.” Journal of Marketing 71 (April), 150–168. Erdem, Tulin, and Michael Keane. 1996. “Decision-Making Under Uncertainty: Capturing Dynamic Brand Choice Processes in Turbulent Consumer Goods Markets.” Marketing Science 15 (1), 1–20. Erdem, Tulin, Michael Keane, Sabri Oncu, and Judi Strebel. 2005. “Learning about Computers: An Analysis of Information Search and Technology Choice.” Quantitative Marketing and Economics 3 (3), 207–246. Erdem, Tulin, Michael Keane, and Baohong Sun. 2008. “A Dynamic Model of Brand Choice when Price and Advertising Signal Product Quality.” Marketing Science 27 (6), 1111–1125. Gentzkow, Matthew, and Jesse Shapiro. 2008. “What Drives Media Slant? Evidence from U.S. Daily Newspapers.” Working paper, University of Chicago. Gowrisonkaran, G. Viviantto, and Robert J. Towh. 2006 “Casulity, Learning, and Forgetting in Surgery,” Working Paper. Hamilton, James. 1994. Time Series Analysis. Princeton, NJ: Princeton University Press.
LEARNING MODELS
81
Heilman, Carrie, Douglas Bowman, and Gordon Wright. 2000. “The Evolution of Brand Preferences and Choice Behaviors of Consumers New to a Market.” Journal of Marketing Research 37 (2), 139–155. Hitsch, Gunter. 2006. “An Empirical Model of Optimal Dynamic Product Launch and Exit Under Demand Uncertainty.” Marketing Science 25 (1), 25–50. Hogarth, Robin, and Hillel Einhorn. 1992. “Order Effects in Belief Updating: The Belief-Adjustment Model.” Cognitive Psychology 24, 1–55. Jedidi, Kemal, Carl Mela, and Sunil Gupta. 1999. “Managing Advertising and Promotion for Long-Run Profitability.” Marketing Science 18 (1), 1–22. Jensen, Richard. 1982. “Adoption and Diffusion of an Innovation of Uncertain Profitability.” Journal of Economic Theory 27, 182–193. Kahneman, Daniel, and Amos Tversky. 1979. “Prospect Theory: An Analysis of Decision Under Risk.” Econometrica 47, 263–292. Kalman, R.E. 1960. “A New Approach to Linear Filtering and Prediction Problems.” Journal of Basic Engineering 82, 34–45. Lovett, Mitchell. 2008. “The World is Changing, Isn’t it? Implications for Consumer Learning, Accuracy, and Choice.” Working Paper. Mehta, Nitin, Surendra Rajiv, and Kannan Srinivasan. 2003. “Price Uncertainty and Consumer Search: A Structural Model of Consideration Set Formation.” Marketing Science 22 (1), 58–84. Mehta, Nitin, Surendra Rajiv, and Kannan Srinivasan. 2004. “Role of Forgetting in Memory-Based Choice Decisions.” Quantitative Marketing and Economics 2, 107–140. Meinhold, Richard, and Nozer Singpurwalla. 1983. “Understanding the Kalman Filter.” American Statistician 37 (2), 123–127. Narayanan, Sridhar, Puneet Manchanda, and Pradeep Chintagunta. 2005. “Temporal Differences in the Role of Marketing Communication in New Product Categories.” Journal of Marketing Research 42 (August), 278–290. Narayanan, Sridhar, Pradeep Chintagunta, and Eugeno Miravete. 2009. “The Role of Self Selection, Usage Uncertainty, and Learning in the Demand for Local Telephone Service.” Quantitative Marketing and Economics 5, 1–34. Narayanan, Sridhar, and Puneet Manchanda. 2009. “Heterogeneous Learning and the Targeting of Marketing Communications for New Products.” Marketing Science 28 (3), 424–44. Roberts, John, and Glen Urban. 1988. “Modeling Multiattribute Risk and Belief Dynamics for New Consumer Durable Brand Choice.” Management Science 34 (2), 167–185. Sriram, S., Pradeep Chintagunta, and Ramya Neelamegham. 2006. “Effects of Brand Preference, Product Attributes, and Marketing Mix Variables in Technology Product Markets.” Marketing Science 25 (5), 440–456. Sriram, S., and Manohar Kalwani. 20097. “Optimal Advertising and Promotion Budgets in Dynamic Markets with Brand Equity as the Mediating Variable.” Management Science 53 (1), 46–60. Tversky, Amos, and Daniel Kahneman. 1974. “Judgment Under Uncertainty: Heuristics and Biases.” Science 185, 1124–1131. West, Mike, and Jeff Harrison. 1994. Bayesian Forecasting and Dynamic Models. New York: Springer. Xiao, Ping, Tat Chan, and Chakravarthi Narasimhan. 2006. “Product Bundles Under Three-Part Tariffs.” Working paper, Washington University, St Louis. Zhang, JuanJuan. 2006. “The Sound of Silence: Evidence of Observational Learning from the U.S. Kidney Market.” Working paper.
Context/Industry
Consumer durables
Packaged goods
Packaged goods
Packaged goods
Pharmaceuticals
Packaged goods
Pharmaceuticals
Computers
Reference
Roberts and Urban (1988)
Erdem and Keane (1996)
Ackerberg (2003)
Mehta, Rajiv, and Srinivasan (2003)
Coscelli and Shum (2004)
Mehta, Rajiv, and Srinivasan (2004)
Crawford and Shum (2005)
Erdem, Keane, Oncu, and Strebel (2005)
Summary of Selected Studies on Learning
Appendix 3.1
Consumer
Physician
Consumer
Physician
Consumer
Consumer
Consumer
Consumer
Agent
Quality
Symptomatic and curative Efficacy
Quality
Quality of a new drug for treating different diagnoses
Quality and Price
Quality and advertising intensity
Quality
Quality
Unknown Entity
Individual level
Individual level (stated preference and willingness to recommend)
Type of Data
Retail stores, publications, ads, and word of mouth
Consumption
Consumption
Prescription and patient feedback
Brand consideration and consumption
Individual level
Individual level
Individual level
Individual level
Individual level
Consumption and watch- Individual level ing advertisements
Consumption and ads
Word of mouth
Signals
Forwardlooking
Forward looking
Myopic
Myopic
Myopic
Forward looking
Forward looking
Myopic
Decision Horizon
82
Pharmaceuticals (prescription)
Telephone
Packaged goods
Kidney
Discount airlines
Pharmaceuticals (prescription)
Wireless
Pharmaceuticals (prescription)
Pharmaceuticals (prescription)
Narayanan, Manchanda, and Chintagunta (2005)
Narayanan, Chintagunta, and Miravete (2007)
Hitsch (2006)
Zhang (2006)
Dixit and Chintagunta (2007)
Narayanan and Manchanda (2007)
Xiao, Chan, and Narasimhan (2006)
Ching (2008)
Chintagunta, Jiang, and Jin (2008)
Physician
Patient and physician
Consumer
Physician
Airlines
Patient
Firm with a new product
Consumer
Physician
Consumption
Patient feedback, detailing, other marketing activities by pharmaceutical companies Individual level
Aggregate
Overall quality of drug and patientdrug match
Quality of generics
Preference for voice and text
Quality of drugs
Attractiveness of markets
Quality
Patient satisfaction
Consumption
Consumption or usage
Individual level
Aggregate
Individual level
Physician level
Firm level
Market demand Patient feedback, detailing
Individual level
Silent word of mouth (observation of other patients’ choices)
Demand and profit- Observed market shares Individual prodability potential uct level
True consumption rate
Quality of drugs
Myopic
Myopic
Myopic
Myopic
Myopic
Myopic
Forward looking
Myopic
Myopic
83
Chapter 4
CUSTOMER CO-CREATION A Typology and Research Agenda Matthew S. O’Hern and Aric Rindfleisch
Abstract Traditional marketing thought and practice largely view new product development (NPD) as an internal, firm-based activity in which customers are relatively passive buyers and users. This traditional paradigm is currently being challenged by a new perspective in which customers are active co-creators of the products they buy and use. This chapter identifies the origins of this paradigm shift, presents a conceptual typology of four different types of co-creation activity, and offers an agenda for future research of this emerging paradigm. New product development (NPD) is an important driver of corporate growth and profitability (Sorescu, Chandy, and Prabhu 2003; Wind and Mahajan 1997). Unfortunately, most new products fail to deliver on their objectives (Christensen 1997). Hence, marketing scholars and practitioners have duly devoted substantial attention toward improving NPD processes. This attention has led to several important advances, including the specification of the Stage-Gate model (Cooper 1990), the formulation of sophisticated NPD tools such as conjoint analysis and premarket launch forecasting (Rangaswamy and Lilien 1997), and advances in knowledge about how best to organize and manage NPD teams (Sethi, Smith, and Park 2001). These core topics of NPD research and practice share an important but often unstated assumption that NPD is essentially an internal, firm-based activity. As recently observed by von Hippel (2005, p. 19), “The idea that novel products and services are developed by manufacturers is deeply ingrained in both traditional expectations and scholarship.” Hence, NPD research and practice largely operate under a firm-centered paradigm in which customers are viewed as having little active influence upon NPD activity. While this paradigm may have served academics and practitioners well in the past, it is currently being challenged by the emergence of empowered customers seeking greater input and control over NPD activity (Seybold 2006). This challenge is ushering in a new paradigm in which firms can enhance corporate growth and profitability by allowing customers to take a more active role in NPD activity (Prahalad and Ramaswamy 2000; von Hippel 2005). In this newly emerging co-creation paradigm, customers are central and vital participants in the NPD process and, in some cases, are capable of creating new products with little help from firms. For example, many of today’s most successful computer applications, including Apache, 84
CUSTOMER CO-CREATION: A Typology and Research Agenda
85
Figure 4.1 Growth of the Open Source Software Movement
2,000,000
Number of Registered Users
1,800,000 1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0 Mar-00 Mar-01 Mar-02 Mar-03 Mar-04 Mar-05 Mar-06 Mar-07 Mar-08 Source: SourceForge.net.
Linux, and Firefox, are open source projects that are managed by self-organizing communities of volunteer programmers. As shown in Figure 4.1, the open source movement has experienced tremendous growth in recent years. Likewise, many computer game modifications are developed by players rather than manufacturers (Jeppesen and Molin 2003). Customers are also becoming actively involved in the creation of traditional manufactured products. For instance, over 120,000 individuals around the world served as voluntary members of Boeing’s World Design Team and contributed ideas and input regarding the design of its new 787 Dreamliner airplane (www.newairplane .com). Likewise, Arduino, an Italian microcontroller manufacturer, provides open access to its software and schematics and actively encourages customers to tinker with its product design (www.arduino.cc). Although the literature on this topic is sparse, the evidence marshaled thus far suggests that customer co-creation is positively associated with several NPD metrics, including increased new product creativity, decreased time to market, and reduced development costs (Grewal, Lilien, and Mallapragada 2006; Shah 2006; von Hippel 2005). This new paradigm has attracted the attention of the Marketing Science Institute, which has identified customer co-creation as a top research priority (Marketing Science Institute 2008). Likewise, Vargo and Lusch (2004) recognize customer co-creation as a foundational premise underlying marketing’s new service-dominant logic. Moreover, the importance of encouraging and utilizing customer-generated solutions has been noted by several leading innovation researchers and practitioners, including Cook (2008), Evans and Wolf (2005), Prahalad and Ramaswamy (2004), Seybold (2006), and von Hippel (2005), among others. However, to date, marketing scholars have devoted scant attention to customer co-creation and, instead, continue to focus on NPD as largely a firm-based activity. Consequently, little is known about the nature of this phenomenon or its
86
Matthew S. O’Hern and Aric Rindfleisch
implications for marketing thought and practice. To date, the limited body of marketing scholarship on customer co-creation has largely focused on specific exemplars of this phenomenon, such as assembling a store-bought bookshelf (Bendapudi and Leone 2003), modifying a computer game (Prügl and Schreier 2006), or developing open source software (Grewal, Lilien, and Mallapragada 2006; Rajagopalan and Bayus 2008). Although these studies provide an important contribution by examining the motives and mechanisms underlying customer co-creation, a focus on specific exemplars provides only a narrow look at a complex and multifaceted phenomenon (Cook 2008; Seybold 2006). Our research seeks to address this gap by offering a broader examination of various forms of customer co-creation, which we define as a collaborative NPD activity in which customers actively contribute and/or select the content of a new product offering.1 In accordance with this definition, customer co-creation involves two key processes: (1) contribution (that is, submitting content) and (2) selection (choosing which of these submissions will be retained). Using these two processes as our foundation, we offer a conceptual typology of four different forms of customer co-creation as well as an agenda for future research in this domain. We believe that this typology and agenda have the potential to enhance both marketing scholarship and practice. Specifically, we provide scholars with a nuanced understanding of the commonalities and distinctions among these various types of co-creation while offering practitioners an examination of their payoffs and challenges. In order to provide a contextual backdrop, our thesis begins with an examination of the trends fueling the rise of customer co-creation. The Rise of Customer Co-Creation Successful NPD requires two essential types of information: (1) information about customer needs and (2) information about how best to solve these needs (Thomke and von Hippel 2002; von Hippel 2005). Typically, customers (or users) have the most accurate and detailed knowledge about the first type of information, while manufacturers (or providers) have the most accurate and detailed knowledge about the second type. This disparity creates a condition of information asymmetry (von Hippel 2005). Traditionally, firms have attempted to manage this asymmetry by engaging in various forms of marketing research to obtain better information about their customers’ needs. Under this approach, “Successful innovation rests on first understanding customer needs and then developing products to meet those needs” (Hauser, Tellis, and Griffin 2006, p. 3). Unfortunately, customer needs are often idiosyncratic and tacit in nature and, hence, hard to accurately measure and coherently implement (Franke and Piller 2004; Simonson 2005). As suggested by von Hippel (2005), consumers have deep and complex (“high fidelity”) needs; however, traditional market research methods often provide managers with only a cursory (“low fidelity”) signal of what customers want or need. As a result, most new product failures are attributed to a firm’s inability to accurately assess and satisfy customer needs (Ogawa and Piller 2006). As recently noted by von Hippel and colleagues (Thomke and von Hippel 2002; von Hippel 2005; von Hippel and Katz 2002), an alternative and emerging means of bridging this asymmetry is to provide customers with information and tools that enable them to take a more proactive role in the NPD process. As we detail subsequently, an increasing number of firms are employing this new approach in various manifestations. This movement toward providing consumers with greater autonomy over NPD activity has witnessed tremendous growth in recent years due to the rise of customer empowerment. In the remainder of this section, we briefly discuss these trends and their implications for co-creation.
CUSTOMER CO-CREATION: A Typology and Research Agenda
87
As noted earlier, the traditional NPD paradigm largely views consumers as rather passive entities who are highly dependent upon firms to help satisfy their needs (Carpenter, Glazer, and Nakamoto 1994; Simonson 2005). Although many individuals still conform to this traditional role definition, recent cultural developments are empowering a growing number of end users to play a more active role in developing the products they buy and use. One important cultural development is consumers’ growing suspicion and distrust of marketing communications. For example, a considerable body of research suggests that consumers are quite skeptical of marketing communication in general and claims about new product performance in particular (Darke and Ritchie 2007; Obermiller and Spangenberg 1998; Wright 2002). In recent years, this inherent skepticism, fueled by increased news coverage of corporate scandals (such as Enron, ImClone), muckraking documentaries of big business (for example, SuperSize Me, The Corporation), and anticorporate websites (for example, adbusters.org, spacehijackers. org), has ignited more active forms of consumer resistance such as anticorporate blogging, brand avoidance, and culture jamming (Klein, Smith, and John 2004; Kozinets and Handelman 2004; Thompson, Rindfleisch, and Arsel 2006). Hence, an increasing number of consumers are engaging in direct action to alter corporate marketing activities that they find objectionable. This increased consumer agency represents a significant strategic challenge and has led several large firms, including Wal-Mart, Nike, and McDonalds, to be more cognizant of and open to customer input (Kalaignanam and Varadarajan 2006). For example, in order to appease consumer activists, Nike has taken steps to actively engage customers in many facets of its strategic planning and execution (Seybold 2006). In addition to a growing suspicion and heightened activism, consumers also appear to be increasingly less fulfilled by the act of consumption itself (Firat, Dholakia, and Venkatesh 1995). The notion that material objects are unable to satisfy intrinsic psychological needs has been strongly established by consumer researchers (Belk 1985; Richins and Dawson 1992), and these findings have recently been disseminated to the broader public (Kasser 2003; Kohn 1999; Schor 1998). According to cognitive psychology, intrinsic needs are more likely to be met via creative pursuits (Csikszentmihalyi 1996; Deci and Ryan 1985). Thus, through their creative contributions, customers may reap psychological benefits they would normally be unable to achieve via consumption alone. Indeed, many of today’s popular television programs glorify the creative process (for example, American Chopper, Trading Spaces, This Old House), and creative pursuits such as cooking, crafts, and home improvement are rapidly growing in popularity (Ebenkamp 2005; Pietrykowski 2003). Recent research on customer-led innovation reveals that users often find this activity highly enjoyable (Lakhani and Wolf 2005; Shah 2006).2 Spurred by these cultural influences, an increasing number of consumers are seeking a more active role in the creation of the products they consume (Handelman 2006; Roberts, Baker, and Walker 2005). For example, nearly one third of the members of extreme sports communities (such as sailplaning, canyoning, and snowboarding) engaged in some form of product innovation, and almost a quarter of these innovations were later incorporated into existing products by manufacturers (Franke and Shah 2003). Moreover, research on the creative potential of brand communities suggests that consumers are willing and able to introduce new offerings even after a product (for example, Apple Newton) is long abandoned by the firm that sold it (Muñiz and Schau 2005). The ability of consumers to take a more active role in NPD has been significantly enhanced by recent technological advances, most notably the development and growth of the Internet. According to several researchers, consumers have traditionally lacked the technical skills and capabilities that NPD requires (Christensen 1997; Randall, Terweisch, and Ulrich 2005; Simonson 2005). However, the Internet has helped ameliorate this deficiency and empower customers in at least three ways.
88
Matthew S. O’Hern and Aric Rindfleisch
First, the Internet increases access to knowledge that can enhance consumers’ ability to engage in creative pursuits. For example, consumers interested in learning how to build an electric car can find several websites that contain detailed technical information and user-friendly tutorials on this topic (for example, www.evadc.org; www.makezine.com; www.evsupersite.net). Hence, through these electronic archived data sources, knowledge that was once tacit and remote has now become codified and proximate (Jeppesen and Molin 2003). Second, the Internet also facilitates consumers’ ability to apply their knowledge by providing access to a variety of online design tools (Prahalad and Ramaswamy 2004; Thomke and von Hippel 2002). For example, fans of popular computer games such as Half-Life and The Sims can access Internet-based programs that enable them to create their own modifications and extensions to these games. Similar types of online design tools can also be found for website development, podcasting, and digital audio/video production. According to von Hippel (2005, p. 123), these tools “are often as good as those available to professional designers,” and research suggests that they are instrumental in encouraging end users to experiment with and improve their own products (Prügl and Schreier 2006). As a result, in many fields, an increasing number of consumers are acquiring skills and knowledge that nearly equal those of a firm’s internal NPD team (Leadbeter and Miller 2004; Prahalad and Ramaswamy 2004). Third, in addition to enriching the creative capabilities of individual consumers, the Internet enhances collective co-creation by connecting individual consumers with others (both consumers and manufacturers) in a manner that allows them to participate effectively in a co-creation community (Moon and Sproull 2001; Prahalad and Ramaswamy 2000; Sawhney, Verona, and Prandelli 2005). These communities enable consumers to learn from (and teach) other consumer-creators (Prügl and Schreier 2006) and help form collective knowledge and memory systems that transcend the information and skills of any single individual (Jeppesen and Molin 2003; Leadbeter and Miller 2004). For example, open source computer software is typically developed via self-organized communities of thousands of contributors who work in a highly collaborative manner and play a variety of different roles. This collective information exchange enables these co-creation communities to create offerings that can equal or surpass traditional firm-based NPD activity in terms of development speed, creativity, and marketplace success (Shah 2006). In sum, growing customer empowerment appears to be rapidly promoting customer co-creation by motivating consumers to play a more active role in the NPD process, enhancing their NPD knowledge and skills, and connecting them with proactive communities of like-minded individuals. This emerging trend presents an exciting opportunity for marketing researchers and practitioners to employ co-creation as a potential alternative to the traditional NPD paradigm. A Typology of Customer Co-Creation In recent years, the rise of co-creation has garnered considerable attention across a broad range of fields, including information systems, economics, management, and marketing (e.g., Sharma, Sugumaran, and Rajagopalan 2002; Etgar 2008; Evans and Wolf 2005; von Hippel and Katz 2002; Pitt et al. 2006; Prahalad and Ramaswamy 2004; Vargo and Lusch 2004). This research has uncovered a variety of different forms of co-creation, ranging from enhanced customer sensing techniques embedded within largely firm-driven NPD processes to open source innovation occurring beyond the boundaries of the firm. In this section, we conceptually synthesize this diverse array of co-creation initiatives into a coherent typology. As previously noted, a growing number of consumers are seeking increased autonomy and displaying higher levels of empowerment over the NPD process. According to the organizational
CUSTOMER CO-CREATION: A Typology and Research Agenda
89
creativity literature, a high degree of autonomy enhances creativity (Amabile et al. 1996; Velt house 1990; Woodman, Sawyer, and Griffin 1993). For example, Amabile and her coauthors find that mid-level managers with substantial autonomy are considerably more likely to generate creative projects compared to mid-level managers with limited autonomy. The positive effect of autonomy is largely believed to be due to its ability to cultivate high levels of intrinsic motivation and psychological ownership, which in turn facilitates creativity by making the creative task more enjoyable and rewarding (Csikszentmihalyi 1996; Deci and Ryan 1985). A release of control should have a similar (and perhaps even greater) effect on customer co-creation, as customers (unlike employees) typically receive little or no direct financial compensation for their creative efforts and, thus, must possess high levels of intrinsic motivation in order to engage in and sustain such activity (Seybold 2006). Hence, releasing control of NPD activities should, theoretically, lead to more creative and successful offerings. Although a few firms are beginning to recognize the potential benefits of releasing NPD control (Seybold 2006), many are reluctant to enhance customer autonomy due to concerns about leaking valuable proprietary information, ceding managerial power, and losing control over their brands (Pitt et al. 2006; von Hippel 2005). For example, Intuit’s chairman, Scott Cook, recently revealed that upper-level managers at his firm resisted customer co-creation because of its “challenge to long unquestioned beliefs about the role of management, the value of experts, the need for control over customer experience, and the importance of quality assurance” (Cook 2008, p. 68). This type of reluctance may be well placed, as the marketing strategy literature suggests that tight and systematic managerial controls (such as the Stage-Gate model) enhance NPD success (Cooper 1990; Song and Parry 1997). Consequently, co-creation initiatives display considerable variance in the degree to which they empower customers by allowing them greater autonomy over the NPD process, especially at its early stages (that is, fuzzy front end). The NPD literature suggests that the early stages of developing a new product entail two essential activities: (1) the contribution of novel concepts and ideas, and (2) the selection of which specific concepts and ideas should be pursued (Kahn 2005). In many firms, both of these activities are closely guarded and typically conducted by a small number of employees (that is, an NPD team) (Sethi, Smith, and Park 2001; Song and Parry 1997). In most cases, customers are not actively engaged in either activity. Thus, firms can engage in customer co-creation by releasing control of either the contributions made to the NPD process and/or the selection of these contributions. Consequently, the degree of customer autonomy across these two activities forms the conceptual basis for our typology. Our typology is depicted in Figure 4.2. As shown in this figure, we depict contribution and selection as two distinct NPD activities that vary in the degree to which a firm releases control and empowers its customers as active participants. Our depiction acknowledges that the balance between control and empowerment lies along a continuum from low to high. Specifically, we suggest that the type and format of NPD contributions can range from being essentially fixed by a firm to wholly open to customer input and that the selection of these contributions can be either directed by a firm or directed by customers. When arranged along two dimensions, these activities allow us to derive four distinct types of customer co-creation: (1) collaborating, (2) tinkering, (3) co-designing, and (4) submitting, with submitting at one extreme (fixed contribution and firm-led selection) and collaborating at the other (open contribution and customer-led selection). Although non-exhaustive, we believe that this typology classifies a considerable body of co-creation activity.3 In the remainder of this section, we define each of these types of co-creation, identify their key features, provide relevant exemplars, and discuss their benefits and limitations (see Table 4.1 for a summary).
Submitting
Firm-led
Fixed
Co-designing Customer-led Fixed
Open
Increased access to Retaining and motivating novel customer ideas existing co-creators
Acquiring knowledgeable new co-creators
Defending against new entrants
Decreased risk of product failure Shortened product development cycles
Attracting a critical mass of designers
Policing the content of rogue co-creators Creating new competitors
Attracting a critical mass of collaborators
Protecting intellectual property
Key challenges
Reduced development costs
Enhanced differen tiation Virtual test markets for new products
Firm-led
Tinkering
Contribution activity Key payoffs Reduced development costs Continuous product improvement
Selection activity
Collaborating Customer-led Open
Type of co-creation
Characteristics of Co-Creation Types
Table 4.1
Company-sponsored design competitions
Online voting on customer-generated content and designs
Modified computer games
Open source software
Füller et al. (2004) Sawhney et al. (2005)
Cook (2008)
Ogawa and Piller (2006)
Nieborg (2005) Prügl and Schreier (2006)
Jeppesen and Molin (2003)
Grewal et al. (2006) Lakhani and Wolf (2005) von Krogh et al. (2003)
Prototypical application Key studies
90
CUSTOMER CO-CREATION: A Typology and Research Agenda
91
Customer-led
Collaborating
Co-designing
Tinkering
Firm-led
SELECTION ACTIVITY
Figure 4.2 Four Types of Customer Co-Creation
Submitting
Fixed
Open
CONTRIBUTION ACTIVITY
Collaborating We define collaborating as a process in which customers have the power to collectively develop and improve a new product’s core components and underlying structure. As shown in Figure 4.2, we conceptualize collaborating as the form of co-creation that offers customers the greatest power to contribute their own ideas and to select the components that should be incorporated into a new product offering. At present, the best examples of collaborating can be seen in open source software initiatives such as Linux, Apache, and Firefox. In contrast to commercial software, which places considerable restrictions on consumer usage, open source software empowers users to make fundamental changes to a program’s basic structure (that is, source code). This openness also influences the way intellectual property is managed, as many open source licenses dictate that program changes be made freely available to other users. In recent years, open source applications have gained widespread adoption and market success. For example, Apache, an open source application, dominates the worldwide market for web-server software with over a 70 percent market share (Grewal, Lilien, and Mallapragada 2006). In addition to software development, collaborators are making important and innovative contributions in several other areas, including agriculture (www.cambria.org), pharmaceutical products (www.tropicaldisease.org), medical devices (www. designthatmatters.org) and architecture (www.architectureforhumanity.org). While many collaborator-based projects are managed by nonprofit organizations, the principles underlying this form of co-creation may be usefully employed by for-profit firms (Evans and Wolf 2005; Shah 2006). In fact, some software firms release the source code for selected commercial
92
Matthew S. O’Hern and Aric Rindfleisch
products into the open source community in hopes of establishing a dominant technological platform. Sun Microsystems, for example, employed this strategy when it made the source code for its NetBeans development tool freely available. Today, NetBeans has been downloaded more than eight million times and enjoys contributions from over 300,000 collaborators around the world (www.netbeans.org). Similarly, IBM uses collaborating as a key part of its competitive strategy by employing hundreds of open source programmers dedicated to enhancing the Linux operating system (Harris 2004). IBM also actively encourages open source development outside the firm by serving as a founding member of the Open Invention Network, a company that supports open source innovation by purchasing software patents and making them available on a royalty-free basis. We suggest that collaborating provides customers with a high degree of latitude to contribute their own new product improvements and gives them substantial freedom to select the new product improvements they find most valuable. In the case of open source software development, the underlying offering is based on open standards that grant all customers (who have the requisite skills) the ability to fully customize the product to better satisfy their own unique needs (Lakhani and Wolf 2005; Shah 2006). Hence, collaborating grants customers almost unlimited freedom to alter the underlying product, thereby transforming customers from passive users to active contributors (Pitt et al. 2006; Evans and Wolf 2005; von Krogh, Spaeth, and Lakhani 2003). Moreover, collaborators are often responsible for forming their own project teams. These teams exist outside the traditional boundaries of the firm, are organized in a non-hierarchical fashion, and rely on a community-based governance system to evaluate and select the inputs made by fellow collaborators (Grewal, Lilien, and Mallapragada 2006; Lakhani and Wolf 2005; von Krogh, Spaeth, and Lakhani 2003). Hence, this customer-led selection mechanism stands in stark contrast to traditional NPD teams in which the selection process is largely confined to a few select employees. Although some individuals engage in collaborating for extrinsic rewards (such as enhancing their career opportunities or gaining status or recognition), most collaborators appear to be intrinsically motivated by a strong philosophical belief in the importance of their work as well as by a deep enjoyment of contributing their thoughts and ideas (Evans and Wolf 2005; Hertel, Niedner, and Herrmann 2003; Sharma, Sugumaran, and Rajagopalan 2002). Due to this high level of intrinsic motivation, collaborating has the potential to generate high levels of co-creator involvement (Lakhani and Wolf 2005; von Hippel 2005). Moreover, recent research suggests that collaborating can improve NPD performance by accessing novel sources of customer-held knowledge held across a diverse set of individual contributors (Grewal, Lilien, and Mallapragada 2006). By generating a high degree of involvement and accessing diverse knowledge domains, collaborating appears to be a highly effective means of generating innovative and successful new products. For example, the open source–based Firefox web browser competes successfully against Microsoft’s dominant Internet Explorer and is widely regarded as the most innovative browser currently available (Vogelstein 2008). In addition to spurring innovation, collaborating can dramatically lower the costs of NPD by using unpaid customers to replace salaried employees. Moreover, unlike traditional NPD projects, which have finite start and end dates, collaborating is an ongoing process. This quality should help firms stay on the leading edge by providing a mechanism for continuous product improvement and enhance customer welfare by accelerating the pace at which new innovations can be created and distributed to users (von Hippel 2005). While collaborating may produce substantial benefits, this form of co-creation also faces a number of challenges. Most importantly, collaborating appears to best suited for informationrich applications (for example, software development, medical research, digital graphic design) and, thus, may be the most challenging form of co-creation for more traditional industries such as packaged goods or consumer durables. Moreover, in order to drive innovation, collaborating
CUSTOMER CO-CREATION: A Typology and Research Agenda
93
requires that at least a small cadre of co-creators obtain a high level of skill and knowledge. This entry requirement may discourage lower-skilled and less knowledgeable customers (who may nevertheless have interesting ideas) from fully participating in the innovation process. Moreover, successful collaboration requires firms to cede managerial authority over NPD and loosen their control over their intellectual property (Cook 2008). Hence, collaborating may be especially difficult for firms with highly centralized organizational structures and large investments in proprietary assets. Thus, firms may be forced to strike a difficult balance between control and openness. For example, Shah (2006) finds that when firms engaged in collaboration initiatives restrict individuals’ freedom to alter and share new product improvements, customers’ willingness to contribute decreases and the risk of customer defections increases. Tinkering We define tinkering as a process in which customers make modifications to a commercially available product and some of these modifications are incorporated into subsequent product releases.4 Tinkering is similar to collaborating in terms of allowing customers a relatively high (but somewhat lower) degree of autonomy over NPD contributions. However, firms that employ tinkering usually retain a considerable degree of control over the selection of these contributions. At present, tinkering is most apparent in the computer game industry, where user-generated contributions (that is, modifications) are not only widely tolerated, but actively encouraged. For example, many game manufacturers invite users to make alterations ranging from incremental changes, such as edits to a character’s physical appearance, to more radical innovations, such as the creation of a completely new computer game. In order to assist tinkerers in making these changes, several computer game manufacturers provide customers with free or low-cost design tools that are similar or even identical to those used by their in-house software developers (Moon and Sproull 2001; Nieborg 2005). This strategy often leads to unexpected and innovative creations, widespread adoption by other gamers, and marketplace success for the firm that produced the base game. For example, over 90 percent of the content of the widely successful computer game, The Sims, is derived from tinkerer-based modifications (called “mods” in the parlance of gamers) (Leadbeter and Miller 2004). The contributions of tinkerers are not, however, limited to computer gaming. Tinkering is also quite common in other information-based products, such as customized digital music and individually tailored web-based applications. For example, leading Internet firms such as Google.com and Amazon.com offer open access to their application program interfaces (APIs). Consumers can combine these open APIs with data from third parties or self-created content to generate innovative hybrid creations known as “mashups.” One impressive user-generated mashup is the website Chicagocrime.org, which melds information from the Google Maps API with a database from the Chicago Police Department (CPD). This co-created website allows users to create their own customized visual display of reported crimes for any street, neighborhood, travel route, or time period they wish to select. This mashup is considerably more visually appealing and interactive than CPD’s traditional database and provides customers with enhanced product functionality and a more enjoyable online experience. At present, Chicagocrime.org receives over 500,000 hits per month. Google also directly benefits from the efforts of tinkerers like the creators of Chicagocrime .org, as the changes they make provide a nearly continuous stream of new content that enhances product functionality and helps differentiate Google Maps from its competitors. Like collaborating, tinkering begins with the release of a basic building block (such as base commercial product and development tools). However, in contrast to collaborators, tinkerers do not
94
Matthew S. O’Hern and Aric Rindfleisch
have unfettered access to a product’s source code. This is an important point, because firm-based control over the underlying product and its source code limits the scope of the product improvements that tinkerers can develop. In addition, tinkerers must typically sign end-user licensing agreements. This means that firms have the power to revoke tinkerers’ rights to use development tools and can limit tinkerers’ ability to share their creations with other users. Thus, customer-created content that was once freely available may be restricted after the firm launches a new product release that includes these customer-led improvements. As a result, tinkering allows a firm to exert substantial control in determining which new customer-generated product improvements are selected to appear in its official new product releases. These characteristics are depicted in Figure 4.2, which positions tinkering as exhibiting less openness (compared to collaborating) in terms of customer contribution and a heavily firm-led selection approach. Although knowledge regarding the outcomes of tinkering is at a formative stage, it appears that this type of co-creation may deliver several benefits to firms. In crowded markets with similar offerings, tinkering may provide a basis for product differentiation. For example, the powerful development tools included in the Unreal Tournament 2004 (UT 2004) Special Edition DVD distinguished this product from similar offerings and helped UT 2004 become a highly successful computer game (Nieborg 2005). Firms may also benefit from the activities of tinkerers in terms of enhancing customer satisfaction and extending their market reach. By supporting and leveraging the contributions of tinkerers, firms can assist customers in satisfying their own needs and sharing their solutions with other customers who may have similar needs. Firms can also benefit from the actions of tinkerers in terms of enhancing market acceptance of their in-house NPD efforts. For example, LucasArts allowed images and music from its Star Wars films to be incorporated into the customer-led mod Galactic Conquest, even though it was simultaneously developing its own proprietary Star Wars–themed computer game (Nieborg 2005). By visiting websites where Galactic Conquest devotees congregated, LucasArts was able to identify and contact users who downloaded this popular mod and determine which customer-generated content they found most appealing. This approach provided LucasArts with both direct access to customer-created innovations as well as a virtual test market for its fledgling commercial product, which ultimately incorporated many of the mod’s most popular features. Although tinkering may provide a number of benefits, it also poses several challenges. First, in most (but certainly not all) cases, the act of tinkering requires a considerable degree of user knowledge and expertise about the product to be modified as well as its underlying technology. However, with the increasing availability of user-friendly development tools, consumers who are not expert users can readily acquire basic tinkering capabilities with moderate learning costs.5 Tinkering also presents the risk that high-quality (and freely available) mods may dissuade customers from purchasing a firm’s future new releases. Thus, firms that actively encourage tinkering may find that their customers have become a formidable source of competition (Cook 2008). In this sense, tinkering can raise the NPD bar, as a firm must ensure that new releases surpass both the functionality of its existing products while also demonstrating superiority over versions that have been created and made freely available by tinkerers. It is also possible that the actions of tinkerers can have a negative impact on a firm’s brand equity. For example, some tinkerers may modify computer games in ways that are especially violent and/or sexually explicit. Imagine, for instance, the case of a customer transforming a World War II combat game into a mod set in an American high school, in which players amass points by gunning down their fellow students and teachers. Most consumers would be appalled by such a game, and the surrounding media attention would undoubtedly reflect very negatively on the firm that created the base product. Thus, the level of contribution autonomy provided by tinkering may
CUSTOMER CO-CREATION: A Typology and Research Agenda
95
be a double-edged sword, as firms that rely heavily on tinkering may be particularly vulnerable to the negative actions of rogue co-creators. Co-designing We define co-designing as a process in which a relatively small group of customers provides a firm with most of its new product content or designs, while a larger group of customers helps select which content or designs should be adopted by the firm. As shown in Figure 4.2, co-designing is characterized by a relatively fixed contribution approach but a high degree of customer autonomy over the selection of these contributions. One of the best examples of co-designing is the online clothing manufacturer Threadless.com. This firm actively solicits original T-shirt designs from current and potential customers and then invites its extensive network of online customers to evaluate and select a short list of prospective new products (Chafkin 2008). Similarly, both the online news service Digg.com (www.digg.com) and the cable television channel Current TV (www.current.tv) acquire much of their content directly from their users. In contrast to the standard approach used by commercial news organizations, Digg.com eschews hierarchical editorial control and instead allows its community of over 300,000 registered reviewers to vote on the stories they deem worthy for display. Likewise, Current TV provides amateur film makers with the opportunity to upload their homemade documentaries and gives viewers the chance to select the clips that air on the network.6 This co-creation approach has been utilized across a wide variety of product categories, including sporting goods, household products, home appliances, and consumer packaged goods (Ogawa and Piller 2006). For example, Jones Soda (www.jonessoda.com) uses co-designing to differentiate its products by displaying customer-submitted photographs (which are rated online by its users) on its product labels. The co-designing process begins when customers create new designs and submit their original content to a central hub (such as a company website). Next, a network of interested customers evaluates these submissions and selects (typically via online voting) those they find most appealing. Based on these evaluations, the firm then decides which products it will produce and market. In contrast to tinkering, where co-creators have considerable latitude in terms of altering the base product, firms engaged in co-designing usually dictate the precise format that co-created contributions must follow. For example, contributions to Threadless must be submitted using a company-issued template, can contain only limited text, and are constrained to eight sanctioned colors. Due to these mandates, co-designing contributions are considerably more fixed and constrained compared to either collaborating or tinkering. In contrast, co-designing provides customers considerable autonomy in terms of the selection process. For example, the contributions that Threadless selects to print as new T-shirt designs are almost exclusively determined by ratings provided by its customers. However, at times, Threadless invites particular designers to submit designs, and thus bypasses its typical selection process. Consequently, it appears that co-designing entails a level of customer autonomy over content selection that falls somewhere between collaborating and tinkering. From the perspective of a firm, co-designing appears to offer several advantages. Most importantly, this approach should dramatically reduce a firm’s cost of developing its own original designs or creative content, as this function is largely outsourced to customers. In addition, because customers actively assist a firm by both contributing new content and selecting the content that should appear in future product releases, firms should reduce their cycle times and launch new products more quickly compared to traditional NPD processes. For example, Threadless typically introduces several new T-shirts each week. Moreover, in contrast to collaborating and tinkering,
96
Matthew S. O’Hern and Aric Rindfleisch
co-designing is an approach in which both highly skilled (design contributors) and lower-skilled (design selectors) customers can freely participate. Thus, by providing its broader customer base with a strong voice in NPD selection and a sense of collective identity, co-designing should allow a firm to attain higher levels of customer satisfaction and commitment (Hertel, Niedner, and Herrmann 2003). In addition, by subjecting designs to prelaunch evaluation by a large network of customers, co-designing should reduce the risk of product failure, avoid drastic price markdowns, and minimize inventory holding costs (Ogawa and Piller 2006). Although the benefits of co-designing are intriguing, there are also substantial challenges in implementing this approach. First, firms may encounter difficulty in terms of attracting a critical mass of designers large enough to ensure that they obtain a sufficient amount of high-quality content (Cook 2008). This challenge may be especially acute for firms whose competitors have already established co-designing initiatives. In addition, although customers may be initially intrigued with this approach, the novelty of being able to submit and vote on designs may quickly wear off. Hence, a firm may find its pool of evaluators shrinking over time. Moreover, because this approach is relatively easy to imitate, firms that base their value proposition primarily on codesigning may end up lacking a distinctive core competence as competitors copy their approaches. To combat these challenges, firms that employ co-designing should establish strong lines of mutual communication with their co-designers and devote substantial effort toward fostering a collective sense of community (Cook 2008). Submitting We define submitting as a process in which customers directly communicate ideas for new product offerings to a firm. Submitting is differentiated from traditional forms of customer inquiry (for example, focus groups, satisfaction surveys, tracking studies, and so forth) by both the degree of customer effort required and by the nature of the input that customers provide to the firm. In contrast to most traditional forms of customer inquiry, which simply ask customers to provide responses to a set of prearranged queries, submitting requires customers to expend considerable energy developing (either in isolation or as part of a team) tangible ideas for new product offerings. In addition, while traditional inquiry approaches typically involve customers solely in concept ideation and evaluation, submitting often requires customers to translate general ideas into welldefined processes, detailed graphic depictions, or working new product prototypes. As shown in Figure 4.2, we conceptualize submitting as the form of co-creation that is characterized by the least amount of customer autonomy in terms of both NPD contribution and selection. Although submitting resembles co-designing (that is, both types of co-creation allow customers to directly contribute their own novel ideas and solutions), it differs from co-designing because in submitting, the firm retains full control over the NPD selection process. Firms that employ submitting-based co-creation actively solicit input from either current or potential customers. This solicitation often (but not always) occurs in the form of online invitations for customer-generated content. For example, the Swedish appliance manufacturer Electrolux sponsors an annual submitting competition called “Designlab” in which participants are asked to submit technical designs and product prototypes for cutting-edge household appliances. This initiative attracts thousands of entries across dozens of countries. From these, Electrolux selects a small set of finalists and invites them to a six-day, company-sponsored retreat, where they participate in workshops, present their inventions, and compete for cash prizes (www.electrolux. com/designlab). The Italian motorcycle manufacturer Ducati Motors employed a similar approach via its recent “Design Your Dream Ducati” contest that encouraged Ducati enthusiasts to submit
CUSTOMER CO-CREATION: A Typology and Research Agenda
97
innovative artistic and technical ideas to an executive team, which then selected the winning contributions (Sawhney, Verona, and Prandelli 2005). In addition to these firm-sponsored initiatives, submitting can also be brokered by third parties. For example, InnoCentive (a spinoff of Eli Lilly) is a brokering agent that links firms seeking solutions to complex scientific problems to a network of thousands of individual scientists. Firms anonymously post their NPD problems on the InnoCentive website, and highly trained specialists from around the world can submit solutions. Successful InnoCentive submitters receive financial compensation that can total tens or even hundreds of thousands of dollars, and their contributions have led to innovative breakthroughs for a variety of major corporations, including Procter and Gamble, Boeing, and DuPont. As described in the above examples, submitting begins when customers contribute detailed new product ideas, solutions, or prototypes. Based on these inputs, a firm then decides which concepts to further develop, test, and eventually launch. Compared to the three prior types of co-creation, submitting represents the lowest level of customer empowerment (as the firm dictates the format that contributions must follow and also has full power to select which customer contributions to adopt). In addition, many firms seek to retain control by insisting that submitters release their legal rights to the product improvements they help develop (Wells 2005). Compared to more traditional forms of customer input, however, submitting provides consumers with a much stronger voice in the NPD process and allows them to share their knowledge and creative skills directly with firm-based NPD teams. Firms may derive several benefits from submitting. First, case study evidence from large firms such as Intuit and Procter and Gamble suggests that this approach can result in a significant reduction in the time required to develop a new product and an increase in its degree of innovativeness (Cook 2008; Huston and Sakkab 2006). Also, because submitting offers customer benefits (for example, enhanced self-image and increased social status) largely absent from more traditional customer input methods, this form of co-creation should allow firms to engage in richer dialogues with customers who would normally be unresponsive to more traditional research inquiries. Thus, this engagement should result in improved market-sensing capabilities and enhanced customer relationships. Like other co-creation approaches, submitting also entails a number of challenges. Compared to the three other forms of co-creation in our framework, submitting may be least likely to result in truly innovative products because of its minimal level of customer empowerment. Due to these conditions, submitters may feel less connected with both the firm and other customers compared to collaborators, tinkerers, or co-designers, and may lack sufficient intrinsic motivation to actively cooperate with the firm on an ongoing basis. Consequently, firms interested in using this approach may experience difficulty in retaining active customer participation. Hence, it is important that firms duly recognize the contributions of submitters (with, for example, financial rewards, words of praise, explicit recognition). Without this type of reciprocity, it is possible that submitters may feel exploited and come to view submitting as a one-sided exchange (Fournier, Dobscha, and Mick 1998) rather than as a mutually satisfying bidirectional relationship (Oliver 2006). Perhaps more importantly, firms seeking to adopt this approach may find it quite difficult to continuously attract new contributors to their submitting initiatives. Because each customer may have only a limited number of new product solutions to offer, attracting new submitters may be even more important than retaining established contributors. Moreover, the successful retention of existing contributors should also enhance a firm’s ability to identify and recruit new submitters via positive word of mouth (Mathwick, Wiertz, and De Ruyter 2007).
98
Matthew S. O’Hern and Aric Rindfleisch
Future Research Agenda As seen in the preceding typology, customer co-creation represents a dramatic departure from traditional NPD practice both in terms of how customers are viewed as well as the roles they play. In brief, firms that abide by the traditional paradigm seek to solve NPD’s information asymmetry problem by first researching consumer preferences and then using this information to develop new products in-house. In contrast, firms following the emerging co-creation paradigm seek to solve this problem by actively soliciting consumer contributions and incorporating selected contributions into their new product offerings. As evidenced by our many examples, a paradigm shift is currently under way. While the full ramifications of this transition for marketing thought and practice are not yet clear, they are likely to be quite substantial. In the words of Wind and Rangaswamy (2001, p. 20), the emergence of customer co-creation has the potential to alter “everything we take for granted in marketing.” In this final section, we outline an agenda for future research in this domain. This agenda focuses on the impact of customer co-creation on six distinct domains of inquiry: (1) organizational culture, (2) organizational learning, (3) organizational dynamics, (4) resources and capabilities, (5) customer valuation, and (6) brand communities. Organizational Culture Because co-creation is fundamentally different from traditional NPD practice, successful implementation of this new paradigm will likely require significant changes in organizational culture (Sawhney, Verona, and Prandelli 2005; Vargo and Lusch 2004; von Hippel 2005). According to Thomke and von Hippel (2002, p. 78), “Turning customers into innovators requires no less than a radical change in management mind-set.” Specifically, co-creation’s highly interactive nature may require firms to adopt an open culture in which their goals, activities, and processes are highly transparent and collaborative (Ogawa and Piller 2006; Prahalad and Ramaswamy 2004). In particular, firms seeking to harness the benefits of co-creation may find it necessary to relax control over their intellectual property. For example, in 2005, IBM took the radical step of declaring that it would not enforce hundreds of its software patents in an effort to stimulate open source innovation and increase the market size for its Linux-based servers (Lohr 2005). This type of openness should encourage co-creation activity and enable customers to contribute innovative solutions to help a firm meet its goals and objectives. However, many firms with traditional organizational cultures are quite reluctant to relax control over their intellectual property (Cook 2008; von Hippel 2005). These firms may be willing to entreat customer contribution but will likely seek to retain control over the selection of these contributions. Thus, the degree to which the benefits (such as increased new product creativity, decreased time to market, and reduced development costs) of customer co-creation depend upon releasing organizational control over each of these two key processes (contribution and selection) is an important issue for future research. Along with increased openness, firms seeking to reap the benefits of customer co-creation will also likely need to adopt a more emergent strategic perspective (Mintzberg 1994). According to Jeppesen and Molin (2003, p. 377), under co-creation, “The management issue is not to enforce ideas, but to make room for them to emerge and channel them into an innovation.” This diminished focus on planning, forecasting, and control runs directly counter to the well-planned logic of the traditional NPD paradigm and, thus, is likely to meet with considerable resistance from managers who strongly believe in a more traditional approach. Indeed, their resistance may be well founded, as the substitution of improvisation for planning can be potentially harmful to NPD success (Moorman and Miner 1998). Of the four types of co-creation identified in our
CUSTOMER CO-CREATION: A Typology and Research Agenda
99
typology, co-designing appears to strike the best balance between improvisation and planning, as it encourages active customer participation within defined constraints. Future research is needed to determine the effectiveness of planned forms of co-creation such as co-designing and submitting versus more emergent forms of co-creation such as collaborating and tinkering. Organizational Learning Research on the drivers of innovation success has a long tradition in marketing scholarship (see Henard and Szymanski 2001, and Hauser, Tellis, and Griffin 2006 for reviews). Historically, the bulk of this research has focused on internal (that is, firm-based) drivers of NPD success, such as a firm’s degree of marketing and technological skills, level of market orientation, and amount of cross-functional integration. In recent years, this research has been supplemented by a growing number of studies that examine drivers of successful innovation beyond the immediate firm, such as the influence of acquired entities or alliance partners (see, for example, Chandy, Prabhu, and Ellis 2005; Rindfleisch and Moorman 2001; Sividas and Dwyer 2000). This research broadly suggests that the success of these collaborative NPD efforts strongly depends upon the degree to which a firm is able to acquire, assimilate, and apply information and know-how from its partners. In sum, collaborative NPD has been largely viewed from the perspective of a firm’s ability to learn. Customer co-creation adds a new dimension to this emerging literature by suggesting that NPD success strongly depends not only on a firm’s ability to learn but also on its ability to teach. Specifically, successful co-creation rests heavily upon the degree to which a firm is able to enhance its customers’ NPD-related knowledge and skills via such actions as allowing access to its source code, providing toolkits that allow customers to directly alter their products in creative ways, and engaging in direct, two-way communication with co-creators. This education imperative should be most important for collaborating and tinkering, as these co-creation approaches typically require a high level of technical skill. However, co-creation education may also be valuable in enhancing the quality of the contributions of customers engaged in co-designing and submitting. The importance of customer education has received some attention within the emerging literature on customer toolkits (such as Franke and Piller 2004; Prügl and Schreier 2006; Thomke and von Hippel 2002). However, several intriguing questions remain unanswered. For example, what is the relative value of educating existing customers versus recruiting new customers who already possess co-creating skills? Likewise, what types of customer education efforts are most valuable for each of the four types of customer co-creation identified in our typology? Organizational Dynamics The sense of empowerment that co-creators enjoy, combined with the increased knowledge and skills they are likely to acquire, may give customers a strong sense of psychological ownership over their contributions (Pitt et al. 2006; Prahalad and Ramaswamy 2000). In turn, this sense of ownership may complicate a firm’s internal NPD plans and activities. Because co-creation shifts tasks that were formerly conducted by managers down to customers, it may blur the boundary between these two groups (Evans and Wolf 2005; Prahalad and Ramaswamy 2000). Thus, customers may begin to see themselves not only as consumers but also as producers. Research on organizational dynamics (that is, stakeholder theory) suggests that firms that are heavily engaged in customer co-creation could find it more difficult to quickly alter their product lines or radically change their NPD processes in order to respond to competitive pressures (Donaldson and Preston
100
Matthew S. O’Hern and Aric Rindfleisch
1995). In essence, co-creation initiatives could create a new class of organizational stakeholders, many of whom may vigorously oppose NPD-related changes with which they disagree. These challenges are likely to be especially pronounced for those forms of co-creation that involve a high degree of customer-led contribution activity (tinkering and collaborating). However, anecdotal evidence suggests they can also occur in co-designing settings. For example, Threadless is currently facing a sense of ownership struggle, as several members of its online community are upset about its selection procedures and the recent opening of a brick-and-mortar store in Chicago (Chafkin 2008). Future research is needed to shed light on both the positive and negative impact that empowered co-creator stakeholders have upon NPD performance. In particular, the ability of these stakeholders to effectively respond to market challenges such as the entry of a new competitor or the emergence of a discontinuous technological is a topic ripe for empirical investigation. Resources and Capabilities The ability of a firm to achieve and sustain a competitive advantage is widely believed to rest upon its resources and capabilities such as sensing market trends, developing strong customer relationships, and creating innovative new products (Day 1994; Hunt and Morgan 1995; Moorman and Slotegraaf 1999). These resources and capabilities have been broadly viewed as assets that are internal to a firm and reside in its organizational culture, operating procedures, and human resources. For example, Day (1994, p. 38) suggests that “Capabilities are deeply embodied within the fabric of the organization.” The emergence of customer co-creation suggests that this strict focus on internal-based resources and capabilities ignores an important source of potential competitive advantage: the knowledge and skills embodied in a firm’s customer base (Jeppesen and Molin 2003; Prahalad and Ramaswamy 2000; von Hippel 2005). As suggested by Prahalad and Ramaswamy (2000, p. 80), when firms adopt co-creation, “consumers become a new source of competence for the corporation.” In essence, firms that succeed in establishing co-creation can gain access to a rich external source of NPD-related resources and capabilities that can supplement their internal value creation ability. Thus, the emergence of customer co-creation suggests that marketing scholars should view a firm’s resources and capabilities from a broader network-based (embodied) perspective rather than focusing narrowly on its internal (embedded) assets. As with their internal counterparts, the value of co-creation–based capabilities is likely to depend upon the degree to which they are distinctive and non-imitable. These qualities may be especially difficult to attain via a sharing approach because this form of co-creation appears to have few barriers to entry. Future research could make an important contribution by identifying the degree to which our four types of cocreation possess these desired qualities. Customer Valuation Understanding and assessing customer value is currently an extremely important topic of marketing scholarship (Marketing Science Institute 2008). To date, research on this topic has primarily focused on identifying the characteristics of profitable customers based largely upon their purchase behavior over time (for example, Gupta, Lehmann, and Stuart 2004; Reinartz and Kumar 2003; Rust, Zeithaml, and Lemon 2001). This focus reflects marketing’s traditional belief that the transaction itself is the primary mechanism of value exchange (Vargo and Lusch 2004). According to this belief, firms are responsible for creating value, and customers reward this value by purchasing their products (Srivastava, Shervani, and Fahey 1998). However, this perspective
CUSTOMER CO-CREATION: A Typology and Research Agenda
101
seems inconsistent with the logic of co-creation, which suggests that customers are not only targets of a firm’s value proposition, but also active value creators in their own right (Prahalad and Ramaswamy 2004; Vargo and Lusch 2004). In sum, it seems as if the boundary between producers and consumers is clearly shifting. The emergence of customer co-creation is a prime example of these shifting boundaries, as consumers provide firms with value in the form of their purchase activity as well as their production activity. Thus, it is possible that a customer who is an infrequent and low-volume purchaser, but a highly active contributor or selector of new product ideas, may be one of the firm’s most valuable assets. Unfortunately, the worth of this customer would go unrecognized by existing customer valuation perspectives, which do not account for the productive aspects of customer behavior. Hence, the emergence of customer co-creation calls for updated and enhanced customer value metrics that adopt a broader view of consumers and the spectrum of value they bring to the firm. For example, new metrics that focus on assessing the value of a collaborator’s contributions using market-based criteria such as the number of times their contributions have been downloaded, viewed, or further modified by fellow collaborators would be an important refinement of the existing customer valuation frameworks. Brand Communities Historically, marketing scholars have focused on consumer behavior from the perspective of the individual. In recent years, however, researchers have highlighted the growing phenomenon of collective consumer behavior in the form of brand communities (McAlexander et al. 2001; Muñiz and O’Guinn 2001; Muñiz and Schau 2005). These communities bring together (either physically or virtually) individuals who share a common affinity for a particular brand such as Apple computers, Saab automobiles, or Harley Davidson motorcycles. Thus far, this literature has emphasized the potential of these communities as vehicles for forging deep and enduring customer-brand relationships. However, they may also serve as especially fertile ground for co-creation activity. Recent research by Muñiz and Schau touches on this potential by observing that members of the abandoned Apple Newton community develop applications for this product and actively share their creations with fellow members. Indeed, the basic features of brand communities (for example, a collection of dedicated and knowledgeable individuals who exchange information about their beliefs, interests, and insights) should provide a means of incubating customer co-creation. Moreover, because customers may engage in co-creation as contributors as well as selectors, community-based appeals may resonate with one kind of co-creator but not the other. In the case of Threadless, for example, customers who regularly engage in selecting T-shirt designs are perhaps more likely to feel a greater sense of community than those who contribute these designs, because this latter group may be more self-directed and extrinsically motivated by Threadless’s $2,000 cash reward and the chance to garner greater professional exposure. Thus, the role of brand communities as a catalyst for co-creation (and vice-versa) is an intriguing topic for future research. Conclusion In today’s highly competitive marketplace, a growing number of firms are placing increased reliance upon innovation as a means of achieving growth and profitability. Unfortunately, most new products fail because they do not adequately satisfy customer wants or needs. Thus, as a means of minimizing market failure and enhancing financial performance, an increasing number of firms are empowering customers and allowing them to actively participate in the NPD process. As we have
102
Matthew S. O’Hern and Aric Rindfleisch
shown, this customer empowerment can take a number of distinct forms. Our typology classifies four emerging forms of co-creation delineated by the degree to which customers are empowered to contribute and select new product ideas. Although each of these co-creation types has its own potential benefits and challenges, they all lie outside the boundary of NPD’s traditional worldview and, hence, are contributing to the emergence of an exciting new paradigm. Millions of empowered customers around the world have embraced the co-creation ethos. Thus, firms should look beyond the confines of their traditional NPD approaches and develop effective strategies for identifying and harnessing the ideas, skills, and talents of their customers. We hope this typology and research agenda fosters greater appreciation for and investigation into this important phenomenon. Notes 1. For ease of exposition, we employ the term “product.” However, we acknowledge that customer cocreation is highly congruent with marketing’s service-dominant logic, which posits that collaborating with customers creates value by harnessing the power of customer learning and leveraging the service-based benefits embedded in products (Vargo and Lusch 2004). 2. Although we make no claim that these creativity-based benefits supersede consumption-based benefits, the extant literature on customer co-creation nevertheless suggests that benefits that are based on the creative work of customers can play an important role in enhancing customer satisfaction (Lakhani and Wolf 2005; Shah 2006). 3. Some scholars also identify lead users as a form of co-creation (for example, Urban and von Hippel 1988). Our typology does not explicitly consider lead users as a specific form of co-creation unto itself. Instead, we highlight the role that lead users play in some of the types of co-creation identified in our typology. In addition, while a few scholars view mass customization as a form of co-creation (Wind and Rangaswamy 2001), a large number of scholars disagree and feel that mass customization does not sufficiently incorporate customer input into the actual creative process (for example, Jeppesen and Molin 2003; von Hippel 2005). Thus, our framework does not explicitly consider mass customization as a form of co-creation. 4. Many examples of lead-user alterations (such as the first homemade windsurfers developed by surfing enthusiasts) bear a resemblance to tinkering. Although these modifications may result in creative outcomes, firms rarely if ever assist end users in making these alterations (Franke and Shah 2003; Luethje, Herstatt, and von Hippel 2002). In contrast, firms engaged in tinkering actively encourage customers to alter their products, establish forums for tinkerers to share their creations, and specifically design their products to allow for easy customer modification. 5. It is estimated that approximately one-third of all computer games offer these types of toolkits to their users (Jepperson and Molin 2003). 6. Current TV was founded in 2005 by former vice president Al Gore. Its tagline (“The TV network created by the people who watch it”) nicely reflects its co-creation ethos.
References Amabile, Teresa M., Regina Conti, Heather Coon, Jeffrey Lazenby, and Michael Herron. 1996. “Assessing the Work Environment for Creativity.” Academy of Management Journal 39 (5), 1154–1184. Belk, Russell W. 1985. “Materialism: Trait Aspects of Living in the Material World.” Journal of Consumer Research 12 (December), 265–280. Bendapudi, Neeli, and Robert P. Leone. 2003. “Psychological Implications of Customer Participation in Co-Production.” Journal of Marketing 67 (April), 14–28. Carpenter, Gregory S., Rashi Glazer, and Kent Nakamoto. 1994. “Meaningful Brands from Meaningless Differentiation: The Dependence on Irrelevant Attributes.” Journal of Marketing Research 31 (August), 339–350. Chafkin, Max. 2008. “The Customer Is the Company.” Inc. 30 (June), 88–97. Chandy, Rajesh K., Jaideep Prabhu, and Mark Ellis. 2005. “The Impact of Acquisitions on Innovation: Poison Pill, Placebo, or Tonic?” Journal of Marketing 69 (January), 114–130. Christensen, Clayton M. 1997. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Boston: Harvard Business School Press.
CUSTOMER CO-CREATION: A Typology and Research Agenda
103
Cook, Scott. 2008. “The Contribution Revolution.” Harvard Business Review 86 (October), 60–69. Cooper, Robert G. 1990. “Stage-Gate Systems: A New Tool for Managing Products.” Business Horizons 33 (May-June), 44–53. Csikszentmihalyi, Mihalyi. 1996. Creativity: Flow and the Psychology of Discovery and Invention. New York: HarperCollins. Darke, Peter R., and Robin J.B. Ritchie. 2007. “The Defensive Consumer: Advertising Deception, Defensive Processing, and Distrust.” Journal of Marketing Research 44 (February), 114–127. Day, George S. 1994. “The Capabilities of Market-Driven Organizations.” Journal of Marketing 58 (October), 37–52. Deci, Edward L., and Richard M. Ryan. 1985. Intrinsic Motivation and Self-Determination in Human Behavior. New York: Plenum Press. Donaldson, Thomas, and Lee E. Preston. 1995. “The Stakeholder Theory of the Corporation: Concepts, Evidence, and Implications.” Academy of Management Review 20 (January), 65–91. Ebenkamp, Becky. 2005. “Scrapbooking: Crafters Create their Back Pages.” Brandweek 46 (24), 24. Etgar, Michael. 2008. “A Descriptive Model of the Consumer Co-Production Process.” Journal of the Academy of Marketing Science 36 (Spring), 97–108. Evans, Philip, and Bob Wolf. 2005. “Collaboration Rules.” Harvard Business Review 83 (July-August), 96–104. Firat, A. Fuat, Nikhilesh Dholakia, and Alladi Venkatesh. 1995. “Marketing in a Postmodern World.” European Journal of Marketing 29 (1), 40–56. Fournier, Susan, Susan Dobscha, and David Glen Mick. 1998. “Preventing the Premature Death of Relationship Marketing.” Harvard Business Review 76 (January-February), 42–51. Franke, Nikolaus, and Frank Piller. 2004. “Value Creation by Toolkits for User Innovation and Design: The Case of the Watch Market.” Journal of Product Innovation Management 21, 401–415. Franke, Nikolaus, and Sonali Shah. 2003. “How Communities Support Innovative Activities: An Exploration of Assistance and Sharing Among End-Users.” Research Policy 32 (1), 157–178. Füller, Johann, Michael Bartl, Holger Ernst, and Hans Mühlbacher. 2004. “Community Based Innovation: A Method to Utilize the Innovative Potential of Online Communities.” Proceedings of the 37th Annual Hawaii International Conference on System Sciences,Big Island, Hawaii: IEEE. Grewal, Rajdeep, Gary Lilien, and Girish Mallapragada. 2006. “Location, Location, Location: How Network Embeddedness Affects Project Success in Open Source Systems.” Management Science 52 (July), 1043–1056. Gupta, Sunil, Donald R. Lehmann, and Jennifer Stuart. 2004. “Valuing Customers.” Journal of Marketing Research 41 (February), 7–18. Handelman, Jay M. 2006. “Corporate Identity and the Societal Constituent.” Journal of the Academy of Marketing Science 34 (Spring), 107–114. Harris, Steven E. 2004. “Inside IBM: Dan Frye and the Linux Technology Center.” Available at www.consultingtimes.com/articles/ibm/frye. Hauser, John, Gerard J. Tellis, and Abbie Griffin. 2006. “Research on Innovation: A Review and Agenda for Marketing Science.” Marketing Science 25 (November-December), 686–717. Henard, David H., and David M. Szymanski. 2001. “Why Some New Products Are More Successful than Others.” Journal of Marketing Research 38 (August), 362–375. Hertel, Guido, Sven Niedner, and Stefanie Herrmann. 2003. “Motivation of Software Developers in Open Source Projects: An Internet-Based Survey of Contributors to the Linux Kernel.” Research Policy 32 (July), 1159–1177. Hunt, Shelby D., and Robert M. Morgan. 1995. “The Comparative Advantage Theory of Competition.” Journal of Marketing 59 (April), 1–15. Huston, Larry, and Nabil Sakkab. 2006. “Connect and Develop: Inside Procter & Gamble’s New Model for Innovation.” Harvard Business Review 81 (March), 58–66. Jeppesen, Lars Bo, and Mans J. Molin. 2003. “Consumers as Co-developers: Learning and Innovation Outside the Firm.” Technology Analysis & Strategic Management 15 (September), 262–283. Kahn, Kenneth. 2005. The PDMA Handbook of New Product Development, 2nd ed. New York: Wiley. Kalaignanam, Kartik, and Rajan Varadarajan. 2006. “Customers as Co-Producers: Implications for Marketing Strategy Effectiveness and Marketing Operations Efficiency.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 166–179. Armonk, NY: M.E. Sharpe.
104
Matthew S. O’Hern and Aric Rindfleisch
Kasser, Tim. 2003. The High Price of Materialism. Cambridge, MA: MIT Press. Klein, Jill Gabrielle, N. Craig Smith, and Andrew John. 2004. “Why We Boycott: Consumer Motivation for Boycott Participation.” Journal of Marketing 68 (July), 92–109. Kohn, Alfie. 1999. “In Pursuit of Affluence: At a High Price.” New York Times, February 2, D2. Kozinets, Robert V., and Jay M. Handelman. 2004. “Adversaries of Consumption: Consumer Movements, Activism, and Ideology.” Journal of Consumer Research 31 (December), 691–704. Lakhani, Karim R., and Robert Wolf. 2005. “Why Hackers Do What They Do: Understanding Motivation and Effort in Free/Open Source Software Projects.” In Perspectives on Free and Open Source Software, ed. J. Feller, B. Fitzgerald, S. Hissam, and K. Lakhani, 3–22. Cambridge, MA: MIT Press. Leadbeter, Charles, and Paul Miller. 2004. The Pro-Am Revolution. London: Demos. Lohr, Steve. 2005. “Sharing the Wealth at IBM.” New York Times, April 11, C1. Luethje, Christian, Cornelius Herstatt, and Eric von Hippel. 2005. “User Innovators and ‘Local’ Information: The Case of Mountain Biking.” Research Policy 34 (August), 951–965. Marketing Science Institute. 2008. 2008–2010 Research Priorities. Cambridge, MA: Marketing Science Institute. Mathwick, Charla, Caroline Wiertz, and Ko De Ruyter. 2007. “Social Capital Production in a Virtual P3 Community.” Journal of Consumer Research 34 (April), 832–849. McAlexander, James H., John W. Schouten, and Harold Koenig. 2002. “Bulding Brand Community.” Journal of Marketing. 66 (January), 38–54. Mintzberg, Henry. 1994. The Rise and Fall of Strategic Planning. New York: Free Press. Moon, Jae Yun, and Lee Sproull. 2001. “Turning Love into Money: How Some Firms May Profit from Voluntary Electronic Customer Communities.” Working Paper, Stern School of Business, New York University, New York. Moorman, Christine, and Anne S. Miner. 1998. “The Convergence of Planning and Execution: Improvisation in New Product Development.” Journal of Marketing 62 (July), 1–20. Moorman, Christine, and Rebecca J. Slotegraaf. 1999. “The Contingency Value of Complementary Capabilities in Product Development” Journal of Marketing Research 36 (May), 239–257. Muñiz, Albert M., Jr., and Thomas C. O’Guinn. 2001. “Brand Community.” Journal of Consumer Research 27 (March), 412–432. Muñiz, Albert M., Jr., and Hope Jensen Schau. 2005. “Religiosity in the Abandoned Apple Newton Brand Community.” Journal of Consumer Research 31 (March), 737–747. Nieborg, David B. 2005. “Am I a Mod or Not? An Analysis of First Person Shooter Modification Culture.” Available at www.gamespace.nl/content/DBNieborg2005_CreativeGamers.pdf. Obermiller, Carl, and Eric Spangenberg. 1998. “Development of a Scale to Measure Consumer Skepticism Toward Advertising.” Journal of Consumer Psychology 7 (2), 159–186. Ogawa, Susumu, and Frank T. Piller. 2006. “Reducing the Risks of New Product Development.” Sloan Management Review 47 (Winter), 65–72. Oliver, Richard L. 2006. “Co-Producers and Co-Participants in the Satisfaction Process: Mutually Satisfying Consumption.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 166–179. Armonk, NY: M.E. Sharpe. Pietrykowski, Bruce. 2003. “You Are What You Eat: The Social Economy of the Slow Food Movement.” Review of Social Economy 62 (3), 307–309. Pitt, Leyland F., Richard T. Watson, Pierre Berthon, Donald Wynn, and George Zinkhan. 2006. “The Penguin’s Window: Corporate Brands from an Open Source Perspective.” Journal of the Academy of Marketing Science 34 (Spring), 115–127. Prahalad, C.K., and Venkat Ramaswamy. 2000. “Co-opting Customer Competence.” Harvard Business Review 78 (January-February), 79–87. ———. 2004. “Co-Creation Experiences: The Next Practice in Value Creation.” Journal of Interactive Marketing 18 (Summer), 5–14. Prügl, Reinhard, and Martin Schreier. 2006. “Learning from Leading-Edge Customers at The Sims: Opening up the Innovation Process Using Toolkits.” R&D Management 36 (June), 237–250. Rajagopalan, Balaji, and Barry L. Bayus. 2008. “Exploring the Open Source Software Bazaar.” Review of Marketing Research, vol. 5, ed. Naresh K. Malhotra. Armonk, NY: M.E. Sharpe, forthcoming, 58–74. Randall, Taylor, Christian Terweisch, and Karl T. Ulrich. 2005. “Principles for User Design of Customized Products.” California Management Review 47 (Summer), 68–85.
CUSTOMER CO-CREATION: A Typology and Research Agenda
105
Rangaswamy, Arvind, and Gary L. Lilien. 1997. “Software Tools for New Product Development.” Journal of Marketing Research 34 (February), 177–184. Raymond, Eric S. 1999. The Cathedral & the Bazaar. North Sebastopol, CA: O’Reilly. Reinartz, Werner, and V. Kumar. 2003. “The Impact of Customer Relationship Characteristics on Profitable Lifetime Duration.” Journal of Marketing 67 (January), 77–99. Richins, Marsha L., and Scott Dawson. 1992. “A Consumer Values Orientation for Materialism and Its Measurement.” Journal of Consumer Research 19 (December), 303–316. Rindfleisch, Aric, and Christine Moorman. 2001. “The Acquisition and Utilization of Information in New Product Alliances: A Strength-of-Ties Perspective.” Journal of Marketing 65 (April), 1–18. Roberts, Deborah, Susan Baker, and David Walker. 2005. “Can We Learn Together? Co-Creating with Consumers.” International Journal of Market Research 47 (4), 407–427. Rust, Roland T., Valarie Zeithaml, and Katherine N. Lemon. 2001. Driving Customer Equity: How Customer Lifetime Value Is Reshaping Corporate Strategy. New York: Free Press. Sawhney, Mohanbir, Gianmario Verona, and Emanuela Prandelli. 2005. “Collaborating to Compete: The Internet as a Platform for Customer Engagement in Product Innovation.” Journal of Interactive Marketing 19 (Autumn), 4–17. Schor, Juliet B. 1998. The Overspent American: Upscaling, Downshifting, and the New Consumer. New York: Basic Books. Sethi, Rajesh, Daniel C. Smith, and C. Whan Park. 2001. “The Effect of Cross-Functional Teams on Creativity and the Innovativeness of New Products.” Journal of Marketing Research 37 (February), 73–85. Seybold, Patricia. 2006. Outside Innovation: How Your Customers Will Co-Design Your Company’s Future. New York: Collins. Shah, Sonali. 2006. “Motivation, Governance and the Viability of Hybrid Forms in Open Source Software.” Management Science 52 (July), 1000–1014. Sharma, Srinarayan, Vijayan Sugumaran, and Balaji Rajagopalan. 2002. “A Framework for Creating Hybrid– Open Source Software Communities.” Information Systems Journal 12 (January), 7–25. Simonson, Itamar. 2005. “Determinants of Customers’ Responses to Customized Offers: Conceptual Framework and Research Propositions.” Journal of Marketing 69 (January), 32–45. Sivadas, Eugene, and F. Robert Dwyer. 2000. “An Examination of Organizational Factors Influencing New Product Success in Internal and Alliance-Based Process.” Journal of Marketing 64 (January), 31–49. Song, X. Michael, and Mark E. Parry. 1997. “A Cross-National Comparative Study of New Product Development Processes: Japan and the United States.” Journal of Marketing 61 (April), 1–18. Sorescu, Alina, Rajesh Chandy, and Jaideep Prabhu. 2003. “Sources and Financial Consequences of Radical Innovations: Insights from Pharmaceuticals.” Journal of Marketing 67 (October), 82–102. Srivastava, Rajendra K., Tasadduq A. Shervani, and Liam Fahey. 1998. “Market-Based Assets and Shareholder Value: Framework for Analysis.” Journal of Marketing 62 (January), 2–18. Thomke, Stefan, and Eric von Hippel. 2002. “Customers as Innovators: A New Way to Create Value.” Harvard Business Review 80 (April), 74–81. Thompson, Craig J., Aric Rindfleisch, and Zeynep Arsel. 2006. “Emotional Branding and the Strategic Value of the Döppelganger Brand Image.” Journal of Marketing 70 (January), 50–64. Urban, Glen L., and Eric von Hippel. 1988. “Lead User Analyses for the Development of New Industrial Products.” Management Science 34 (May), 569–582. Vargo, Stephen L., and Robert F. Lusch. 2004. “Evolving to a New Dominant Logic for Marketing.” Journal of Marketing 68 (January), 1–17. Velthouse, Betty A. 1990. “Creativity and Empowerment: A Complementary Relationship.” Review of Business 12, 13–18. Vogelstein, Fred. 2008. “The Mozilla CEO on His Firefox Strategy, His Google Gambit, and Working with Apple.” Wired, 16 (August). Available at www.wired.com/techbiz/people/magazine/16–08/ ff_lilly. von Hippel, Eric. 2005. Democratizing Innovation. Cambridge, MA: MIT Press. von Hippel, Eric, and Ralph Katz. 2002. “Shifting Innovation to Users via Toolkits.” Management Science 48 (7), 821–833. von Krogh, Georg, Sebastain Spaeth, and Karim R. Lakhani. 2003. “Community, Joining, and Specialization in Open Source Software Innovation: A Case Study.” Research Policy 32 (7), 1217–1241. Wells, Melanie. 2005. “Have It Your Way.” Forbes, February 14, 78–86.
106
Matthew S. O’Hern and Aric Rindfleisch
Wind, Jerry, and Vijay Mahajan. 1997. “Issues and Opportunities in New Product Development.” Journal of Marketing Research 34 (February), 1–12. Wind, Jerry, and Arvind Rangaswamy. 2001. “Customerization: The Next Revolution in Mass Customization.” Journal of Interactive Marketing 15 (Winter), 13–22. Woodman, Richard W., John E. Sawyer, and Ricky W. Griffin. 1993. “Toward a Theory of Organizational Creativity.” Academy of Management Review 18 (2), 293–322. Wright, Peter. 2002. “Marketplace Metacognition and Social Intelligence.” Journal of Consumer Research 28 (March), 677–682.
Chapter 5
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT Combining Research and Practice Perspectives Koen Pauwels and Dave Reibstein
Abstract Return on Marketing Investment (ROMI) holds great promise as a metric to increase the accountability for marketing spending. Unfortunately, many organizations experience considerable roadblocks to fulfilling the appealing promise of measuring ROMI and using it to enable better marketing decisions and higher performance. We discuss ten such roadblocks, give examples, and critically examine how research has addressed and should further address these issues. Introduction Return on Marketing Investment (ROMI) is defined as the incremental margin generated by a marketing program, divided by the cost of that program at a given risk level (Powell 2002). The typical formula is displayed in equation 1:
Return on Marketing Investment = [Incremental Margin – Marketing Investment] / Marketing Investment
(1)
Use of this metric promotes accountability for marketing spending, enables comparison across alternatives to decide on the best action, and furthers organizational learning and cross-functional teamwork. Unfortunately, managers are struggling to define and calculate ROMI (Woods 2004), especially outside the price/promotions domain (Bucklin and Gupta 1999). A survey of over one thousand C-level managers (CMO Council 2004) revealed that over 90 percent of marketing executives viewed marketing performance metrics as a significant priority, but that over 80 percent were unhappy with their current ability to measure performance. Only 17 percent of marketing executives have a comprehensive system to measure marketing performance. The companies they work for outperformed other firms in revenue growth, market share, and profitability. Thus, most organizations experience considerable roadblocks to fulfilling the appealing promise of measuring ROMI and using it to enable better marketing decisions and higher performance. Since most financial decisions in the non-marketing domain are typically based on their return on investment, the 107
108
Koen Pauwels and Dave Reibstein
absence of ROI calculations for marketing activities makes it harder to obtain funding for them. Several reasons underlie these difficulties, from the improper use of the term “return on investment” for measures that do not include profits/margins or investment costs (Lenskold 2003) to the lack of research into how return on marketing investment can be measured and how it can be used to enhance performance (Pauwels et al. 2008). Indeed, while many marketing practitioners and academics have expressed concern about marketing accountability and return on investment, the current push has come largely from outside the field, notably top management and finance (Lehmann and Reibstein 2006). Unfortunately, CEOs and CFOs have been disappointed by the most common responses of the marketing field, from “it is hard to judge the impact of marketing spending since so many factors come into play between the spending and the ultimate financial result” (marketing practice) to “we already show it through our sales response functions” (marketing academia). The authors’ experience in recent years demonstrates that such positions are of little help in bridging the gap between marketing and finance fields, enabling joint understanding and trust in ROMI calculations and ROMI-based decisions and building the standing of marketing in the C-suite. Previous authors have already laid the conceptual frameworks for return on marketing investment (Lehmann 2005; Lehmann and Reibstein 2006; Rust et al. 2004; Sheth and Sisodia 2002; Srivastava, Shercvani, and Fahey 1998). True to the focus of Review of Marketing Research on “implementing new marketing research concepts and procedures,” this chapter discusses ten conceptual and implementation issues that complicate measurement and use of return on marketing investment. First, the “incremental margin” in equation 1 (hereafter “return”) needs to be forecast, in terms of magnitude but also timing and associated risk. Second, the investment could involve a combination of marketing actions and must be considered from the point of decision perspective. Once the components of returns and investment are measured, it is still unclear whether they should be combined for a focus on impact versus efficiency and realized versus potential return on marketing investment. Finally, acting upon measured return on marketing investment requires clarity on how to weigh multiple objectives and an understanding of whether high ROMI means the marketing action should get more or less investment in the future. Often, spending more on programs with a high ROI will lower the ROI percentage but raise the total return, given that we are generally at the diminishing-returns stage of the response curve. Figure 5.1 presents our framework for organizing these issues, while Table 5.1 summarizes what we already know and what we still need to learn from further research. The remainder of this chapter discusses all ten challenges in detail, giving examples and critically examining how research has addressed and should further address these issues. Ten Challenges to Measure and Act upon Return on Marketing Investment Challenge 1: ROI Framework Devised for Predictable Timing of Returns Issue Return on Investment was devised for comparing capital projects (such as building a larger factory) in which an investment is made once and the returns flow predictably during the following predictable years. In contrast, spending on a marketing campaign may extend across multiple periods and result in the building of brand and/or the value of customer assets with a less predictable duration. These are the two biggest assets marketing brings to the firm. The effects of brand and customer asset building will entail wear-in and wear-out patterns that are hard to measure,
Decision perspective
When?
RETURN ON INVESTMENT
Synergy?
Risk?
Competition?
Intervening factors?
Realized vs. Potential?
Impact vs. Efficiency?
Figure 5.1 Framework: Ten Challenges to Measure and Act upon Return on Marketing Investment
Invest more or less?
Multiple objectives?
109
Sales effects of promotions, advertising, and new products Wear-in and wear-out patterns in mature (Western) markets
Estimate uncertainty, discount factor, finance portfolio theory Look forward to decide among projects, backward to evaluate
Timing of returns
Risk of returns
Nature and size of competitive response to promotions Internal decision rules dominate net impact of marketing actions
Include in model if available Scenario building and decisions trees if data are not available NPV, DCR, and EVA focus on impact; ROI focuses on efficiency
Reaction of market players
Intervening factors
Realized ROMI, next-product-to buy, share of requirements
Incorporate in objective function Optimal spending with concave and S-shaped response functions
Realized vs. potential ROMI
Multiple objectives Invest more or less in marketing
Impact vs. efficiency
Advertising synergy matters Need for integrated marketing strategies
Marketing synergy
Point of decision perspective
What we know now
Challenges
Relation with marketing actions How to merge different data sets with different periodicity How to achieve both and when to focus on impact vs. efficiency Implications for budgeting Calculate underinvestment ratio Identify new opportunities Optimal weighting Weigh efficiency versus growth Experimentation
ROMI impact of response Response to strategic actions Information value of investor reaction for marketing decisions
Effects on marketing assets How and when marketing assets affect sales, margins, and risk Managerial uncertainty; how to include it in marketing budgets Divisibility of marketing actions How much to experiment? Experiments to decouple collinearity and to estimate marketing interactions
Sales effects of distribution
What we still need to know
Summary of Challenges, What We Know Now, and What We Still Need to Know
Table 5.1
110
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
111
especially when tracking the impact of several activities. As a result, the mapping of the return for any one marketing investment will be very difficult to establish. Even other marketing investments that are not necessarily directed at brand or at attracting customers will have varying degrees of longevity and unpredictable impact on the firm’s return. Example An advertising campaign may involve spending the money in the first month, seeing the execution in the second and third month, observing initial sales effects in the third month and the peak sales impact in the fifth month. The same campaign may also have increased the willingness to pay for the brand, which does not show up immediately, as the company does not raise prices. One year later, though, the company is able to raise the brand’s price by 3 percent and lose only 2 percent of unit sales, versus 4 percent of unit sales that would have been lost without the ad campaign. Current Research While Little (1970) pointed to the possibility of wear-in times for marketing campaigns, empirical evidence remains limited to sales effects of advertising, new product introductions, and point-ofpurchase actions. The peak sales effect of advertising occurs relatively fast, typically within two months (Pauwels 2004; Tellis 2004), while the wear-in times for mindset metrics (such as awareness, liking, and consideration) are just over two months (Srinivasan, Vanheule, and Pauwels 2010). In contrast, new product introductions typically take several months or years to take off (Golder and Tellis 1997). As can be expected, point-of-purchase actions work right away or not at all (Pauwels 2004), with price promotions standing out as the most studied marketing action (Srinivasan et al. 2004). A major gap in the literature is the timing of improvements in distribution, with only Srinivasan et al. (2010) reporting it takes an average of 2.1 months for increases in distribution coverage to reach their highest impact—the longest wear-in time of all studied marketing actions. Further investigation of this issue is important because distribution stands out as the most effective marketing action (Bronnenberg, Mahajan, and Vanhonacker 2000; Srinivasan et al. 2010). Finally, we know very little of the timing of returns to investments in new (electronic) media such as paid search, banner ads, and word-of-mouth referrals. On the latter, Trusov, Bucklin, and Pauwels (2009) report that wear-out times are substantially higher for word-of-mouth referrals as compared to traditional marketing actions for a social networking site. Moreover, Drèze and Bonfrer (2008) calculate the timing of e-mail returns to optimize e-mail frequency and maximize customer equity. While considerable work has been done on brand and the value of customers, there has been little research that reflects how spending directly adds to the brand’s value and when the payout will be realized. More efforts are being made to track the connection between spending and customer acquisition and the value they bring to the firm (Rust et al. 2004). Even here it is often difficult to allocate the proportion of spending that contributes to customers’ acquisition versus retention. Priorities for Further Research Empirical generalizations on wear-in and wear-out effects are necessary for managerial advice in cases where data are (for now) missing (Lehmann 2005). As literature has focused on certain marketing actions in U.S. markets, we are in need of studies analyzing return timing for investments in new media and new (emerging) markets. Moreover, the timing of returns may systematically vary by medium and target audience, which should be taken into consideration when deciding
112
Koen Pauwels and Dave Reibstein
between campaigns. Considerable research still is required to determine the contribution of marketing spending on a brand’s value and when this value is realized by the firm. Challenge 2: How to Adjust Projected Returns for Risk? Issue While marketing is all about change and risk, marketing managers are reluctant to calculate riskadjusted ROI due to a combination of fear of statistics and the perceived need to be a cheerleader for marketing actions (Marketing NPV 2007). Further, we have few measures of risks. Marketers often look at the projected return from alternative programs and select the ones that have the overall highest projected return. This treats each program as having the same levels of risk in their projections. Different levels of risk for the same expected return should lead risk-averse managers to select the action with the lowest risk. In addition, the kind of assets created by certain programs may reduce the vulnerability of the company in the future (Fornell et al. 2006; Srinivasan and Hanssens 2009). A typical example is a marketing program that builds customer loyalty. As the loyalty gets larger, the risk levels are reduced. Hence, the programs could be assessed based on their short-term ROMI or on the resulting reduction in risk for subsequent sales. Example A firm has a history of running price promotions every fall. Their years of experience have shown there will be a 10 percent lift in sales, resulting in a 15 percent ROMI. A manager suggests the alternative of running a banner ad on several Internet sites, which are preferred by other firms claiming sales lifts ranging from 2–30 percent with an average of 15 percent and an expected ROMI of 20 percent, and a range of –5 to 40 percent. If the firm would invest strictly in the program with the highest return, it would clearly select the banner ads with the higher expected return. However, many risk-averse managers would instead prefer to continue running the price promotion. The picture further changes if the proposed new media increases brand equity not captured in the ROMI calculation. As the brand grows in value, it stabilizes future sales. Current Research In academic articles, risk assessment typically involves reporting standard errors around model estimates and forecast error bands, sometimes noting the caveat that the future is expected to be like the past. Practitioner-oriented articles (e.g., Dhar and Glazer 2003; MarketingNPV 2007) and books that speak to both academic and practitioner audiences (e.g., Blattberg, Kim, and Neslin 2008; Jagpal 1999, chapter 6) go beyond these statements to consider how risk should be incorporated in marketing decision making. The discount rate applied to returns is key in this respect and can be determined based on the opportunity cost of capital and as well as on the source of risk. The former approach holds that the firm should use as a discount rate the rate of return an investor could make on a project of similar risk, calculated as the weighted average cost of capital (WACC) (Brealey, Meyers, and Marcus 2004). This number differs per company, depending on their betas (the ratio of the firm’s return variability over the variability of the stock market) and the firm’s amounts of debt and equity. For instance, using these data from the 1997–2002 period, Amazon had a WACC of 16.45 percent, while Pfizer had one of only 6.88 percent (ibid.). Because it is calculated for the whole company, the WACC is the appropriate discount rate when the firm
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
113
is considering investing in a marketing campaign similar to its past campaigns. For a very different campaign, managers should identify the source of risk (that is, which component has a higher risk level than that of typical projects), which requires them to consider the reasons for atypically high risk (and ways to mitigate them). Finally, investing in a portfolio of marketing campaigns and customer segments may reduce risk through diversification, and modern portfolio theory may help firms balance expected returns and variance (Sharpe 2000). Priorities for Further Research Several of the recent advances in finance (as noted above) have yet to be applied and adapted to marketing investments. Identifying sources of risk is enabled with tools such as Strengths-Weaknesses-Opportunities-Threats (SWOT) analyses, process flows, and contributing factor diagrams (MarketingNPV 2007). Investing in portfolios of marketing campaigns and customer segments is a prime area for future research. For instance, the correlation among returns from different segments should be inspired by academic research on how consumers interact across segments (with, for example, one segment acting as opinion leader for the others). Challenge 3: Point of Decision Perspective Needs to Be Captured in ROMI Approach Issue ROI was devised for comparing capital projects in which a given level of investment is made once. Each new adjustment in spending requires a new ROI assessment of the expected returns for that investment. The expected returns are adjusted as new information comes in. Similarly, most marketing investments (for example, ad weight) are divisible (that is, we can spend half of it now, half of it later) and can be revisited as new information comes in. How should companies deal with such potentially continuous updating of ROI assessments? Example from Lenskold 2003 A company launches a new product and is deciding on the marketing support campaign. With a cost of $200K ($100K for market research, $100K for media buying), the marketing campaign is expected to generate $600K in extra revenues at 50 percent contribution margin. The company calculates a ROMI of [$300K–$200K]/$200K = 50 percent and decides to undertake the campaign. After having spent the $100K on market research, research shows that the campaign will generate only $360K in extra revenues. Should the company spend the remaining $100K? A retrospective calculation would now reveal a ROMI of [$180K – $200K]/$200K = –10 percent. However, a forward-looking calculation would consider the $100K spent on market research as a sunk cost and thus reveal a ROMI of [$180K – $100K]/$100K = 80 percent and imply the campaign should go on. Which one is correct? Current Research Marketing academic research is largely silent on this issue. From decision-making disciplines, we know that the forward-looking calculation is right in guiding which course of action to take, as it takes the correct point of decision perspective. However, the retrospective ROMI calculation is
114
Koen Pauwels and Dave Reibstein
insightful for the purpose of tracking the precision of revenue forecasts and addressing specific shortcomings in this regard. Priorities for Further Research It is important to distinguish forward-looking ROMI for making specific decisions from ROI of the entire marketing budget (or even marketing department), which includes maintenance marketing and other costs not attributable to specific campaigns. Moreover, further research is needed into the true divisibility of strategic marketing investments to answer tough questions that CMOs face regarding the required level of investment. Finally, it appears worthwhile to regularly experiment with new marketing ideas using small investments in limited settings and then to scale up the successful actions quickly to the rest of the organization (Eechambadi 2005). Challenge 4: Synergy in Marketing Spending Issue Here the issue is one of isolating the impact of any one element of the marketing program. Often, the goal is trying to identify which components of the program are working and producing the greatest return on the marketing investment. As such, it would be possible to eliminate the components that are less efficient. However, the components are often interdependent, which occurs in two different forms: multi-collinearity and interactions. Multi-collinearity is where two or more variables or programs are run in conjunction with each other, and as such it is difficult to assess which of the programs is really causing the result: is it one, the other, or really only when done in conjunction. If it is really just one of the programs that is causing the effect, yet it is hard to assess which is the driver of the result, this is not really synergy, but it poses the same problem of assessing which is the critical expenditure that is yielding the result. The case of interaction is really one of synergy. The two components combined have a much greater impact than the one alone. In this situation, it is difficult to assess the impact of a single program, as it has an impact beyond its direct effect. The entire notion of an “integrated marketing program” is built on the premise that there is a synergy among the various different marketing activities. Example Once a year, Bloomingdale’s runs a special price promotion for its credit card holders. It stays open late for an exclusive shopping evening at reduced prices. To do this without letting their credit card holders know of the event would be foolish, so the store sends out several special mail pieces to all of their credit card holders. The net result is a huge boost in sales for that evening. It is difficult for the store to tell if the sales hike is the result of the later than usual hours, the reduction in price, or the special mail pieces that were sent out. It could be merely the result of multi-collinearity, and because these three activities always happen in conjunction, it is impossible to assess which is the real causal factor for the sales spike. Current Research Integrated Marketing Communications enjoys a long history in practitioner-oriented literature, but relatively few academic papers have demonstrated marketing synergy. Those that do typically
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
115
find large synergy effects (Naik and Raman 2003; Srinivasan et al. 2009). For one, advertising is nine times more effective when paired with a new-to-the world car introduction versus paired with an existing car model (Srinivasan et al. 2009). Priorities for Further Research The above-cited papers provide models to demonstrate and quantify synergy based on past data. However, perfect collinearity among marketing actions in these past data (as in our example) prohibits a model from distinguishing such effects. In that case, the best way to assess the impact of the individual components of the marketing program would be to run an experiment or test market. Taking the Bloomingdale’s example above, it would be possible to try an evening with extended hours and test the impact. Similar isolation of the effects could be run for the other two components of the program. In this manner it would be possible to test the effect of each as well as to assess the synergistic impact. Challenge 5: ROMI Depends on the Reaction of Competitors and Other Market Players Issue The return on marketing investment is influenced by the reaction of competitors, other market players (such as retailers), and by employees of the company itself (for example, through decision rules that favor repeating past successes (Dekimpe and Hanssens 1999). As a result, marketing managers are urged to consider the net long-term impact of their decisions, which includes dynamic response of such market players (Chen 1996; Dekimpe and Hanssens 1999; Day and Reibstein 1997; Jedidi, Mela, and Gupta 1999). This makes it very difficult to know what the return will be on any proposed marketing spending prior to knowing all important reactions. To complicate matters, managers often have little insight into their own company’s decision rules (for example, supporting a new product with advertising and price cuts) or its inertia (an initially successful action gets prolonged and/or repeated). Marketing literature has so far focused on estimating customer and competitor response to marketing actions, instead of the response of other players in the market system. Examples The absence of a significant post-promotion sales dip in several empirical studies is mostly due to the fact that prices do not return to their regular levels for several weeks (Pauwels 2004, 2007; Srinivasan et al. 2004). A plausible reason for such prolonged company action, as confirmed in experiments, is the managerial tendency to weigh past prices when setting future prices (Krishna, Mela, and Urbany 2000; Nijs, Srinivasan, and Pauwels 2007). Second, advertising may fail to affect sales due to its inability to generate consumer response for established brands (Abraham and Lodish 1990), or due to competitive retaliation campaigns that cancel any demand gain (Bass and Pilon 1980). Current Research Research is abundant as to competitive reaction, including its nature (aggressive, accommodating, or neutral), its speed, and absence due to competitor’s unawareness or inability to react (Chen 1996).
116
Koen Pauwels and Dave Reibstein
Recent studies have begun to assess the importance of competitive reaction on the return on the initial marketing investment and report small effects on average, including both positive and negative impact (Steenkamp et al. 2005). Pauwels (2007) finds that competitor reaction lowers price promotion benefits by 10 percent. Only a few papers have focused on company decision rules, and they have consistently found that their impact on the net performance effects of price promotions dominates that of competitive reaction (Horváth, Leeflang, and Wittink 2001; Pauwels 2004, 2007). Similarly, the reaction of retailers is most often analyzed in the context of price promotions (pass-through) and new product introductions (retailer adoption and shelf space allocation). As for other market players, recent research has focused on analyzing investor reaction to product innovation, price reductions, and advertising (e.g., Pauwels et al. 2004; Joshi and Hanssens 2004; Srinivasan and Hanssens 2009). Priorities for Further Research The reaction of market players can be assessed by dynamic system modeling in data-rich environments (e.g., Pauwels 2004) and by role play in data-scarce environments, such as one-shot negotiations (Armstrong 2001). Marketing researchers have become fascinated with the reaction of financial markets to marketing actions, as evident from the attention of the Marketing Science Institute and the Journal of Marketing in a forthcoming special issue on marketing and finance. Further research is needed to assess whether investors react appropriately to marketing actions and thus how valuable the information of investor reaction is for marketing decision making. A key research priority is to go beyond documenting a reaction and toward understanding the impact of that reaction on the return on investment of the initiating action. For marketing mix actions, is it really the case (Pauwels 2004) that the majority of the net sales impact derives not from customer reaction but from support from other marketing actions (for advertising and new products) or from prolonging the initial action for several weeks (in the case of feature and price cuts)? For strategic marketing actions, how does one assess likely competitive reaction in deciding on location, product quality, and regular price level (for example, start or avoid starting a price war)? Challenge 6: Intervening Variables Mask the True Impact of Marketing for the Firm Issue Several factors intervene between marketing actions and when the customer ultimately buys the product, as well as the true impact on the value of the firm. Supply disruptions (especially critical in emerging economies) can lower the financial returns to marketing actions, even when sufficient consumer demand was generated. Macro-economic changes, such as the recent credit crunch and financial meltdown, can wreak havoc on carefully planned and executed campaigns. Consumer trends concerning diet habits can change as a result of specific media exposure. Such intervening factors appear especially important for products with a long sales cycle, as considerable time passes from awareness to interest to consideration to preference to purchase, assessment, and ultimately repurchase. Moreover, information on such intervening factors typically comes from non-marketing datasets (such as government or industry sources), raising the issue of how to merge different data sets with different periodicity. As a result of intervening factors, it becomes difficult to directly assess the impact of the marketing actions and spending on the value of the firm, even if there is a highly positive impact created by the spending.
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
117
Example As in the example in Challenge 1 above, advertising spending occurs in the first month. In the second month, the economy takes a downturn, and operations has trouble shipping the product in a timely manner. The question is, what would have been the resulting sales had these other factors not occurred? Sales may even be down, yet in order to assess the return on the marketing investment, the real question is, how much more would the sales have been down had there been no marketing effort? Current Research When the intervening factors are known and measurable, researchers prefer to include them as variables in their models. For instance, Pauwels and his coauthors (2004) included the S&P 500 index, the construction cost index, and the dollar–yen exchange rate in their analysis of the U.S. performance impact of product innovation for Japanese car manufacturers. If historical data are unavailable, managers can assess the likelihood of intervening factors and their impact on the return on marketing investment, for instance through scenario building (Armstrong 2001; MarketingNPV 2007). Combining such managerial judgment with estimates from past data offers a promising way to get the best out of model and manager (Blattberg and Hoch 1990). Priorities for Further Research A key research goal is to assess the impact of marketing, separating out all of the extraneous and intervening factors. If there are varying regions or time periods in which there were no or at least different intervening factors, by running a cross-sectional or time series analysis and including the intervening variables as other variables in the model or as co-variants, it should be possible to isolate the marketing effects. This, of course, assumes two things—the intervening factors are independent of the marketing efforts, and there are sufficient observations of these factors. Unfortunately, often these two conditions do not exist. For example, when we run a promotion or an advertisement, if sales boom, then running out of stock is directly correlated with the promotion or marketing. This would clearly understate the true potential impact of the marketing effort. As for the number of observations, many times the extraneous factors are episodic and are one-time events, such as a labor strike, and would be hard to build into a model with such limited observations. The other alternative is to run an experiment or test market. In a tightly controlled experiment in a limited geographic area, it would be possible to avoid such intervening factors. Here the challenges are twofold: 1) running the experiment for a sufficient time period for measuring the long-term effects, and 2) avoiding the intervening variables that result from the expenditures, such as advertising and stock-outs. This way, the internal validity would be high in showing the effect of marketing on performance in the absence of intervening factors. Of course, such experiment or test market would have limited external validity: in the real world, managers do have to anticipate and account for key intervening forces. Challenge 7: Impact vs. Efficiency of Marketing Spending Issue Even with consensus on measuring the return and investment components of marketing spending, it is unclear whether they should be combined to measure effective spending versus efficient spending. Normally, when we refer to return on marketing dollar or ROI, it is in reference to efficiency.
118
Koen Pauwels and Dave Reibstein
With efficiency as the goal, it almost always results in a reduction in budget, as the way to increase the efficiency of the total marketing spending is to eliminate those programs that are less efficient. Instead, the firm may be more interested in the effectiveness (impact) of a marketing action, which may be better expressed as [return – investment], without dividing by the investment as indicated by the ROI formula. As an illustration, compare two mutually exclusive projects, with returns of respectively $100M and $10M and investment costs of respectively $80M and $2M, at the same level of risk. The first project has the larger net return [$20M > $8M], but the second project has the larger return on investment [25 percent < 400 percent]. Which project should a manager prefer? Example IBM once had a supplies division, in which it offered a variety of services. In order to maximize their efficiency, they starting looking at each service and eliminating them from the portfolio. Not surprisingly, the size of the division continued to shrink until they decided they did not have critical mass to warrant the division. So, while it was efficient, they lacked enough effectiveness in what they were providing to keep the business. Current Research This issue has not been much debated in academic articles, but more so in books and practitioner-oriented publications. Lenskold (2003) argues that the goal should be efficiency, that is, to maximize the return per dollar spent. In contrast, Ambler (2003) argues that impact (effectiveness) is more important: it is most closely related to Net Present Value, Discounting Cash Flow, and Economic Value Added metrics and focuses the organization on maximizing long-term firm value instead of short-term efficiency. Priorities for Further Research Research could analyze the conditions under which an organization would and should focus on impact versus efficiency goals. For instance, growing, cash-rich firms in boom times favor impact goals, while cash-strapped (small) firms in recession times favor efficiency goals. In general, though, there should be a balance between effectiveness and efficiency goals. For instance, Diageo displays marketing actions on a 2 by 2 matrix juxtaposing their impact (on defined objectives) and their efficiency (ROMI). Actions without sufficient impact are likely to be canceled, no matter how high their ROMI, while impactful but inefficient actions are reexamined to improve efficiency in the future. Probably there should be some threshold, perhaps the cost of capital, that marketing spending should exceed, and all programs that exceed this threshold should be supported. Research could investigate what these thresholds for impact and efficiency should be. As this chapter indicates, measuring the effectiveness or the efficiency is not an easy task. The point is that it is not only important to measure the percentage return of any spending amount but also the magnitude. The goal should be to maximize the total impact once a certain threshold is achieved, even if that reduces the overall efficiency. Challenge 8: Realized ROMI vs. Potential ROMI Issue While sophisticated companies like Procter and Gamble have a good feel for the return on investment of their actions (“realized ROMI”), they have yet to find how much return they could have
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
119
had (“potential ROMI”). This is the spirit behind the Pioneering Research for an In-Store Metric (PRISM) project: by tracking aisle-specific traffic and conversion (or “closing rate”), companies get a better feel for how much extra revenues they could have obtained. Indeed, most companies do not invest in promising marketing actions until diminishing returns set in: history (non-zero budgeting, conservative bias) and modesty prevent this. Example When focus groups revealed that its target customers loved ice cream, Unilever wanted to fill store freezer doors with ads for its new Sunsilk hair care product (Neff 2008). However, PRISM data revealed that only 10 percent of the target market (young women age 18–24) actually go down the ice cream aisle during a shopping trip. Instead, end-aisle and hair care shelf ads reached a much larger portion of the target group. In contrast to ice cream, some categories obtain high aisle traffic but low conversion. Butter, yogurt, cereal, and coffee are examples. How can we explain such patterns, and how can companies capture some of this potential? Current Research Research on marketing potential has focused on customer relationship management, that is, identifying how specific customers can be upgraded to higher usage and retention (e.g., Neslin et al. 2006). Using only firm-specific records, next-product-to-buy models attempt to infer the current share of requirements and thus quantify the potential for increasing revenues from a customer (Knott, Hayes, and Neslin 2002). Du, Kamakura, and Mela (2007) augment these internal records with insights into customers’ relationships with competing firms to estimate the size of each customer’s wallet and the firm’s share of it. Finally, hidden Markov models have been applied to estimate household life cycles and their impact on budgetary allocation (Du and Kamakura 2006). Priorities for Further Research Marketing could gain a “better seat” in the boardroom if CEOs and CFOs understood how much money they left on the table by under-investing in marketing and marketing skills (such as optimal pricing). Based on models of customer requirements and the firm share of wallet (static focus) or life cycle and next-product-to-buy models (dynamic focus), researchers could calculate “underinvestment quotients” to quantify this. Companies can then set up a system at the highest level that includes metrics allowing experiments and scaling them up. For example, provided with better in-store metrics, researchers can analyze which vehicles have the biggest impact on aisle traffic and shopper conversion to capture the full potential of that aisle traffic. Challenge 9: How to Deal with the Multiple Objectives? Issue Besides profits, managers at different levels of the organization care about a multitude of objectives, including stock price/market capitalization, sales volume, the share of a specific market, market share, brand equity and other customer mindset metrics, retailer equity, and so forth. How do these objectives relate to each other? Should they be incorporated into an “integrative” ROMI calculation, and how much weight should be given to different objectives?
120
Koen Pauwels and Dave Reibstein
Example Natter and his coauthors (2007) optimized dynamic pricing and promotion planning for a retailing company, having agreed to optimize profits. When they recommended higher prices to increase company revenues, they met with substantial resistance from the purchasing managers, whose supplier discounts depend on sales volume, and from local branch managers, who insisted on keeping a market share leadership position in their city. After further discussion, they decided to combine profits, total sales volume, and local market share objectives in an overall goal function for the model to optimize. Current Research With the exception of the work of Natter and his colleagues (2007), we know of little academic research on this issue. Forrester research reports and academic work in progress does focus on the measurement of new media, such as user-generated content and blogs. A big question here is what the objectives of the organization should be. Priorities for Further Research Bridging market perspectives across functional and geographical boundaries is an important objective of marketing in general (e.g., Jaworski and Kohli 1993) and marketing dashboards in particular (Pauwels et al. 2008). Eliciting these opinions and furthering consensus are underdeveloped areas in research and practice. Moreover, research could investigate the “optimal” weighting of objectives based on hard performance measures, similar to combining model and managerial judgment (Blattberg and Hoch 1990). Challenge 10: Invest More or Less in High ROMI Actions? Issue If a firm is able to assess the return on their marketing dollars, how should they use this information for future budgeting decisions? It is not as simple as shifting dollars to the actions with the highest past ROMI. Just because this is how the dollars produced in the past does not mean this is how they will in the future. Further, that a certain level of return was obtained does not mean this is the optimal level; that is, perhaps better performance could be achieved if a different level of spending or different implementation was employed. Example As Harrah’s implemented their dashboard and determined the efficiency of their marketing spending, they reallocated their dollars to spend only on more productive programs. One might speculate that with greater productivity around their marketing spending, they would be more likely to spend more. That was not the case, however (Reibstein et al., 2005). At least, they did not decrease their marketing spending, but rather felt they could get more done with the same overall marketing budget. Current Research Empirical generalizations on sales response functions provide some guidance as to optimal spending rules (Mantrala, and Zoltners 1992).
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
121
Figure 5.2 Sales Response Function
Sales
Marketing Stock
Imagine a firm with a sales response function as shown in Figure 5.2. What is shown is the sales response resulting from a certain level of marketing “stock.” This “stock” reflects the impact of current spending plus whatever impact there is from the previous years’ spending. In other words, advertising’s impact accumulates. Part of the “stock” that has been built by past spending erodes each year and is added to with each new level of spending. A couple of issues arise: 1. The implication of stock is if the same amount of money was spent in the same way this year as last, the results will probably be different, even if all other conditions were the same, since the firm would be at a different level of “stock.” 2. Looking at the “X” on the graph, what is clear is the firm could have had a higher return had it spent more, although at a lower level of ROMI, since it is past the second inflection point. Priorities for Further Research If a firm would measure its response function and knew where it was on that function, it would be in a better position to know whether to be spending more or less than in previous years. The alternative would be to run experiments to assess alternative levels of expenditure and different programs and their resulting impact. Beyond the ten challenges from the framework in Figure 5.1, multinational companies face several additional issues, such as global versus local branding and ROMI measurement. For instance, Samsung has successfully revamped their entire marketing effort to reflect a global optimization of resources, while Avaya has standardized ROMI measurement globally but encourages local branding and marketing actions. Conclusion It is critical to measure the return on marketing investments. In order to get budgeting support, marketing will find it necessary to make the case that investments into marketing programs warrant the expenditure, relative to other opportunities facing the organization.
122
Koen Pauwels and Dave Reibstein
Unfortunately, the estimate of ROMI is faced with many challenges, as outlined above. Perhaps most critical is the time delay between when the marketing dollars are spent and the actual and residual results. Making the direct association between marketing spending and sales results is all the more difficult because of the numerous other factors that change due to the marketing investment. Sales are reflected in the number of units sold, but there also is the impact of the marketing spending on the margins that are commanded. Are the incremental unit sales or the added margin the result of the marketing spending, or are they coming from the research and development dollars that enhanced the quality of the product or service? Did the R&D spending get its insights and direction from the marketing research that was done and was honed to the customers’ needs during the new-product development process? These factors and others cited above make the problem of estimating ROMI exceedingly difficult. That said, it should not be an excuse for estimating the return. Comparable issues face other investment opportunities facing the organization. When one invests in plant and equipment, one is never fully certain how long the equipment will actually last and how productive the new facilities will be, or even when they will be completed. Yet estimates are made. When investing in financial instruments or foreign currencies, an estimate is made of the likely return, but this is always done with uncertainty. The same holds for estimating the return of marketing dollars. In this chapter we have specified the types of research that could be done to help reduce some of the uncertainties and help in ROMI estimation. These steps will not eliminate the uncertainty but should help in estimating the likely return. In this way, marketing expenditures can be compared with the other choices facing the organization. Acknowledgments We thank Sunil Gupta, Dominique Hanssens, Scott Neslin, Shuba Srinivasan, and Russ Winer for most excellent comments on a previous version. All remaining errors are our own. References Abraham, Magid, and Leonard M. Lodish. 1990. “Getting the Most Out of Advertising and Promotion.” Harvard Business Review 68 (3) (May-June), 50–60. Ambler, Tim. 2003. Marketing and the Bottom Line, 2nd ed. Financial Times Prentice Hall. Armstrong, Scott, ed. 2001. Principles of Forecasting: A Handbook for Researchers and Practitioners. Boston, MA: Kluwer Academic Publishers. Bass, Frank M., and T.L. Pilon. 1980. “A Stochastic Brand Choice Framework for Econometric Modeling of Time Series Market Share Behavior.” Journal of Marketing Research 17, 486–497. Blattberg, Robert C., and Stephen J. Hoch. 1990. “Database Models and Managerial Intuition: 50% Model + 50% Manager.” Management Science 36 (8), 887–899. Blattberg, Robert C., Byung-Do Kim, and Scott Neslin. 2008. Database Marketing: Analyzing and Managing Customers. International Series in Quantitative Marketing. Springer. Brealey, R.A., S.C. Meyers, and A.J. Marcus. 2004. Fundamentals of Corporate Finance. New York: McGraw-Hill/Irwin. Bronnenberg, Bart J., Vijay Mahajan, and Wilfred Vanhonacker. 2000. “The Emergence of Market Structure in New Repeat-purchase Categories: A Dynamic Approach and an Empirical Application.” Journal of Marketing Research 37 (1), 16–31. Bucklin, Randolph E., and Sunil Gupta. 1999. “Commercial Use of UPC Scanner Data: Industry and Academic Perspectives.” Marketing Science 18 (3), 247–273. Chen, Ming-Jer. 1996. “Competitor Analysis and Interfirm Rivalry: Toward a Theoretical Integration.” Academy of Management Review 21 (1) 100–134. CMO: Chief Marketing Officers Council’s Marketing Measures Performance Audit. 2004. White Paper. Day, George, and David Reibstein, eds. 1997. Wharton on Dynamic Competitive Strategy. New York: John Wiley & Sons.
CHALLENGES IN MEASURING RETURN ON MARKETING INVESTMENT
123
Dekimpe, Marnik, and Dominique Hanssens. 1999. “Sustained Spending and Persistent Response: A New Look at Long-Term Marketing Profitability.” Journal of Marketing Research 36 (November), 397–412. Dhar, R., and R. Glazer 2003. “Hedging Customers.” Harvard Business Review 81 (5), 86–92. Du, Rex, and Wagner Kamakura. 2006. “Household Life Cycles and Lifestyles in the United States.” Journal of Marketing Research 43 (1), 121–132. Du, Rex, Wagner Kamakura, and Carl Mela. 2007. “Size and Share of Customer Wallet.” Journal of Marketing 71(2), 94–113. Drèze, Xavier, and Andre Bonfrer. 2008. “An Empirical Investigation of the Impact of Communication Timing on Customer Equity”, Journal of Interactive Marketing, 22 (1), 36–50. Eechambadi, Naras. 2005. High Performance Marketing: Bringing Method to the Madness of Marketing. Kaplan Business. Fornell, Claes, Sunil Mithas, Forrest V. Morgeson III, and M.S. Krishnan. 2006. “Customer Satisfaction and Stock Prices: High Returns, Low Risk.” Journal of Marketing 70 (January), 3–14. Golder, Peter N., and Gerald J. Tellis. 1997. “Will It Ever Fly: Modeling the Takeoff of Really New Consumer Durables.” Marketing Science 16 (3), 256–270. Horváth, Csilla, Peter S.H. Leeflang, and Dick R. Wittink. 2001. “Dynamic Analysis of a Competitive Marketing System.” Working Paper, Yale School of Management. Jagpal, Sharan. 1999. Marketing Strategy and Uncertainty. New York: Oxford University Press. Jaworski, B., and A. Kohli. 1993. “Market Orientation: Antecedents and Consequences.” Journal of Marketing 57, 53–70. Jedidi, Kamel, Carl F. Mela, and Sunil Gupta. 1999. “Managing Advertising and Promotion for Long-Run Profitability.” Marketing Science 18 (1), 1–22. Joshi, Amit, and Dominique Hanssens. 2004. “Advertising Spending and Market Capitalization.” Marketing Science Institute, Report 04–110. Knott, Aaron, Andrew Hayes, and Scott Neslin “Next-Product-to-Buy Models for Cross-Selling Applications.” Journal of Interactive Marketing 16 (3), 2002. Krishna, Aradhna, Carl Mela, and Joel Urbany. 2000. “Inertia in Pricing.” Working paper, University of Notre Dame, Indiana. Lehmann, Donald R. 2005. “The Metrics Imperative: Making Marketing Matter.” In Review of Marketing Research, vol. 2, ed. Naresh K. Malhotra, 177–202. Armonk, NY: M.E. Sharpe. Lehmann, Donald R., and David J. Reibstein. 2006. Marketing Metrics and Financial Performance. Marketing Science Institute monograph. Lenskold, James D. 2003. Marketing ROI. New York: McGraw-Hill. Little, John D.C. 1970. “Models and Managers: The Concept of a Decision Calculus.” Management Science 16 (8), Application Series (April), B466–B485. Mantrala, Sinha, and Zoltners, 1992. “Impact of Resource Allocation Rules on Marketing Investment-Level Decisions and Profitability.” Journal of Marketing Research 29 (2), 162–175. MarketingNPV. 2007. “How Much Unnecessary Risk Is in Your Marketing Plan?” 1 (1), 4–8. Naik, Prasad, and Kalyan Raman. 2003. “Understanding the Impact of Synergy in Multimedia Communications.” Journal of Marketing Research 40 (4), 375–388. Natter, Martin, Thomas Reutterer, Andreas Mild, and Alfred Taudes. 2007. “An Assortmentwide DecisionSupport System for Dynamic Pricing and Promotion Planning in DIY Retailing.” Marketing Science 26 (4), 576–583. Naik, Prasad, and Kalyan Raman. 2003. “Understanding the Impact of Synergy in Multimedia Communications.” Journal of Marketing Research 40 (4), 375–388. Neff, Jack. 2008. “Seeing Shoppers Through a Prism.” Research World 1 (September), 84–85. Neslin, Scott, Dhruv Grewal, Robert Leghorn, Venkatesh Shankar, Marije L. Teerling, Jacquelyn Thomas, and Peter C. Verhoef. 2006. “Challenges and Opportunities in Multichannel Customer Management.” Journal of Service Research 9 (2), 95–112. Nijs, Vincent, Shuba Srinivasan, and Koen Pauwels. 2007. “Retail-Price Drivers and Retailer Profits.” Marketing Science 26 (4), 473–487. Pauwels, Koen. 2004. “How Dynamic Consumer Response, Competitor Response, Company Support and Company Inertia Shape Long-Term Marketing Effectiveness.” Marketing Science 23 (4), 596–610. ———. 2007. “How Retailer and Competitor Decisions Drive the Long-Term Effectiveness of Manufacturer Promotions for Fast Moving Consumer Goods.” Journal of Retailing 83 (2).
124
Koen Pauwels and Dave Reibstein
Pauwels, Koen, Tim Ambler, Bruce Clark, Pat LaPointe, David Reibstein, Bernd Skiera, Berend Wierenga, and Thorsten Wiesel. 2008. “Dashboards and Marketing: Why, What, How and What Research Is Needed?” Marketing Science Institute, Special Report 08–203. Pauwels, Koen H., Jorge M. Silva-Risso, Shuba Srinivasan, and Dominique M. Hanssens. 2004. “New Products, Sales Promotions and Firm Value, with Application to the Automobile Industry.” Journal of Marketing 68 (October), 142–156. Powell, Guy. 2002. Return on Marketing Investment. RPI Press. Reibstein, David, David Norton, Yogesh Joshi, and Paul Farris. 2005. “Marketing Dashboards: A Decision Support System for Assessing Marketing Productivity.” Presentation at the Marketing Science Conference, Atlanta, Georgia. Rust, Roland.T., Tim Ambler, Greg Carpenter, V. Kumar, and Raj Srivastava. 2004. “Measuring Marketing Productivity: Current Knowledge and Future Directions.” Journal of Marketing 68 (4), 76–89. Sharpe, W.F. 2000. Portfolio Theory and Capital Markets. New York: McGraw-Hill. Sheth, Jagdish N., and Rajendra S. Sisodia. 2002. “Marketing Productivity: Issues and Analysis.” Journal of Business Research 55 (5), 349–362. Srinivasan, Shuba, and Dominique Hanssens. 2009. “Marketing and Firm Value: Metrics, Methods, Findings, and Future Directions.” Journal of Marketing Research, 46(3), 293–312. Srinivasan, Shuba, Koen Pauwels, Dominique Hanssens, and Marnik Dekimpe. 2004. “Do Promotions Benefit Retailers, Manufacturers, or Both?” Management Science 50 (5), 617–629. Srinivasan, Shuba, Koen Pauwels, J. Silva-Risso, and Dominique Hanssens. 2009. “Product Innovations, Marketing Investments and Stock Returns.” Journal of Marketing forthcoming, January. Srinivasan, Shuba, Marc Vanheule, and Koen Pauwels. 2010. “Customer Mindset Metrics and Marketing Mix Models: An Integrative Approach.” Journal of Marketing Research, forthcoming. Srivastava, Rajendra, Tasadduq A. Shercvani, and Liam Fahey. 1998. “Market-Based Assets and Shareholder Value: A Framework for Analysis.” Journal of Marketing (January), 1–18. Steenkamp, Jan-Benedict, Vincent Nijs, Dominique Hanssens, and Marnik Dekimpe. 2005. “Competitive Reactions to Advertising and Promotion Attacks.” Marketing Science 24 (1), 35–54. Tellis, Gerard J. 2004. Effective Advertising: Understanding When, How, and Why Advertising Works. Thousand Oaks, CA: Sage. Trusov, Michael, Randy Bucklin, and Koen Pauwels. 2008. “Estimating Online Word-of-Mouth Elasticity.” Working Paper, UCLA. ———. 2009. “Effects of Word of Mouth versus Traditional Marketing: Finding for an Internet Social Networking Site,” Journal of Marketing, forthcoming October 2009. Woods, Julie A. 2004. “Communication ROI.” Communication World (January/February), 14.
Chapter 6
Service-Dominant Logic A Review and Assessment Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Abstract The emergence and evolution of service-dominant (S-D) logic (Vargo and Lusch 2004a) has drawn increasing attention toward the integration of resources, especially intangible and dynamic resources and interdependent processes that drive the creation of value. The core of this developing mindset fundamentally shifts the focus of marketing and, more generally, business away from the production and distribution of goods (goods-dominant logic) toward service, the application of operant resources (knowledge and skills), as the basis of exchange. S-D logic’s advancement is driven by the collaboration and contributions of marketing and non-marketing-related (and other) disciplines. Currently, evidence of this joint effort can be found in a variety of journal special issues and conference special sessions, a number of journal articles and other presentations, several dedicated conferences, and one book with fifty contributing scholars. This review consolidates the S-D logic writings of Vargo, Lusch, and their coauthors, as well as the related work (and viewpoints) of other scholars, to examine the implications of the S-D logic mindset for marketing. It (1) explores the need for S-D logic and summarizes its current state of development, (2) provides an S-D logic perspective of the market and marketing, (3) clarifies major theoretical misconceptions, (4) reviews the extension of S-D logic and its integration with existing marketing knowledge , (5) provides an assessment of the role of S-D logic in the evolution of academic marketing, and (6) offers directions for future research. Introduction What has become known as service-dominant (S-D) logic was introduced by Stephen Vargo and Robert Lusch (2004a) in a Journal of Marketing (JM) article titled, “Evolving to a New Dominant Logic for Marketing.” However, in a very real sense, its beginning was much earlier and more deeply rooted in marketing and marketing-associated literature. S-D logic is intended to capture and extend a convergence of apparently diverse thought that has been shifting the dominant logic of marketing and economic thought away from a primary concern with tangible resources, output in the form of firm-created value (goods), and transactions. It points toward a revised logic for 125
126
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
marketing based on the application of often intangible, dynamic resources, inputs for co-created value, and relational economic and social processes. As the title of the original article implies, S-D logic represents an evolution, rather than a revolution, in marketing thought. The central tenet of S-D logic is that reciprocal service, defined as the application of competences for the benefit of another party, is the fundamental basis of economic exchange. That is, service is exchanged for service. The discussion and debate about S-D logic was initiated by the simultaneous publication of “Evolving . . .” (Vargo and Lusch 2004a) and the commentaries of seven renowned marketing scholars invited by Ruth Bolton, the editor of JM at that time. In her introduction to the commentaries, Bolton (Bolton et al. 2004, p. 18) observed that “the new dominant logic has important implications for marketing theory, practice, and pedagogy, as well as for general management and public policy. . . . The ideas expressed in the article and commentaries will undoubtedly provoke a variety of reactions.” It appears her foresight was correct. The original article and commentaries have stimulated a lively and ongoing worldwide discussion about the nature of the market and marketing within the discipline and beyond. Interest in and evidence of this continuing discussion—as well as evidence of the spirit of collaboration and knowledge generation in which S-D logic is grounded—can be seen in the number and variety of publications, special issues, conferences, and conference special sessions with an S-D logic focus that have followed the publication of the original article. The “Otago Forum on Service-Dominant Logic” (organized by David Ballantyne) was held in November 2005 to discuss and further codevelop S-D logic and its challenge to marketing’s dominant, goods-centered paradigm and was followed by “Otago Forum II” in December 2008 to focus on moving the theoretical foundations of S-D logic into marketing and marketing-related practice. Both forums were tied to special issues of journals—Marketing Theory and Industrial Marketing Management, respectively. About the time of the first Otago forum, The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions (Lusch and Vargo 2006a) was published. This book brought together insights from fifty top marketing scholars from around the world and represented various viewpoints and positions in relation to S-D logic. Since the 2004 article, there have also been special sessions on S-D logic at the American Marketing Association (AMA) summer and/or winter conference every year to date (2004–2008). Additionally, S-D logic has been the focus of special sessions at the 2004 Academy of Marketing Science (AMS) Cross-Cultural Conference held in Mexico, the 2007 AMS World Marketing Congress in Italy, the 2005 European Marketing Academy Conference (EMAC), the 2005 Australia and New Zealand Marketing Academy Conference (ANZMAC), and the 2008 Global Marketing Conference in Shanghai (coordinated by the Korean Academy of Marketing Science), as well as a joint ANZMAC/EMAC symposium in 2007 that resulted in a special issue of the Australasian Marketing Journal (2007). Also in 2008, an S-D logic Doctoral Colloquium was held in conjunction with the Logic and Science of Service (also known as the Art and Science of Service), which focused on S-D logic–related topics. Moreover, S-D logic is the central topic of a special issue of the Journal of Academy of Marketing Science (2008), which received approximately seventy submissions. Further evidence of the impact of and continuing dialogue about S-D logic can be found in a Google search for “Vargo and Lusch” on the World Wide Web, which reveals hundreds of forums, journal publications, conference presentations, books and book chapters, and marketing course syllabi, as well as weblogs (blogs), and other websites. Much of the feedback surrounding S-D logic has been favorable, and many scholars have embraced the opportunity to discuss the evolution of marketing and how their work contributes to this developing mindset. However, as is often the case with dramatic shifts in thought, some
Service-Dominant Logic: A Review and Assessment
127
hesitation, misunderstanding, and skepticism have emerged as well. These diverse viewpoints enrich the discussion. The purpose of this chapter is to review and consolidate the S-D logic writing of Vargo and Lusch,1 to review the S-D logic work (and viewpoints) of others, to examine the implications of the S-D logic mindset, and to point toward directions for future research. First, an overview of the need for and development of a revised worldview of marketing and the foundational bases of S-D logic are provided. Next, S-D logic’s unique market and marketing perspectives are offered. The theoretical integration and conceptual expansion of a service-centered perspective of marketing are then examined, with special attention to the clarification of common misunderstandings. The major elaborations and extensions of S-D logic that have been contributed by marketing-related scholars are then explored. Next, an assessment of the role of S-D logic in the evolution of marketing knowledge is provided. We conclude with a discussion of the major research frontiers and opportunities that S-D logic offers. S-D Logic: An Alternative Worldview A dominant worldview is rarely clearly stated or specifically promoted; rather it permeates into the mindsets of researchers and practitioners through individual beliefs and collective paradigms. Kuhn (1970, p. 10) defines paradigms as “accepted examples of actual scientific practice that provide models from which spring particular coherent traditions of scientific research.” Drawing on Kuhn, Arndt (1985, p. 11) views paradigms as “social constructs reflecting the values and interests of the dominant researchers in a science and their reference groups.” Hunt (1991) uses the term in a broad sense to connote a “worldview.” Although S-D logic is not a paradigm (Lusch and Vargo 2006b; Vargo and Lusch 2006; Vargo 2007b), according to the criteria above, it functions at a paradigmatic level and provides an alternative lens, a mindset, through which phenomena can be examined. Therefore, it could become a paradigm if, by definition, it becomes a worldview. But worldviews are determined bottom up rather than top down, and, thus, it is the discipline that will make this determination over time. Although the current underlying paradigm, or dominant logic of marketing, as well as economic science and its other derivatives, remains goods-dominant, a paradigm shift is possible. Goods-Dominant (G-D) Logic As the label suggests, G-D logic (Vargo and Lusch 2004a; Lusch and Vargo 2006b) focuses on goods—or more generally, “products,” encompassing both tangible (goods) and intangible (services) units of output—as the basis of exchange. G-D logic can be paraphrased as follows (see Vargo and Lusch 2004a): 1. Economic exchange is fundamentally concerned with units of output (products). 2. These products are embedded with value during the manufacturing (or agricultural, or extraction) process. 3. For efficiency, this production ideally (a) is standardized, (b) takes place in isolation from the customer, (c) can be inventoried to even out production cycles in the face of irregular demand. 4. These products can be sold in the market by capturing and stimulating demand in order to maximize profits.
128
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
In essence, G-D logic says that the purpose of the firm is to make and sell things. It has also been called the “neoclassical economics research tradition” (e.g., Hunt 2000), “manufacturing logic” (e.g., Normann 2001), and “old enterprise logic” (Zuboff and Maxmin 2002). The foundation for G-D logic is grounded in economic science, as developed from the work of Adam Smith (1776/1904), “the father of economics.” Smith did not literally invent economics, nor was that even his purpose. Smith was a moral philosopher whose intention was to identify the normative activities that would lead to national wealth rather than to provide a positive framework for economic science. Smith initially built his political-economic views on the foundational proposition of the efficiency of the “division of labor,” resulting in the necessity of “exchange.” For Smith (1776/1904, p. 1), labor was the “fund which originally supplies (the nation) with all the necessities and conveniences of life which it annually consumes.” Thus labor, the application of mental and physical skills—that is (essentially) service (see Vargo and Lusch 2004a)—provided the foundation for exchange. However, having established labor/service as central to exchange and well-being, and the central metric of exchange as value-in-use—benefit in relation to the labor required to achieve it—Smith partially abandoned this model. He was not inherently concerned with all of exchange or with economic exchange in general. As noted, he was seeking a normative explanation about which types of service should be promoted in order to advance national wealth. He thus shifted the focus to value-in-exchange (nominal value, market price), rather than value-in-use, which he felt was easier to understand and also simplified his task of the identification of activities that contributed to the creation of national wealth. In Smith’s eighteenth-century world, with limitations on personal travel and the nonexistence of electronic communication, the primary route to wealth creation was the export of tangible goods, and the source of these goods was manufacturing. Thus, his underlying model was centered on the product—surplus tangible goods that could be exported. This narrowed focus on the exchange value of tangible goods can be seen in his extended discussion of the distinction between “productive” and “unproductive” activities (see Vargo and Morgan 2005). For Smith, only those activities that contributed to the creation of surplus tangible goods were deemed “productive.” Other activities, though useful and essential to individual well-being, were called “unproductive” because they did not create exportable, tangible goods. The economic philosophers (e.g., Say 1821; Mill 1848/1929) who followed generally disagreed with Smith’s productive versus unproductive distinction, reasoning that all activities that contributed to well-being were productive; but, having done so, they also generally acquiesced. Smith’s (1776/1904) productive/unproductive distinction had taken solid root, and, over time, “products” (tangible goods that could be exported) became the focus of economics; value morphed from usefulness to an embedded property of goods (essentially value-in-exchange); unproductive morphed into services (intangible goods); and a clear distinction between producers (creators of value) and consumers (destroyers of value) was established. This product-, or goods-, based model of economic activity was also convenient because it was compatible with the increasing desire of the economists to turn economic philosophy into economic science. The model of “science” at that time was Newtonian Mechanics, a model of matter embedded with properties. Therefore, an economic model of products embedded with utility had natural compatibility and appeal. At least partly because of the desire for scientific respectability, the goods‑centered paradigm survived and flourished. Economics and the derivative business disciplines—as well as more general, societal understandings of commerce—emerged and developed from this G-D paradigm. Services (usually plural), from this G-D perspective, are
Service-Dominant Logic: A Review and Assessment
129
seen as either (1) a restricted type of good (i.e., as intangible units of output) or (2) an add-on that enhances the value of a good. The dominance of a goods-centered paradigm can be recognized in Shostack’s (1977, p. 73) statement, “The classical ‘marketing mix,’ the seminal literature, and the language of marketing all derive from the manufacture of physical goods,” and the terminology used to describe the related marketing phenomena, such as “producer,” “consumer,” “goods and services,” “supply chains,” “channels of distribution,” “value-added,” etc. Arguably, this G-D paradigm is increasingly viewed as severely restricted, as evidenced by the call for a more encompassing and solid paradigmatic foundation by a number of scholars (e.g., Gronroos 1994; Gummesson 1995; Hunt and Morgan 1995; Schlesinger and Heskett 1991; Shostack 1977). Service as a Unifying Concept S-D logic moves the understanding of markets and marketing from a product or output-centric to a service or process-centric focus. The most distinguishing difference between G-D logic and S-D logic can be seen in the conceptualization of service. As mentioned, in S-D logic service is defined as the application of competences (knowledge and skills) for the benefit of another party (Vargo and Lusch 2006). The use of the singular “service” as opposed to the plural “services,” as traditionally employed in G-D logic, is intentional and significant. It signals a shift from thinking about value creation in terms of operand resources—usually tangible, static resources that require some action to make them valuable—to operant resources—usually intangible, dynamic resources that are capable of creating value (Constantin and Lusch 1994). That is, whereas G-D logic sees services as (somewhat inferior to goods) units of output, S-D logic sees service as the process of doing something for and with another party, and thus always as a collaborative process. In S-D logic, the purpose of economic exchange is to provide service in order to obtain reciprocal service—that is, service is exchanged for service. While goods are sometimes involved in this process, they are appliances for service provision; they are conveyors of competences. In either case—service provided directly or indirectly through a good—it is the knowledge and skills (competences) of the providers and beneficiaries that represent the essential source of value creation, not the goods, which are only sometimes used to convey them. Importantly, S-D logic represents a shift in the logic of exchange, rather than a shift in the type of product that is under investigation. It is a shift that Vargo and Lusch (2004a) insist is already taking place. They point out that evidence of this “new logic” can be found in somewhat diverse academic fields such as information technology (e.g., service-oriented architecture), human resources (e.g., organizations as learning systems), marketing (e.g., service and relationship marketing, network theory), and the theory of the firm (e.g., resource-based theories), etc., as well as in practice. Additionally, this “new logic” is actually an old logic in the sense that it recaptures the foundational ideas of value creation through the reciprocal application of knowledge and skills that Smith (1776/1904) established before abandoning them to discuss national wealth creation. It also can be seen in the work of Bastiat (1848/1964, p. 162), a nineteenth-century economist who claimed, “Services are exchanged for services . . . it is the beginning, the middle, and the end of economic science.” S-D logic does not imply that goods-based models of exchange should be modified to transition to a service orientation. Rather, it suggests that a service-based foundation, built upon servicedriven principles, establishes a generalizable logic for understanding all economic activity (i.e., even when goods are involved) and provides a more robust logic for transitioning from a goodscentered to a service-centered perspective.
130
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Thus, S-D logic plays a unifying role; it not only accounts for goods in exchange, it actually gives them an important role as service-delivery vehicles. The same is not the case with the treatment of service(s) in G-D logic, in which service(s) has traditionally been all but ignored and has only recently been treated as somewhat inferior (i.e., lacking in tangibility, inability to be standardized, inability to be created separate from the customer, and incapable of being inventoried). This “goods-versus-services” rift has required separate streams of marketing literature. Arguably, more generally, G-D logic has created additional conceptual bifurcations, such as the producer-consumer and the related business-to-consumer versus business-to-business distinctions, and necessitated the development of other (sub)disciplines (see Vargo and Lusch 2008c). As we will discuss, these distinctions largely vanish in an S-D logic conceptualization of exchange, markets, and marketing, creating at least the potential for a unified theory of marketing and exchange. Foundations of S-D Logic Many of the concepts (e.g., value co-creation and operant resources) for which S-D logic argues are neither exclusive to nor invented by S-D logic itself. S-D logic captures shifting contemporary marketing thought, in which marketing is seen as a facilitator of ongoing processes of voluntary exchange through collaborative, value-creating relationships among social and economic actors (e.g., individuals and organizations). S-D’s logic development began largely through the unification of existing and emerging views of exchange that stray from the traditional goods-centered logic. This service-centered view draws on historical arguments such as Smith’s (1776/1904) definition of “real value” rooted in labor, Say’s (1821) creation of utility rather than matter, Mill’s (1848/1929) classification of productive labor, and Bastiat’s (1860) criticism of tying value to tangible resources. S-D logic is grounded in the convergence of historical ideas and existing literature in marketing, economics, and management (e.g., Gummesson 1995; Normann and Ramirez 1993; Shostack 1977), as well as influential marketing theories in services and relationship marketing (e.g., Gummesson 1995; Gronroos 1994), resource-advantage theory (e.g., Hunt 2000), core competency theory (e.g., Day 1994; Prahalad and Hamel 1990), and network theory (e.g., Achrol 1999; Hakansson and Snehota 1995; Norman and Ramirez 1993), that bring to light an alternative, service-centered logic of the market (for detailed evolution see Vargo and Lusch 2004a). When the discussion of an S-D logic for marketing emerged (Vargo and Lusch 2004a), eight foundational premises (FPs) were offered to establish a framework for the service-centered mindset. Driven by the spirit of co-creation and continuous evolution, these premises have been revised and extended through the dialogue and discussion among various scholars. Minor revisions and one addition (FP9) occurred in The Service Dominant Logic of Marketing: Dialog, Debate, and Directions (Lusch and Vargo 2006a; Vargo and Lusch 2006). A more complete revision and the addition of FP10 were presented in Vargo and Lusch (2008a), in the special issue of the Journal of the Academy of Marketing Science. The ten FPs, as modified, are shown in Table 6.1 (Vargo and Lusch 2008a, p. 7) and discussed below. Together, they provide the foundation for S-D logic. FP1. Service Is the Fundamental Basis of Exchange FP1 establishes the core premise of S-D logic: the purpose of exchange is mutual service provision. Service, as defined in S-D logic, is the use of one’s competences for the benefit of another party. It is differentiated from the plural “services,” which implies a type of good and is characterized by (often considered inferior) qualities of intangibility, heterogeneity, inseparability, and
Service-Dominant Logic: A Review and Assessment
131
Table 6.1
Service-Dominant Logic Foundational Premises Original foundational premise
Modified/new foundational premise
FP1
The application of specialized skill(s) and knowledge is the fundamental unit of exchange.
Service is the fundamental basis of exchange.
FP2
Indirect exchange Indirect exchange Because service is provided through complex masks the fundamental masks the fundamen- combinations of goods, money, and instituunit of exchange. tal basis of exchange. tions, the service basis of exchange is not always apparent.
FP3
Goods are a distribution mechanism for service provision.
FP4
Knowledge is the Operant resources fundamental source of are the fundamental competitive advantage. source of competitive advantage.
The comparative ability to cause desired change drives competition.
FP5
All economies are services economies.
Service (singular) is only now becoming more apparent with increased specialization and outsourcing.
FP6
The customer is always The customer is Implies value creation is interactional. a co-producer. always a co-creator of value.
FP7
The enterprise can The enterprise cannot only make value propo- deliver value, but sitions. can only offer value propositions.
Enterprises can offer their applied resources for value creation and collaboratively (interactively) create value following acceptance of value propositions, but cannot create and/or deliver value independently.
FP8
A service-centered view is customer oriented and relational.
A service-centered view is inherently customer oriented and relational.
Because service is defined in terms of customer-determined benefit and co-created, it is inherently customer oriented and relational.
FP9
Organizations exist to integrate and transform microspecialized competences into complex services that are demanded in the marketplace.
All social and economic actors are resource integrators.
Implies the context of value creation is networks of networks (resource integrators).
Value is always uniquely and phenomenologically determined by the beneficiary.
Value is idiosyncratic, experiential, contextual, and meaning laden.
FP10
Goods are a distribution mechanism for service provision.
All economies are service economies.
Source: Vargo and Lusch 2008a.
Comment/explanation The application of operant resources (knowledge and skills), “service,” as defined in S-D logic, is the basis for all exchange. Service is exchanged for service.
Goods (both durable and nondurable) derive their value through use—the service they provide.
132
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
perishability (e.g., Zeithaml, Parasuraman, and Berry 1985). Thus, economic exchange involves doing something for another party under the condition that the other party applies its competence reciprocally. The complex and indirect nature of economic exchange, however, makes this rather simple tenet difficult to see (FP2). FP2. Indirect Exchange Masks the Fundamental Basis of Exchange Complex, indirect processes associated with exchange make its service-for-service nature easy to miss. That is, service is often provided through goods (see FP3) and goods can also be used as a form of currency (e.g., held for trade rather than used for self-service). Service exchange also often occurs through the combination of the applied competences of internal micro-specialists (e.g., assembly-line workers) and/or the combination and integration of the micro-specializations (e.g., the combination of all the resources contributed and integrated by the various participants in the “value network or constellation”). Furthermore, in monetized exchange, reciprocal service provision is lagged in relation to the initial transaction until the rights to future service (money) are used in subsequent exchange. But all these exchange vehicles (i.e., goods, money, organizations, and networks) represent intermediaries of complex exchange processes, rather than the essential bases of exchange. Behind these institutions are individuals applying their competences for the ultimate benefit of another party so that they can receive the benefit of applied competences that they do not possess. That is, regardless of its dynamic and complex structure, the essence of market exchange remains the same; to better the circumstances of both parties, people still exchange their applied competences (e.g., knowledge and skills) for the applied competences of others. They exchange service for service. FP3. Goods Are a Distribution Mechanism for Service Provision When service is understood as the basis of all exchange, goods take on a role of vehicles or transmitters for service. The value of a good is not created in a factory and distributed to the market; rather, it is derived and determined through its contribution to the customer’s self-service process—its value-in-use. Thus, rather than “services” representing a special case of “intangible goods,” as they have been conceived under G-D logic, goods are actually a special case of, or a vehicle for, indirect service provision. The basis of exchange is always service provision; goods, when used, are appliances for service provision. FP4. Operant Resources Are the Fundamental Source of Competitive Advantage One of the hallmarks for S-D logic (Vargo and Lusch 2004a) and most critical differences between S-D logic and G-D logic is the distinction between operand and operant resources (Constantin and Lusch 1994). Operant resources produce effects, whereas operand resources need to be acted upon to do so. Operand resources are usually tangible and static, whereas operant resources are usually intangible and dynamic. Almost by definition, G-D logic is centered on operand resources. S-D logic refocuses exchange on operant resources by shifting from units of output to the process of using competences for the benefit of another party—that is, service—so that the other party will reciprocate with its own applied competences. Thus, the ability to compete in the market is a function of knowledge, both individual and collective (e.g., organizational). This does not diminish the importance of operand resources in human well-being. It simply acknowledges that operand resources only become valuable in the context of active resources—for example, modification, combination, and use (see
Service-Dominant Logic: A Review and Assessment
133
Zimmerman 1951). Since S-D logic implies that value is created through activity, it points toward the primacy of the human resources of the firm (Lusch, Vargo, and O’Brien 2007) and underscores the necessity of seeing the customer as endogenous to value creation (see FP6). FP5. All Economies Are Service Economies Contrary to the popular perspective of the “new service economy,” service provision is not just now becoming abundant, nor is it only recently gaining importance. According to S-D logic, what is often seen as an emerging “service economy” is actually an aberration of a G-D logic perspective in which manufacturing (and agriculture and extraction) has been considered primary. What is happening now is that the service nature of exchange is becoming increasingly apparent as specialization increases and as less of what is exchanged fits the dominant manufactured-output classifications of economic activity. In S-D logic, there is no “service revolution,” except for the revelation in service-centered thinking. However, arguably, one revelation that is making the service nature of exchange more apparent is an information revolution (Rust and Thompson 2006). That is, the increase in specialization can be understood in terms of the exponential rate of increase in knowledge and the increasing ability to exchange information (operant resources) in a relatively pure form—that is without being transported by people and/or matter (liquification in Normann’s [2001] terms)—through digitization. FP6. The Customer Is Always a Co-creator of Value Service implies interactivity. In the parlance of the G-D logic vision of services, this is captured in the inseparability of production and consumption. While Vargo and Lusch (2004a, b) argue that inseparability is not a useful distinguishing characteristic of service, it is probably a universal characteristic of value creation. That is, value creation occurs at the intersection of providers and beneficiaries and is always determined by the latter. Stated somewhat differently, value is always created through use, rather than manufactured and then delivered. Thus, use implies the application of the customer’s operant resources in addition to those applied by the provider. All this suggests that the customer is always an active participant of the value-creation process—that is, a co-creator of value. The term “co-creator” requires further explanation. In the original FP6 (Vargo and Lusch 2004a), the term “co-producer” was used. Clearly, “co-producer” has G-D logic connotations. Thus, it was changed to co-creator of value in Vargo and Lusch (2006). However, the term co-production has a useful S-D logic meaning as well. While the “co-creation” of value describes the effect of the process of joint application of operant resources among firms and customers in creating benefit for the customer, “co-production” can be considered a subset of co-creation. Co-production involves the creation of the core offering itself and can occur through shared inventiveness, design, and/or production of the firm’s value proposition. Therefore, from a service-centered perspective, the customer always co-creates value through use and can be, though is not always, a participant in the co-production process. This requirement, that value must be co-created, implies that the firm can only propose its creation. FP7. The Enterprise Cannot Deliver Value, but Can Only Offer Value Propositions Consistent with FP6, FP7 makes explicit the idea that the firm cannot make and deliver value. That is, based on the collaborative nature of value creation and the customer’s determination of value, derived contextually and through use, the firm can offer only value propositions.
134
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
FP8. A Service-Centered View Is Inherently Customer Oriented and Relational In the G-D logic conceptualization of exchange, with its focus on transactions and units of output, “customer orientation” usually means something like making units that the customer will buy, and “relationship” usually refers to multiple transactions occurring over time. That is, they are both normative adjustments to G-D logic necessary for increasing the long-term profitability of the firm through selling more goods (tangible or intangible). However, S-D logic’s central tenets of service being the basis of exchange, service being defined in terms of benefit, and the application of operant resources from both parties to create value, make exchange inherently interdependent and relational, beneficiary (e.g., customer) oriented and beneficiary centered. Stated alternatively, in the consideration of value creation within G-D logic, the firm and the customer are separate, with the former seen as a producer of value and the latter as a destroyer. Within S-D logic, value creation is an interactive process, and thus, value is created in a relational context. FP8 is a positive, rather than normative, statement about how value is created through the relationships of parties in service-for-service exchange. Customer centricity and relationships are not options; they are realities of value-creation processes within markets. The service-for-service nature of exchange extends customer centricity dual (“firm”-“customer”) to “balanced centricity” (see Gummesson 2008). In FP9, the relational orientation is extended to the firm’s network of resources as well as the customer’s. Vargo (2009, forthcoming) more explicitly distinguishes relationship from repeat patronage, or multiple transactions, by associating the former with the more comprehensive, networked process of value creation and transactions as “temporal isolates” in that process. FP9. All Social and Economic Actors Are Resource Integrators The premise that value is co-created through the combined activities of providers and beneficiaries implies that value is determined through the integration of both provider-supplied and beneficiarysupplied resources. That is, the value derived and determined by each actor in an exchange is heterogeneous, based on existing competences (e.g., knowledge and skills), access to other resources (both operant and operand), and the situational context. In the traditional G-D logic–based literature, this integration is most often captured in concepts of supply chains (or more recently, value networks) and manufacturing. In the G-D logic–based “services” literature, it is partially captured in the observation of the heterogeneity of services. S-D logic extends the manufacturing logic from making things to integrating resources to create “densities” (Normann 2001). Normann (p. 27) defines maximum density as a situation in which “the best combination of resources is mobilized for a particular situation—for example, for a customer at a given time in a given place—independent of location, to create the optimum value/cost result.” Think of it as follows: At a given time and place, can a party bring together and integrate all the resources necessary to co-create the best possible value? With S-D logic, the integrating of resources to create densities is extended from the concept of provider and supply networks to the service beneficiary (e.g., “customer”) and beneficiary networks. This network-with-network model, in which each actor is combining resources from multiple parties to create value, is similar to Gummesson’s (2006) “many-to-many” marketing. The integrative, network-with-network model of value creation is not limited to individuals and firms or to economic exchange. It applies equally to all actors and institutions (e.g., families, firms, cities, nations) in their creation of value for themselves through the integration of resources acquired through both economic and social exchange. This foundational premise is a generalized version of the more restricted FP9 introduced by Vargo and Lusch (2006), as modified by Vargo and Lusch (2008a).
Service-Dominant Logic: A Review and Assessment
135
Table 6.2
Conceptual Transitions Goods-Dominant Logic Concepts
Transitional Concepts
Service-Dominant Logic Concepts
Goods Products Feature/attribute Value-added Value-in-exchange Profit maximization Price Equilibrium systems Supply chain Promotion To market Product orientation
Services Offerings Benefit Co-production Value-in-use Financial engineering Value delivery Dynamic systems Value chain Integrated marketing communications Market to Market orientation
Service Experiences Solution Co-creation of value Value-in-context Financial feedback/learning Value proposition Complex adaptive systems Value-creation network/constellation Dialogue Market with Service orientation
Source: Adapted from Lusch and Vargo 2006c, p. 286.
FP10. Value Is Always Uniquely and Phenomenologically Determined by the Beneficiary This tenth foundational premise was added (Vargo and Lusch 2008a) to capture the experiential nature of value more explicitly. Although implicitly suggested by the S-D logic definition of service, various FPs (e.g., FP6, FP8, and FP9), and other, less-formalized, conceptual notions (e.g., consumer’s perceptions, meeting higher-level needs, customer determination), Vargo and Lusch (2008a) formalized the unique and contextual interpretation of value. The word “phenomenological,” rather than “experiential,” was selected because the term “experience” is often interpreted to have positive-only connotations (e.g., something of a “Disney World event”—a “wow” factor), rather than positive, neutral, or negative contextually specific meanings. Conceptual Transitions Embracing a service-centered perspective requires rethinking marketing, if not all of economic science. It requires transitioning from one mental model to another and from one lexicon to another. S-D logic intimates a very different kind of purpose and process for marketing activity and for the firm as a whole: to provide service to stakeholders, including customers, stockholders, and employees. In general, by placing service, rather than goods, at the center of exchange, S-D logic moves the focus of marketing and value creation from tangible (operand) resources to intangible (operant) resources, such as knowledge and skills (Lusch and Vargo 2006c). Table 6.2 (Lusch and Vargo 2006c, p. 286) provides a summary of this and other conceptual transitions associated with moving from the dominant goods-logic toward the emerging service-logic for marketing. The S-D logic–related concepts represent a vision of the language that is needed to further a service-centered logic. No doubt that these will be revised and elaborated as S-D logic evolves (see for example Lusch, Vargo, and Wessels 2008). However, this transition from G-D logic to
136
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Figure 6.1 The Evolution of Marketing.
To Market (Matter in motion)
Through 1950
Marketing To (Management of customers and markets)
Marketing With (Collaborate with customers and partners to produce and sustain value)
1950–2010
2010 onward
Source: Adapted from Lusch, Vargo, and O’Brien (2007), p. 7.
S-D logic–compatible concepts is a difficult one, partly because the former serves not only as the formal language of the goods-centered model of economic exchange, but also as the vernacular foundation for everyday thought about exchange, marketing, and business in general. Thus, first attempts at grasping and expressing S-D logic are often through transitional concepts—G-D logic interpretations of S-D logic (middle column). The emergence of S-D logic can also be seen in another transition, in marketing “with” rather than “to” customers. The shift in primacy of resources, from operand to operant, has implications for how exchange processes, markets, and customers are perceived and, thus, with how they are approached. Focusing on the primacy of operant resources, S-D logic views customers as resources that are capable of acting with other resources and collaborating to co-create value with the firm (Vargo and Lusch 2004a). Thus, S-D logic considers customers as dynamic, knowledge-generating, and value-creating resources. This is a fundamental transition away from G-D logic, which views customers as operand resources that the firm acts upon. From a G-D logic perspective, customers are considered exogenous to the firm and are “segmented” and “targeted,” and often considered “manipulated” in the process of value creation. The primary focus of marketing within G-D logic is to identify customers, and market and sell to them. In the same way that G-D logic considers customers as operand resources, this output-focused paradigm also treats employees and other network partners as static resources that are “managed,” if not manipulated. Alternatively, S-D logic views all exchange partners as operant resources that can, and arguably must, be collaborated within the value-creation process. From this perspective, employees, customers, and other network partners become the primary source of a firm’s innovation, competence, and value. In addition, while the G-D marketing paradigm assumes the external environments (legal, competitive, social, physical, technological, etc.) as largely uncontrollable and forces to which the firm needs to adapt (McCarthy 1960), S-D logic inverts this assumption and views the external environments as resources the firm draws upon for support by overcoming resistances and co-creating these environments. Figure 6.1 (Lusch, Vargo, and O’Brien 2007, p. 7) depicts the shift of marketing philosophies and the evolution from a goods- toward a servicedominant logic.
Service-Dominant Logic: A Review and Assessment
137
Figure 6.2 Service(s) Exchanged for Service(s)
The Market
Party A Performing Service(s)
Money as a medium of exchange Goods as distribution channels Organizations as resource integrators Networks as linkages for exchange
Party B Performing Service(s)
Intermediaries of Service-for-Service Exchange
Market and Marketing Perspectives of S-D Logic S-D logic’s focus on interdependent relationships and reciprocal, service-for-service exchange suggests that markets and marketing have primary societal functions. Society and service exchange are almost synonymous concepts. Human well-being, if not survival, depends on the reciprocal exchange of applied competences (knowledge and skills). This exchange is sometimes purely social and sometimes economic, but most often both, in a complex web of service-for-service exchange (Lusch and Vargo 2006b). Marketing as a Social and Economic Process (Rather than an Outcome) S-D logic’s emphasis on systems of service-for-service exchange suggests that an examination of the market needs to precede marketing analysis (Venkatesh, Penaloza, and Firat 2006). From this perspective, organizations and other social and economic institutions are co-created to facilitate the exchange of applied knowledge and skills among individuals. Furthermore, language, knowledge, norms, culture, money, and scientific paradigms are all part of a network of co-creation activities of individuals and organizations that represent society. In a real sense, society can be viewed as a macro service-provision system. This service-driven society is not a new phenomenon, nor does it lead to a new era of the economy. Ironically, it is the growth and complexity of the institutions of service exchange, particularly those associated with “industrialization” (see Vargo and Lusch 2004b; Vargo and Morgan 2005), that have masked the service-for-service nature of exchange. Figure 6.2 (see also Lusch and Vargo 2006b, p. 410) portrays the masking role of institutions and intermediaries in service-for-service exchange in society. It is important to note that it is only because of the complexities of the market, including (1) money as a medium of exchange, (2) goods as channels of distribution for service(s), (3) organizations as service intermediaries, and (4) networks that link together parties of mutual service provision, that marketing has overlooked the fundamental economic principle that service is exchanged for service. Value does not reside in and is not directly derived from money, goods, organizations, and/or the network; value is found in the joint application of knowledge and skills that generates reciprocal service provision to better
138
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
the circumstances for each other and humankind (society). Money, goods, organizations, and the network merely provide vehicles for exchange. Learning in Competitive Markets S-D logic recognizes service-for-service exchange and competing through service as learning processes. Competing within an S-D logic mindset refocuses the purpose of exchange from the acquisition of tangible, operand resources to the generation and integration of intangible, operant resources. Economic commerce is deeply embedded within social exchange, and the two are difficult, at best, to separate. S-D logic argues that social and economic actors exchange with other actors in an attempt to improve their existing conditions through improving the conditions of others (Lusch et al. 2007). Service-for-service exchange is driven by a simple hypothesis that if the actor takes a certain action (and changes), it will be better off. However, this hypothesis is tested by the perceptions of service rendered and the value derived through an exchange and consequent use. Actors enter into exchange and experience the consequences firsthand. They learn that their hypotheses can be falsified, particularly when the service rendered does not contribute to the desired experience. Each actor has an ongoing desire to improve its condition and thus, via exchange, learns what works and what does not work. The actor then responds by returning to the market to integrate more resources, developing competences that enable it to better adapt the original service rendered in exchange or finding alternatives to the market such as more selfservice, communal sharing, or other institutions for enhancing well-being. The learning process surrounding exchange provides evidence that we live in an adaptive and changing world. In S-D logic, micro entities seek to better their lives by specializing and exchanging their service(s) for the service(s) of others. Macro structures such as organizations, market segments, lifestyle groupings, fashion movements, and legal and government regulations emerge from these individual actions and become more salient. However, behind all these macro and visible trends are individuals seeking to improve their stake in life and engaging in exchange to accomplish this. By participating in exchange, individuals stimulate additional changes that ripple throughout society. As this ripple occurs, we see more and more creative effort because more and more signals are transmitted about what works and what does not work, what results in satisfaction and what creates dissatisfaction, and what results in gain over loss. The system is not perfect, but once the power of individuals exchanging, based upon their micro specializations, starts to roll out throughout the local, regional, national, and world economy and society, more and more change occurs and more variety manifests itself via the creative learning process of exchange (Lusch and Vargo 2006b). Some have argued that the effect is opposite and that the global expansion of markets has resulted in homogeneity. However, with more exchange, there is increasing refinement, division, and reintegration of resources, creating more, rather than less variety. More importantly, this debate signifies another reason why more formal study of the market and market processes is needed in marketing. Theoretical Clarifications of S-D Logic The idea of a service logic is finding increasing acceptance in academic marketing and beyond. This should not be entirely surprising, since what has become known as S-D logic was not so much created by Vargo and Lusch (2004a) as it was reported and extended. That is, S-D logic is both deeply seeded in historical roots related to economic thought (e.g., Bastiat 1848/1964; Delaunay and Gadrey 1992) and representative of the convergence and extension of trends in theory
Service-Dominant Logic: A Review and Assessment
139
development from within marketing, especially its subdisciplines (see Vargo and Lusch 2008c), as well as related disciplines, such as economics and human resources. Additionally, the further development and elaboration of S-D logic since Vargo and Lusch (2004a) has been characterized by collaboration and co-creation. However, S-D logic has also been developed in the paradigmatic context and pervasive lexicon of G-D logic. Thus, in addition to the contentious substantive issues that are expected in any major scientific endeavor, there are issues of communication and interpretation. Most of these issues of understanding can be organized around the following topics: (1) general issues related to the G-D logic lexicon, (2) value as a phenomenological concept, (3) service logic as transcendence versus a goods-services dichotomy, (4) service provision as the common denominator of all exchange, and (5) the role of S-D logic as a foundation for a positive theory of the market versus S-D logic as a normative theory of marketing. The Goods-Dominant Logic Lexicon Influence As mentioned, the deeply seeded roots of a goods-centered lexicon have created difficulties for the communication and development of S-D logic in marketing. That is, the language of G-D logic is the foundational language of marketing and, as such, contributes both directly and indirectly to much of the concern regarding S-D logic. The dominant lexicon reflects more than just words available to talk about marketing; it reflects the underlying paradigm for the thinking about and understanding commerce, the market, and exchange in general. This presents problems for discussing and describing a counter-paradigmatic view, such as S-D logic. Often, there are no alternative, generally acceptable, counter-paradigmatic or even neutral words available. Several misperceptions of S-D logic have been noted (Lusch and Vargo 2006c; Vargo and Lusch 2006; Vargo and Lusch 2008a) that can be directly attributable to language limitations, such as the concepts of “service” versus “services,” “co-creation” versus “co-production,” and “use value” versus “utility.” Arguably, the most critical semantic issue surrounding S-D logic centers on the use of “service” as its designator. Also arguably, no other issue is as tied to the difficulty of using words that have specific G-D logic meanings for explicating the nuances of S-D logic. Some have raised concerns that “service” has too much baggage (e.g., Lehmann 2006). Others have suggested that the S-D logic definition of service is “novel” or “inconsistent” (e.g., Achrol and Kotler 2006; Levy 2006), and still others have argued that it is just the wrong choice and/or it creates a false dichotomy between goods and service (e.g., Brodie, Pels, and Saren 2006). Most of the issues surrounding the use of the term “service” by S-D logic seem to be tied to the fact that in G-D logic, the term “services” is usually intended to refer to units of output, intangible goods. S-D logic, on the other hand, uses the singular term “service” to refer to a process, and is neither faulty nor novel (Vargo and Lusch 2006). We Vargo and Lusch (2006; 2008b) have acknowledged the baggage associated with the term “services,” but emphasize that the term “service” is precisely correct, if not essential, because no other word is more appropriate. Likewise, the use of the term service transcends the old, intractable (see Vargo and Lusch 2004b) debate concerning the difference between goods and services (intangibility, heterogeneity, inseparability, and perishability) by reframing the issue to emphasize the relationship: service is the common denominator of exchange; goods are service-provision mechanisms. Thus, the notion of a false dichotomy between goods and service(s) (Brodie et al., 2006) is not created by S-D logic, but rather was created by G-D logic and is, arguably, resolved by S-D logic (Lusch and Vargo 2006c; Vargo and Lusch 2008b). The problem and difficulty of G-D logic–inspired words’ being inadvertently used to describe S-D logic were evident in the initial selection of the term “co-production” (Vargo and Lusch
140
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
2004a) to capture the collaborative nature of value creation. Perhaps at least in part, the use of co-production led to the argument that S-D logic does not always apply, because customers do not always want to be active participants (e.g., Wilkie and Moore 2006; Rust and Thompson 2006). As we Lusch and Vargo (2006c; Vargo and Lusch 2006, 2008a) have acknowledged, co-production is an inappropriate term to capture the “co-creation” of value, the term used since the modification of FP6, “The customer is always a co-creator of value” (Vargo and Lusch 2006). Nonetheless, “co-production” was retained to describe the involvement of the customer in the creation (e.g., codesign, or shared production, etc.) of a firm’s core offering. This argument is made with the caveat that co-production is an option (for both the provider and the customer), but value is always co-created. That is, co-creation is the common denominator for value creation and superordinate to co-production in the same way that service is superordinate to goods. Value as a Phenomenological Concept For somewhat the same, G-D lexicon–based reasons for the confusion between co-creation of value and co-production, the S-D logic meaning of value is also sometimes misunderstood. Some have suggested that the service in S-D logic implies only utilitarian or “functional” benefits, essentially what has been captured by concepts of “utility” or “value-added” (e.g., Prahalad 2004; Schembri 2006). Typically, in conversation explaining S-D logic, the use of “utility” and “value-added” are avoided. The problem is that, even though “utility” was originally intended to capture “valuein-use,” or “usefulness,” it morphed into a meaning of embedded value, or “value-in-exchange,” essentially the same as value-added, a clearly G-D logic–related concept (see Vargo and Lusch 2006). S-D logic generally supports a “value-in-use” interpretation, but even that term has at least subtle G-D logic connotations. This G-D logic connotation of “value-in-use” might also exacerbate the occasional interpretation of service as referring to value in terms of functional benefits, rather than a phenomenological interpretation by the customer. As Vargo and Lusch (2006, p. 50) noted: We suspect that our emphasis on service satisfying higher-order needs is missed because, as with many misperceptions about S-D logic, the dominant paradigmatic perspective is G-D logic. Arguably, G-D logic implies functional benefits and its dominance is why the literature is just now evolving toward grasping the role of more experiential, expressive, phenomenological, and emotional benefits. To further clarify this issue, FP10 was added (Vargo and Lusch 2008a), which captures S-D logic’s phenomenological view of value and also helps to clarify the misunderstanding that S-D logic is a restatement of the consumer orientation. S-D logic’s emphasis on value-in-use centers on the phenomenological view of value and, thus, is inherently customer oriented and customer centered. Within S-D logic, the identification of consumer orientation becomes redundant. Although S-D logic takes a customer-centric approach to value creation and emphasizes that value is derived through use, it does not suggest that value-in-exchange is not important. Rather, S-D logic recognizes the importance of financial feedback from the market (exchange value) as a learning mechanism and is compatible with the idea that such feedback is tied to accounting systems that capture value-in-exchange. Thus, while S-D logic argues that value-in-exchange could not exist independent of value-in-use, it recognizes the importance of value-in-exchange as feedback to the firm and an intermediary of service provision. The transition in thinking that occurs when one moves from focusing on value-in-exchange to
Service-Dominant Logic: A Review and Assessment
141
value-in-use is illustrated in Table 6.2. We are now beginning to recognize that value-in-use is a transitional concept that takes us from the goods-dominant concept of value-in-exchange toward a service-dominant concept of value. However, “value-in-use” does not fully reflect S-D logic thought—that is, it is transitional—and thus the term “value-in-context” (see Vargo, Maglio, and Akaka 2008) may be more fitting. Value-in-context suggests that not only is value always co-created, it is also contingent on the integration of other resources and is contextually specific. Consider the purchase of a new car. The price paid for the car is the value-in-exchange; the benefits from the use of the car represent the value-in-use. But that value is contingent on integration with other resources (driving ability, maintenance, fuel, roads, ) and the use context—for example, integrating a car with family activities, such as weekend soccer games, establishes a different value-in-use from integrating it with individual needs, such as a daily commute to work. Service-Dominant Logic as Transcendence Versus a Goods-Services Dichotomy As noted, in S-D logic, “service” transcends goods (and “services”) by delineating the relationship between service (a process of using competences for the benefit of another party) and goods (service-provision vehicles), rather than looking for the differences between types of output (goods and “services”). Similarly, rather than replacing goods with service, as some have suggested (e.g., Achrol and Kotler 2006; Brodie et al. 2006) or goods logic with service logic, S-D logic makes service and service logic superordinate to goods and goods logic in terms of classification and function. That is, although some have argued otherwise, S-D logic does not consider service to be a substitute for goods. Rather, S-D logic establishes a nested relationship in which S-D logic transcends G-D logic, meaning that the theoretical and conceptual components of G-D logic are relevant, but are not as deep or broad as those of S-D logic. Thus, it resolves the goods versus service dichotomy that is created by the G-D logic distinctions. Similarly, this notion of transcendence can be used to respond to arguments that a plurality of paradigms is needed—that is, S-D logic and G-D logic should coexist (e.g., Sweeney 2007; Winklhofer, Palmer, and Brodie 2007). At first glance, this pluralistic stance may seem as if it resolves the debatable ideas stemming from the emergence of S-D logic. However, this pluralistic approach is also unnecessary and conceivably incoherent. A service logic and a goods logic can coexist in a nested relationship, as they do in S-D logic, but that is very different from making both service and goods primary (e.g., “dominant”). Essentially, “plurality is what the discipline has had with the separation of goods marketing and services marketing. In S-D logic, that separation is not only unnecessary; it (arguably) is resolved—service and goods coexist with a common purpose (service) in S-D logic” (Vargo 2007b, p. 109). Service Provision as the Common Denominator of All Exchange One of the most consistent restatements and misstatements of the S-D logic thesis is that it is appropriate for marketing to adopt models of “services,” rather than goods, because the former are now dominant in developed economies (see, e.g., Achrol and Kotler 2006; Ambler 2006; Brodie et al. 2006; Shugan 2004). Some have even suggested that S-D logic does not go far enough in reflecting the transition in the market (e.g., Rust 2006). S-D logic does not deny that service dominates exchange today; however the transition-focused perspective does not reflect a full understanding of the central tenet of S-D logic: service is exchanged for service. Thus, service has always been the foundational basis for all exchange. Importantly, goods neither become replaced nor unimportant in S-D logic. Service is just the
142
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
common denominator. The function of goods, when involved, is to enable service—that is, goods represent a special case of service provision. As noted, it is only from the perspective of a model that includes the fundamental assumption that exchange is driven by goods (G-D logic) that the importance of service is just now becoming apparent and that the economy is perceived to be transitioning from goods to service focused. In S-D logic, service provision is the basis of exchange in all economies as well as industries and is the primary function of all organizations. Thus, though some have argued that S-D logic is not likely to apply to all organizations and/or situations (e.g., Day 2006), once service is accepted and understood as the basis of all exchange, S-D logic does not have boundary conditions. That is, it applies equally to what have traditionally (under G-D logic) been differentiated as “services” and “manufacturing” industries and organizations. The Positive vs. Normative Nature of S-D Logic Some (e.g., Venkatesh et al. 2006; Wilkie and Moore 2006) have either implicitly or explicitly indicated that S-D logic might not go far enough because it does not move marketing beyond its present managerial, or firm-centric, orientation and/or does not adequately provide a market focus. Based solely on the original Journal of Marketing article (Vargo and Lusch 2004a), their observations are possibly well founded. Marketing, by definition, is largely a managerial activity, as it should be given its origin and its original focus on application. That is, it has normative connotations. Even the word “marketing” implies doing something—going to market, acting on the market, and so on—as opposed to a more positive term like market science. Also, the conceptualization of S-D logic emerged (Vargo and Lusch 2004a) in the Journal of Marketing, which has an editorial policy of managerial relevance. Thus, Vargo and Lusch (2006) have recognized that some of the initial presentation of S-D logic was couched in managerial terms. However, S-D logic is not inherently managerial, and the non-managerial implications need to be more fully explored (see, e.g., Gummesson 2006; Laczniak 2006; Venkatesh et al. 2006; Wilkie and Moore 2006; and Lusch and Vargo 2006b). More importantly, Venkatesh et al. (2006) argue that what is missing in marketing is an adequate understanding of the market (Vargo and Lusch 2006). In agreement, Lusch and Vargo (2006b; Vargo 2007a; Vargo and Lusch 2008c) suggest that S-D logic offers a foundation for a muchneeded positive theory of markets, on which better normative theories of marketing could be based. That is, the basic premise of S-D logic, the mutual exchange of applied, specialized skills and knowledge, is a more solid foundation for understanding markets and marketing than is the very limited foundation of exchange centered on goods. Thus, S-D logic not only points toward better marketing theory but also possibly points toward a better, process-centered theory of economics and society. Further integration of the literature and generation of knowledge will likely offer insight on social and not-for-profit entities as well as ethical, legal, societal, and ecological issues (Vargo and Lusch 2008c). Such areas are already being examined (see, e.g., Abela and Murphy 2008) and show initial support for S-D logic as a more integrative approach for studying ethical and socially beneficial aspects of marketing. Service-Dominant Logic Knowledge Extensions and Integration Since its introduction in academic marketing, S-D logic (Vargo and Lusch 2004a) has stimulated scholarly dialogue concerning marketing theory and thought in general, as well as more specific phenomena of concern in various subdisciplines (e.g., service marketing, relationship marketing,
Service-Dominant Logic: A Review and Assessment
143
industrial marketing, etc.). Although the dialogue surrounding S-D logic continues to evolve, three subthemes have emerged that underlie much of the elaboration of service-centered concepts and reflect the foundational premises of S-D logic. These subthemes are: (1) the S-D logic meaning of service, (2) a resource-based perspective of the market, and (3) the process-orientation of value co-creation. The growing conversation built upon the three subthemes of S-D logic has contributed back to the research streams from which S-D logic was derived (e.g., service and relationship marketing) and continues to expand into other areas of marketing and marketing-related research (e.g., consumer behavior and business-to-business marketing). The purpose of the following section is to highlight the primary, salient issues and insights emerging from these scholarly conversations. This dialogue integrates S-D logic with existing and developing ideas related to marketing and elaborates and/or extends service-centered concepts. S-D Logic Meaning of Service Growing research focused on service-centered thinking suggests that S-D logic has the potential to provide a foundation for a paradigm shift in marketing. As a result, the existing concepts and models for marketing are increasingly being questioned and reconsidered to reflect the evolutionary transition (e.g., Ekeledo and Sivakumar 2004; Rust 2004; Woodruff and Flint 2006). The discussion surrounding the S-D logic meaning of service, as the application of operant resources (e.g., knowledge and skills) for the benefit of another, has been at the core of this dialogue and evolution. While the clarifications of why the term “service” is appropriate and precise have been discussed above, the following sections provide extensions in the literature that further develop this core concept (cf. Vargo and Lusch 2004b). Rethinking Service Marketing The S-D logic conceptualization of service appears to have a significant impact on the service marketing literature. Ottenbacher et al. (2006, p. 346) argue that S-D logic introduces “a renewed focus on the conceptual fluency between what is relevant in product marketing and what is relevant in services marketing.” As mentioned, “services” are conventionally distinguished from goods by four differentiating characteristics: intangibility, heterogeneity, inseparability, and perishability (Zeithaml et al. 1985), designated as “IHIP” characteristics (Lovelock and Gummesson 2004). Over the past few decades, the IHIP characteristics of services have been widely accepted and applied as the conventional wisdom of service marketing. However, the introduction of S-D logic directly challenges these characteristics by arguing that the IHIP differentiators assume the primacy of goods (i.e., services are what goods are not) and therefore are evidence of a G-D logic (Vargo and Lusch 2004b). Lovelock and Gummesson (2004) have questioned their usefulness in the delineation of services from goods on somewhat similar grounds. The S-D logic conceptualization of service has at least partially redirected the discussion of service from the distinction of “goods versus services” by obviating the need for a “goods versus services” dichotomy because, in S-D logic, “service,” is a transcending concept. That is, service, defined in terms of using competences for the benefit of another party, is an inclusive term, with goods representing a mechanism for service provision. While the discussion has been redirected, in part, the issues are not fully resolved, perhaps somewhat reflecting the paradigmatic power of G-D logic. For example, Edvardsson et al. (2005) question the traditionally accepted definition of services for its managerial and firm-centered focus and argue that services may be defined from the customer’s perspective by incorporating the
144
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
conceptualization of service and the idea of value co-creation supported by S-D logic (Vargo and Lusch 2004a, 2006). Edvardsson et al. (2005) reported preliminary research findings that support the connection between the S-D logic conceptualization of service and the definition of services in the traditional “services industry.” Some scholars continue, however, to consider the goods versus service(s) debate valid and useful. For example, Sampson and Froehle (2006) argue that there is no single, comprehensive, and consistent structure to differentiate goods and services, while others suggest that the characteristics of exchange phenomena should be considered on a case-by-case basis (Laine et al. 2005). Still others continue to believe that it is important to distinguish between goods and services in order to capture the important differences in consumer price fairness perceptions (Bolton and Alba 2006). As S-D logic evolves, its service-centered understanding of exchange continues to develop and raises questions regarding the relative role of goods. In contrast to this continuing debate, and consistent with S-D logic, several marketing scholars have initiated attempts to redefine the role of service marketing with regard to the overall discipline. For example, Gronroos (2006, p. 362) contends that “goods marketing” is a special case of service marketing and proposes three conclusions for marketing’s traditional focus on goods: 1. Concentrating on the product draws the marketer’s attention away from what ultimately is important for the customers: their value-creating processes. 2. Goods can be seen as a platform for services. 3. For the customer to use goods, other resources must accompany them, and the goods are only one resource among others in the process of supporting customers’ value-generating processes. Focusing on the centrality of service in marketing not only informs the marketing of goods, but also sheds light on the management of marketing in general. For example, Brown and Bitner (2006, p. 31) suggest six service-centered best practices that serve as a foundation for contemporary marketing. These best practices are proposed for all types of organizations and include: 1. 2. 3. 4. 5. 6.
Keeping promises to customers Understanding service from the customer’s point of view Recognizing that employees are the product Involving customers in co-producing services Enabling customers to serve themselves Recovering when failures occur.
The emergence of S-D logic has emphasized the centrality of service in marketing and suggests that the theories and models developed in service marketing are applicable to all of marketing (Vargo and Lusch 2008c), including the subset of instances in which goods are involved. The dialogue surrounding service-centered phenomena has marked a starting point to broaden the scope of service marketing literature and reexamine service, in terms of what it is and where it stands in the field of marketing. Solutions and Symbols The service-centered perspective of S-D logic has punctuated the emerging shift from an output-oriented to a solution-oriented approach for marketing (Michel, Vargo, and Lusch 2008).
Service-Dominant Logic: A Review and Assessment
145
Sawhney (2006, p. 365) urges marketers to go beyond the product and focus on service provision that integrates customized solutions for customers by “embracing a solutions mind-set” and focusing on providing “customized outcomes for specific customers.” He further specifies that the solution-centered mindset reflects an S-D logic for marketing and is fundamentally different from G-D logic. Contrasting the goods-centered mindset that starts with products, a solution-centered design begins with an analysis of a customer’s problem and ends with the identification of the resources, both operant and operand, that will be needed to solve the entire problem. The solution-centered mindset emphasizes the central role of the customers and customized experiences in the overall marketing process. Rust and Thompson (2006, p. 284) suggest, “as companies become increasingly service-oriented, marketing strategy will need to accompany this shift and become less product-centered and increasingly customer-centered.” The authors argue that using a customer-based, flexible framework to guide marketing strategy will lead to the greatest payoff for firms. While many are supportive of a solution-orientation for marketing, several researchers have recognized the challenges associated with the transition from a product mindset to a solution-centered approach to marketing. Day (2006, p. 88) notes that it will be difficult for firms to pursue an S-D logic and a solution-oriented marketing strategy because it would entail satisfying five criteria (integration, interaction, co-production, customization (tailored), and customer risk) for a deep relationship that transfers a supplier’s skills and knowledge to a customer who lacks such competences. Extending the movement from a product- toward solution-oriented approach for marketing, several researchers emphasize the symbolic and experiential, rather than purely functional, nature of service. For example, Flint (2006) calls upon symbolic interactionism to discuss the dynamic meanings of customer resources. “Rather than focusing on products as vessels that hold symbolic meaning for the possessor/user [symbolic interactionism] focuses on the dynamic use, interpretation, and changing meanings of symbols within social interaction” (p. 351). Flint explains that symbolic interaction echoes the notion of co-creation in S-D logic by emphasizing the active, rather than reactive, role of customer value. Along the same vein, Venkatesh et al. (2006, p. 253; emphasis in original) extend S-D logic’s view of the service nature of the economy (Vargo and Lusch 2004a) to that of a system of symbolic meaning and explain, “[Vargo and Lusch] use the term service economy in moving from an emphasis on products to services in understanding market exchange. In contrast we put forward the term market, in which the sign is a key in understanding exchange.” In addition, Duncan and Moriarty (2006) apply the notion of service and symbols in the branding literature and argue that service and integrated marketing communications perspectives and brands are correlated and interdependent. The authors suggest using integrated marketing communication touchpoints to operationalize S-D logic. Brodie et al. (2006) recognize the conceptual similarities between S-D logic and the service brand (Berry 2000). Brodie et al. (2006, p. 372) draw on the discussion of sign systems raised by Venkatesh et al. (2006) and suggest that “the service brand is a sign system that symbolizes the value processes.” While the scholarly debate on the S-D logic meaning of service continues throughout various research streams, Anderson (2006) anticipates that future modeling in the marketing discipline will continue to evolve. He argues that marketing models will shift from a goods-based perspective, which examines how marketing can influence individual purchase decisions, to a service-based perspective of customer solutions and interdependent relationships. Through this elaboration of S-D logic’s meaning of service, particularly as a solution and/or symbol (phenomenological interpretation), the intangible and dynamic forces that underlie the creation of value are highlighted, and the operant resources that drive exchange are difficult to ignore.
146
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Resource-Based Perspective of Marketing As noted, S-D logic adopts a resource-based perspective of marketing and argues for the primacy of operant, rather than operand, resources in exchange (Vargo and Lusch 2004a). Since operant resources are usually infinite and dynamic, the sustainable comparative advantages of firms are usually derived from the application and management of such resources (e.g., knowledge, skills, and competences), especially those that are tacit and not easily imitable or transferable (Lusch et al. 2007; Madhavaram and Hunt 2008). This resource-based perspective, focused on operant resources, serves as an instrumental conceptual framework to understand marketing phenomena. Elaborating the Concept of Resources A resource-based perspective on business activities was introduced into marketing largely through resource-advantage (R-A) theory (e.g., Hunt and Morgan 1995; Hunt 2000, 2002). This theory has been recognized as one of the fundamental conceptualizations tied to the emergence of S-D logic (Vargo and Lusch 2004a). R-A theory argues that heterogeneous, imperfectly mobile resources meet heterogeneous demands in the market. This theory implies significant diversity among firms and proposes resource-based comparative advantages. Supporting a resource-based perspective for marketing, Hunt and Madhavaram (2006) suggest using R-A theory to guide business and marketing strategy and further develop an S-D logic for marketing. The authors suggest that R-A theory provides S-D logic with a definition of resources: “tangible and intangible entities available to the firm that enable it to produce efficiently and/or effectively a market offering that has value for some marketing segment(s)” (p. 69). Maintaining a strong focus on the competitiveness of the firm, Hunt and Madhavaram (2006, p. 70) also explain how the value of a resource is determined, For R-A theory, not all resources that have value to the firm have an exchange value or price. That is, relatively immobile resources such as competences are not commonly or easily bought and sold in the marketplace. . . . Therefore, the value of such operant resources is determined not by exchange, but by the extent to which each contributes to the firm’s ability to produce efficiently/effectively market offerings that are perceived by some market segments to have value. While R-A theory’s focus aligns with S-D logic’s emphasis on the exchange and application of operant resources, S-D logic expands the focus of resources beyond the firm to systems of service exchange (Lusch and Vargo 2006c). S-D logic considers the operant resources of customers, employees, and the environment endogenous, rather than exogenous, to the firm. Thus, the competences of customers (e.g., Prahalad and Ramaswamy 2000), employees, and other stakeholders are key components in the competitive advantage of the firm (Lusch et al. 2007). In addition, the distinction of operant versus operand resources in S-D logic has further extended R-A theory by suggesting the primacy of operant resources in achieving competitive advantage. Madhavaram and Hunt (2008) have furthered the integration of S-D logic’s primacy of operant resources with R-A theory by developing a hierarchy of operant resources for the firm: basic, composite, and interconnected. The primacy of operant resources is not limited to those of the firm. The nature and purpose of operant resources have also been elaborated in relation to the customer, through the intersection of S-D logic and consumer culture theory (CCT) (Arnould and Thompson 2005; Arnould 2005;
Service-Dominant Logic: A Review and Assessment
147
Arnould, Price, and Malshe 2006). In particular, Arnould et al. (2006, pp. 93–94) define customers’ operant resources as physical, social, and cultural: Physical Resources: Consumers vary in their physical and mental endowments. This affects their life roles and projects; for example, low literate and physically challenged consumers’ life roles and life projects appear to differ qualitatively from those with average physical endowments. . . . Through understanding customers’ operant physical resources, firms can tailor their offerings including virtual environments that relieve physical constraints. Social Resources: Social operant resources are networks of relationships with others including traditional demographic groupings (families, ethnic groups, social class) and emergent groupings (brand communities, consumer tribes and subcultures, friendship groups) over which consumers exert varying degrees of command (Giddens 1979). Cultural Resources: Consumer culture theorists conceive of cultural operant resources as varying amounts and kinds of knowledge of cultural schemas, including specialized cultural capital, skills, and goals. Arnould et al. (2006) emphasize the role of customers as active players in economic activities and suggest that customers, like businesses, possess different types of resources. These resources are integrated “to co-create value through patterns of experiences and meanings embedded in the cultural life-worlds of consumers” (p. 91). The notion of operant customer resources has stimulated the discussion of value co-creation and resource-integration from a customer’s perspective. Arnould et al. (2006) parallel the role of firms and customers in the value creation process, in which firms deploy operant resources to mold operand resources and value propositions, while customers use operant resources to cocreate value and determine value-in-use. Such interactions present opportunities for developing and enhancing value propositions by leveraging customers’ operant resources. Figure 6.3 (Arnould et al. 2006, p. 92) illustrates the role of the customer’s operant and operand resources in the cocreation of value. Along a similar vein, Etgar (2006) brings the resource-based notion into understanding consumer behavior. He argues that customers need to make economic decisions like those used by managers in firms to optimize the use of resources available to them. Therefore, like managers, customers strive for a balance between minimizing costs and optimizing performance. S-D logic’s perspective that customers and suppliers are operant resources suggests a symmetrical, rather than asymmetrical, relationship among exchange partners (Lusch et al. 2006). This balance of mutual service provision is not limited to the dyadic relationship between a firm and customer; rather, service is continually provided through a network or constellation of value-creating activities. As S-D logic evolves, the central role of networks and interaction in value creation draws increasing attention and continues to be more heavily emphasized (e.g., Lusch and Vargo 2006b; Vargo and Lusch 2008a). Arnould (2008) argues for the further integration of S-D logic with various resource-oriented theories to investigate how operant resources interact and create value for individuals, firms, and society. Developing research on the market as a network and service as the basis of exchange will continue to refocus the understanding of value creation away from a unidirectional, chainlike process to the integration of dynamic and interconnected processes that make up systems, or networks, of service-for-service exchange.
Operand
Goals: Life projects Life roles
Operant
Consumer Allocative capabilities
Economic: Material objects Physical spaces
Source: From Eric J. Arnould, Linda L. Price, and Avinash Malshe, “Toward a Cultural Resource-Based Theory of the Customer,” in The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo (Armonk, NY: M.E. Sharpe, 2006), p. 92. Used by permission.
Physical: Sensorimotor endowment Energy, emotions, strength
Cultural: Specialized knowledge and skills Life expectancies and history Imagination
Social: Family relationships Brand communities Consumer tribes Commercial relationships Authoritative capabilities
Figure 6.3 The Consumer’s Operant and Operand Resources
148
Service-Dominant Logic: A Review and Assessment
149
Systems of Resource Integration The examination of resource-based processes for value creation offers insight concerning how firms should integrate resources through various marketing systems, networks, and intermediaries. Within S-D logic, the “venue” of value creation is found in value configurations—interactions among economic and social actors—and thus, value is created within and among systems of exchange, at various levels of aggregation (Vargo and Lusch 2008a). Based on an S-D logic view, a “value network” or “service ecosystem” (Lusch et al. forthcoming) has been recognized as a “spontaneously sensing and responding spatial and temporal structure of largely loosely coupled value proposing social and economic actors interacting through institutions and technology, to (1) co-produce service offerings, (2) exchange service offerings, and (3) co-create value.” Recent literature integrates this S-D-logic conceptualization of a value network with models of value chain management. Although “value chains” are used to describe several of these related models, it is important to note that while S-D logic recognizes that linear processes exist, they do so within the framework of complex interconnections and interactions with other actors and processes. For example, Flint and Mentzer (2006, p. 140) present an S-D logic model for integrated value chain management with “fully interconnected and smoothly operating supply chains . . . [that] clearly reflect the goal for which supply chain management now strives: a service orientation to multiple enterprise management.” Within a business-to-business context, Flint and Mentzer (2006, p. 139) substantiate the notion of integrated value chain management by emphasizing the use of information about and by all business functions to facilitate the flow of value propositions, which now involve “products, processes, experience (history), and network relationships, all aimed at superior value creation.” The information flow in value chains aids both suppliers and firms to ensure the accuracy and efficiency of the co-production process. In addition, Flint and Mentzer suggest that the integrated supply chains facilitate the process of co-creating knowledge about markets and operations, as well as the knowledge about knowledge generation; that is, firms learn how to learn together. Kalaignanam and Varadarajan (2006) further explore the involvement of customers as co-producers along a firm’s value chain and the implications for marketing strategy effectiveness and marketing operations efficiency. They examine how product, market, customer, and firm characteristics affect the extent of customer involvement in value chain management. In sum, the notion of systems of resource integration has redefined the role of channel members, including customers, in the integrated value chain or value constellation (in S-D logic terms) management system. Mouzas (2006) investigated the underlying processes of marketing action within manufacturerretailer networks and found that companies’ marketing actions may be best understood through their network relationships. Along a similar vein, Lambert and Garcia-Dastugue (2006) present a network framework for supply chain management and address the complexity of value creation by discussing the cross-functional business processes required for implementing S-D logic in an organization. The authors integrate S-D logic with the Global Supply Chain Forum (GSCF) framework for supply chain management. The GSCF framework suggests that “whoever has the relationship with the end user has the power in the supply chain” (p. 153) and focuses the firm on value-in-use for an individual end user. Lambert and Garcia-Dastugue (2006, p. 153) demonstrate the support for an S-D logic framework for managing supply networks by using three steps: (1) mapping the network structure, (2) deciding which customers and suppliers to link with which business process, and (3) deciding the level of management to dedicate to each relationship. Figure 6.4 illustrates the complexity of managing suppliers of resources, which begins at the access of raw materials and continues through to customer use.
150
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Figure 6.4 Supply Chain Network Structure Tier 3 to Initial Suppliers
Tier 2 Suppliers
Tier 1 Suppliers
Tier 1 Customers
Tier 2 Customers
Tier 3 to Consum ers/ End-Users
1 1 1
1
2
2
3
3
n
1 n 1 2 3
1 n
n
1 n
1
n 2
n
n
Focal Company
Consum ers/End-Users ers/End-users
n
Tier 3 to n customers
suppliers Tier 3 to n Suppliers
Initial Suppliers
2 2
Members of the Focal Company’s Supply Chain
Source: Douglas M. Lambert (ed.), Supply Chain Management: Processes, Partnerships, Performance, 2d ed. (Sarasota, FL: Supply Chain Management Institute, 2006), 5. Adapted from Douglas M. Lambert, Martha C. Cooper, and Janus D. Pagh, “Supply Chain Management: Implementation Issues and Research Opportunities,” The International Journal of Logistics Management 9, 2 (1998), 3. Copyright © 2008 Supply Chain Management Institute. See www.scm-institute.org.
The resource-based perspective of marketing supports S-D logic’s premise that value is always co-created through the integration of multiple resources and is largely dependent on individual circumstance. This understanding of resource integration suggests that value is created through a continuous process of knowledge sharing and generation and is largely influenced by culture, competences, and context. This systematic view has been captured by the emergence of service science (see Maglio and Spohrer 2008), which focuses on the examination of service systems— interactive and dynamic interactions among technology, individuals, and firms. Value, within this context, is created through the integration of various resources, including existing knowledge and skills, and determined through experience. These experiences trigger learning and the generation of new operant resources and form pathways for feedback and dialogue among firms, customers, and other social and economic actors are formed. The network structure of value creation suggests that value cannot be created and delivered by any one entity. This process inherently involves customer competences in the co-creation of value, in their own context, but may also include the participation of customers in co-production of the firm’s value proposition or core offering. Etgar (2008) provides a model of the co-production process, in which the firm establishes five stages: (1) development of antecedent conditions, (2) development of motivations that prompt customers to engage in “co-production,” (3) calculation of the co-production cost-benefits, (4) activation when the customers engage in co-producing activities, and (5) generation of outputs and evaluation of the process.
Service-Dominant Logic: A Review and Assessment
151
In addition, Xie, Bagozzi, and Troye (2008) argue for the “productive” nature of “consumption” by exploring a theory of co-creation based on the idea of “prosumption.” The authors argue that “prosumption is a process rather than a single act (e.g., purchase) and consists in an integration of physical activities, mental effort, and socio-psychological experiences” (p. 10). Although this terminology seems reflective of the G-D logic language, with its emphasis on production, the underlying meaning of prosumption appears closely in line with S-D logic’s understanding of the co-creation of value, as “people participate in this process by providing input of money, time, effort and skills” (ibid., p. 10). The Process Orientation of Value Co-creation According to S-D logic, value is “defined by and co-created with the consumer rather than embedded in output” (Vargo and Lusch 2004a, p. 6). This viewpoint highlights the inherent consumerorientation of S-D logic (FP8) and stresses the importance of collaboration and learning from and with customers by being sensitive to ever-changing individual needs. From an S-D logic perspective, value creation is a continuous process focused on the provision of service, and, when production is involved, it is considered as an intermediary step (Vargo and Lusch 2004a). As such, S-D logic stresses a process-oriented value creation model (rather than the output-oriented value creation model derived from G-D logic). This process orientation of value co-creation has been extended by the discussion of relationships and interaction and emphasizes S-D logic’s phenomenological view of value. Relationships and Interaction Once the study of marketing focuses on processes rather than outputs, there is a natural link among value-in-context, value-in-use, and value-in-exchange that points toward the process-oriented and relational nature of exchange. Payne, Storbacka and Frow (2008) provide a framework for the process orientation of value co-creation from an S-D logic perspective. They propose a model that examines the value-creation processes of the firm and those of the customer as well as the interaction that occurs in market encounters. Gummesson (2006) commends the evolution of S-D logic and takes a network approach to value creation by suggesting the implementation of a win-win strategy held by relationship marketing, particularly through the use of “lean production” and “lean consumption” (see Womack and Jones 2005). Gummesson (2006, 2008) argues for balanced centricity among firms and customers and extends this view by widening the context of value creation beyond a dyadic relationship between a firm and a customer and presents a manyto-many theory for creating value. Highlighting the interdependent relationships among firms and customers, Roos, Gustafsson, and Edvardsson (2006) use S-D logic as part of their theoretical framework in defining relationship quality for customer-driven business development. They argue that no objective definition of a company’s service really exists: it is all a question of perspective. If the aim is to strive for relationship strategy, the perspective has to be that of the customer and may have to include an additional behavioral aspect. As such, similar to the notion of value co-creation and inherent to S-D logic, interaction and interdependence are central to developing relationships in the market. Several marketing scholars (e.g., Achrol and Kotler 2006; Gronroos 2006; Gummesson 2006) have pointed out that interaction and networks play a more central role (beyond relationships) in value creation and exchange than was immediately apparent in the initial S-D logic article (Vargo and Lusch 2004a). However, Lusch and Vargo (2006b) have argued that it is not so much that
152
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
S-D logic ignores interaction and networks, but rather they were originally dealt with somewhat implicitly. The centrality of markets as networks and interaction has been made considerably more explicit throughout the development of S-D logic (e.g., Vargo and Lusch 2008c). Through this elaboration, it becomes considerably more evident that S-D logic embraces the idea that value creation is a process of exchanging, integrating, and generating resources, which requires interaction and implies networks. Ballantyne and Varey (2006, p. 224) extend S-D logic’s notion of interaction and suggest that “dialogical” interaction appears to be “an ideal form of communication within the S-D logic because it supports the potential for co-creation of value and sustainable competitive advantage.” The authors advance S-D logic’s focus on operant resources by providing a triad of exchange activities that represent the fundamentals for service-dominant marketing. They suggest that the three strands—communicative interaction, relationship development, and knowledge application— make up “a fundamental conceptual unity of exchange activities,” and it is “difficult to isolate one strand of exchange activity and its effects without reference to the others” (2006, p. 230). Further, they suggest that understanding value co-creation from a triangulated viewpoint supports mutual learning and knowledge renewal. Figure 6.5 (Ballantyne and Varey 2006, p. 231) illustrates this triangular relationship and presents the tripartite fundamentals for S-D logic in marketing. While Ballantyne and Varey (2006) demonstrate the connections between interaction, dialogue, and knowledge in exchange, Berthon and John (2006) focus on the role of interactions in the exchange process. In recognition of interaction as the root of S-D logic, the authors call for a shift in focus from the entities in an exchange process to the interaction between entities. This is because service is codesigned and co-created through interaction, which constitutes “the very fabric of exchange” (p. 196). Applying the notion of interaction and dialogue in the business process, Jaworski and Kohli (2006) suggest that customer needs should be identified by the process in which a firm and its customers co-create the voice of customers. In particular, they suggest that in the needs co-creation process, both firms and customers are engaged in a joint learning process, followed by a mutual understanding concerning customers’ wants and needs. Offering a more critical approach toward the interaction among customers and firms, Wilkie and Moore (2006) emphasize the challenges with engaging in dialogue with customers. They argue that customers are often not aware of their own needs and best options for solving problems, or they are unable to communicate reliable information and are not always straightforward with firms. Further, Rust and Thompson (2006) express concerns that customers cannot, and many times do not, want to keep close, one-to-one relationships with all the firms that they interact with. In this case, an assessment process, such as that proposed by Moller (2006), appears to be crucial. His assessment involves an evaluation of the competency and accessibility of the customer, which is critical to the optimization of performance in the value co-creation process and the creation of a “user friendly” offering. The complexities that arise in the examination of relationships and interaction in marketing emphasize the process-orientation of exchange and point toward both opportunities and challenges that underlie value creation. Arguably, some of these discussions do not make the distinction between co-creation of value and co-production that Vargo and Lusch (2008a) offer, as discussed above. Experiential Nature of Customer Value S-D logic’s notions of value co-creation and value-in-use (or, more recently, value-in-context) are directly tied with literature relating to customer value. Holbrook (2006) draws connections between the original eight FPs of S-D logic and the concept of customer value (CCV) with a
Service-Dominant Logic: A Review and Assessment
153
Figure 6.5 Tripartite Fundamentals for Service-Dominant Marketing
Communicative interaction
Service-dominant exchange activities
Knowledge application
Relationship development
Source: David Ballantyne and Richard J. Varey, “Introducting a Dialogical Orientation to the ServiceDominant Logic of Marketing,” in The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo (Armonk, NY: M.E. Sharpe, 2006), p. 231. Used by permission.
model of a service-logic schema. According to Holbrook, the evolution of S-D logic argues for marketing as resource operant (RO), skills exchanging (SE), performance experiencing (PE), knowledge informed (KI), competence enacting (CE), co-producer involved (CI), value emerging (VE), and customer interactive (CI), which he abbreviates as “ROSEPEKICECIVECI.” Holbrook compares S-D logic with CCV and argues that CCV provides a more compact foundation for a marketing paradigm than S-D logic. In fact, he suggests that CCV provides a “wholesale” version of S-D logic. Vargo and Lusch (2006, p. 182) responded to this claim by explaining S-D logic’s compact or “wholesale” foundation is that “service is exchanged for service,” which offers even more parsimony than “ROSEPEKICECIVECI” and also importantly is more isomorphic with markets and exchange systems. Against the background of the evolving S-D logic, Woodruff and Flint (2006) critically examine the consumer value literature and suggest that the extant marketing research on consumer value has primarily focused on defining and categorizing consumer value typology. They argue, “For
154
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
the S-D logic to succeed as a paradigm shift, marketing thought and practice must be founded on greater in-depth understanding of customer value phenomena” (p. 183). Although S-D logic’s emphasis on the phenomenological nature of value has been addressed, the elaborations in the literature are important here. Arguing for a stronger emphasis on the experiential nature of exchange in S-D logic, Woodruff and Flint (2006) discuss the phenomenological nature of customer value. The authors propose the examination of specific value-related phenomena and offer a research agenda for studying value creation (understand customer value phenomena, understand seller value phenomena, and test theories across contexts). Others have also elaborated and extended the experiential nature of value in S-D logic. For example, in response to the emergence of S-D logic, Prahalad (2004) elaborates the idea of the customer’s involvement and engagement in the value-creation process. He argues that “when we escape from the firm and product-/service-centric view of value creation . . . and move onto an experience-centric cocreation view, new and exciting opportunities unfold” (p. 23). In addition, Schembri (2006) advocates customer experience as “the point of departure for a new service orientation within marketing” (p. 390). Arnould et al. (2006, p. 94) build upon the experiential nature of co-creation and introduce the term “co-consumption.” They extend the phenomenological nature of value co-creation beyond the isolated experience of one customer by explaining four ways that value is increased for a customer through co-consumption: 1. Co-consuming groups represent a form of consumer agency. Enhanced by computermediated communication . . . consumer groups have a greater voice in the co-creation of value than in the more atomistic situations that prevailed in the recent past. 2. Co-consuming communities represent an important information resource for participants. Co-consumer participants in brand fests and other such manifestations can not only easily turn to one another for information about products and brand, but share cultural schema nuances associated with how to consume the product or brand creatively, and interpret these experiences “properly” (Cova and Cova 2001). 3. Co-consuming groups often exhibit a sense of moral responsibility that translates to socialization of other co-consumers . . . Building on the interconnected structure of relationships and sentiments, ritual activities in consumption-oriented groupings facilitate, create and reproduce community. 4. Co-consuming groups tend to bring a relatively celebratory ethos to the consumption context. Similarly, Denegri-Knott, Zwick, and Schroeder (2006) draw from the discursive-power model to depict the increasingly powerful role of customers in the process of exchange and interaction, or the co-creation of value. The discursive-power model attempts to capture the value co-creation process, as it views power as the force that structures the possible interactions and exchanges of free agents (Foucault 1994). This theoretical framework helps illustrate the process of value cocreation and, thereafter, introduces innovation opportunities for marketers. The elaborations and extensions of value co-creation have helped to extend the service-logic perspective toward the phenomenological nature of customer value. In addition, although not directly referring to S-D logic, recent publications in major marketing journals have also reflected and alluded to the fundamental viewpoints that are highlighted in the three subthemes of S-D logic—the S-D logic meaning of service, the resource-based perspective of marketing, and the co-creation of value—presented above. In line with the S-D logic meaning
Service-Dominant Logic: A Review and Assessment
155
of service, Constantinides (2006) pointed out the limitations of a goods-dominant framework, or 4Ps marketing mix, especially the model’s managerial orientation and lack of personalization. Similarly, Iacobucci (2006) observed the adoption of the core ideas of service marketing into the mainstream. Support has been offered for the resource-based perspective of marketing through numerous studies that have recently been published on resource integration or collaboration on both the intra-organizational (e.g., Chimhanzi 2004) and the inter-organizational levels (e.g., Amaldoss and Rapoport 2005; Bradford, Stringfellow, and Weitz 2004; Yaprak, Cavusgil, and Kandemir 2006; Ross and Robertson 2007; Singh et al. 2005; Srinivasan et al. 2004). In particular, these recent publications point toward the complexity, interconnectivity, and multiplicity of the network relationships between firms (e.g., Ross and Robertson 2007; Singh et al. 2005). Finally, the discussions of the process of value co-creation can be found in recent scholarly works on relationship marketing (e.g., Rust and Chung 2006; Singh et al. 2005; Peng and Wang 2006; Palmatier et al. 2006) and customer orientation (Bettencourt et al. 2005; Donavan et al. 2004). The evolution of S-D logic continues as elaborations and extensions are made on servicecentered phenomena in the literature. The ongoing discussions surrounding the S-D logic meaning of service, the resource-based perspective of marketing, and the process orientation of value cocreation are at their early stages of development. However, increasing attention toward the major concepts associated with or supported by S-D logic indicates the need for a unification of emerging ideas. In this sense, S-D logic appears to provide a unifying framework for alternative thinking in marketing that points toward service as the basis of exchange, the primacy of operant resources, and the necessity of interaction and interdependence in value creation. The ongoing development of service-centered research seems to call for a grand theory of marketing (e.g., Gummesson 2006; Hunt and Madhavaram 2006) that connects the complexities of the market with the fundamental drivers of exchange. While S-D logic itself is not a theory (e.g., Lusch and Vargo 2006a; Vargo 2007a), it provides a revised understanding of market-related phenomena that could establish the foundation for the development of a unifying theory of the market and marketing. S-D Logic and the Evolution of Marketing Knowledge: An Assessment The marketing discipline has evolved drastically through decades of integrating research on economic, social, and psychological phenomena. Even a cursory audit of the collective knowledge generation and impact of the evolution of marketing would reveal the impressive exponential expansion of the discipline over the past century. One hundred years ago there were essentially no university professors formally educated in marketing. As late as fifty years ago the discipline was growing through the teachings of a substantial number of PhD-qualified faculty from the social sciences, particularly economics, sociology, and psychology. The hiring of such faculty is now the exception. Since the early twentieth century, formal literature on marketing has expanded to include specialized research in an increasing number of subdisciplines in marketing. Research outside North America has grown, and advances in research methods, particularly quantification, have been significant. A variety of scholarly marketing journals focused on a growing number of subareas (e.g., industrial, international, and service) are now published throughout the world in a number of languages. The most advanced form of study in marketing, the pursuit of the PhD, has flourished beyond North America over the years and is now conducted across the globe and is supplying business schools in North America, Europe, Asia, and South America with marketing-specialized faculty. In addition, marketing as an undergraduate major is often number one or two in popularity for business students, and many MBA students focus their studies on either marketing or finance.
156
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Along with indicators of increased volume of interest and specialization in marketing, there is evidence of the influential quality of marketing research. For instance, the Journal of Marketing is among over seventeen hundred journals tracked by the social science citation index (SSCI) and is often among the top five to ten percent in terms of impact ratings. It is evident that the marketing discipline has made tremendous strides in advancing the study of economic and social exchange phenomena. However, is it possible that this assessment of the collective knowledge and influence of marketing as a discipline may be misguided? The emergence of a service-logic for marketing forces the discipline to question the marketing knowledge base that has been developed over the past century. A meta-analysis of traditional goods-centered literature (e.g., pricing, promotion, placement, and products) raises more questions than answers when approached from an S-D logic lens. An evaluation of the marketing discipline through an S-D logic lens challenges the underlying G-D logic framework for marketing, as well as the purpose of the firm, and asks, “Is the fundamental purpose of marketing really to maximize firm profits by targeting and capturing customers and making decisions based on the four Ps?” Similarly, challenges to the study of customer intent and behavior surface as well, and ask, “Is the market really driven by customers attempting to make purchase decisions in order to maximize and/or optimize utility or satisfaction for themselves?” or, as suggested by FP1, “Is it possible that the market is essentially driven by the application of competences and service-for-service exchange?” If the latter more adequately reflects the purpose and nature of marketing and the market, then, as suggested by FP9, are not all economic and social actors fundamentally integrating resources to create value for themselves, others, and society at large? If such is the case, then perhaps it makes sense to see economics as addressing issues of resource allocation, whereas marketing is addressing issues of resource creation and integration. From this perspective, marketing’s historical tie to exchange is protected, but the means of exchange are viewed as service through resource integration and application, and the end is viewed as value co-creation. It may seem that these challenges to the foundation of marketing thought would most likely come from critics of marketing, who have outside paradigms and biased views of the discipline. However, should it not be the individuals that have participated in the evolution of marketing knowledge to ask questions, criticize, and develop answers regarding these critical issues? The marketing discipline has already begun to acknowledge and address deeply rooted limitations in understanding imposed by the constraints of a G-D logic paradigm. As mentioned, these explorations have been made largely under the hats of a variety of subdisciplines (e.g., service marketing, relationship marketing, industrial marketing, etc.) in the pursuit of solving marketing problems that the goods-centered, production-oriented models and concepts could not (Vargo and Lusch 2008c). The fragmentation of these research streams has called attention to the need for a unifying theory for marketing (e.g., Hunt 2000; Gummesson 2006). In the attempt to develop a grand or general theory of marketing, it seems that the G-D logic paradigm is not conducive to understanding the fundamental basis of exchange and all that results from it (e.g., social and economic systems, higher-order needs, evolution). In fact, this review of the foundation of S-D logic and evolution of marketing has essentially been focused around making this singular point, that the traditional goods-centered paradigm of marketing, focused on the production of units of output (tangible and intangible) to maximize profit, has directed the discipline away from understanding the core purpose and process of exchange. Whereas S-D logic constitutes neither a theory nor a paradigm shift as yet (see Vargo and Lusch 2006; Vargo 2007a), it does appear to provide a more robust perspective and inclusive lens for studying exchange phenomena and the development of market relationships.
Service-Dominant Logic: A Review and Assessment
157
The Formalization of S-D Logic Venkatesh et al. (2006; see also Vargo 2007a) have argued that to develop a rich understanding of the purposes, patterns, and processes of marketing phenomena, a theory of the “market” is needed to serve as the foundation for theories of marketing. In support for a positive service-centered theory of the market, Vargo (2007a, p. 60, emphasis in original) explains, “‘Marketing theory,’ almost by any definition, implies normative theory. A theory of the market on the other hand, suggests a positive theory of exchange.” As Hunt (2002) has stressed, normative theory normatively rests on a positive foundation: “good normative theory is based on good positive theory” (p. 238). S-D logic’s focus on the integration of resources, in systems of service exchange, provides a positive approach for studying exchange relationships. Although S-D logic is not a theory, Vargo and Lusch (2008c; Lusch and Vargo 2006c) have suggested that it could provide a foundation on which a true positive theory of exchange (see also Vargo 2007a) can be built. They have suggested that the development of a positive theory of the market requires a critical shift in thinking, from focusing on how firms should make different goods for different customers and market to them, to studying the meaning and process of value creation for all social and economic actors (Vargo 2008c). They also suggest that it requires getting rid of the producer/consumer distinction. Thus, from an S-D logic viewpoint, positive theory building points toward understanding how operant and operand resources are integrated to benefit individuals, groups of individuals (e.g., an organization), and/or society at large and centers on FP9: All social and economic actors are resource integrators. The development of this positive theory (general theory of the market) could, in turn, provide a foundation for a normative theory (general theory of marketing) to guide managers in effectively and efficiently approaching market-based exchange. In the meantime, even though a full S-D logic–grounded, positive theory of the market and its associated normative marketing theories have not been established, S-D logic can be used to develop empirically testable propositions. For example, the applicability of S-D logic for developing empirical studies has been demonstrated with a set of nine propositions related to competition in the market (Lusch et al. 2007). While these propositions were derived from and reflect S-D logic’s positive foundations, they also have normative implications and are capable of explaining how firms can compete through service. Table 6.3 (Lusch et al. 2007, p. 8) lists the nine derivative, testable propositions that focus on collaboration, interdependent relationships, and knowledge generation, along with the rationale for each. Within S-D logic, market competition becomes a function of how one firm provides service, or applies operant resources, to meet the needs of customers relative to other firms applying such operant resources. This has important implications for the societal contributions firms make in their attempts to achieve sustainable competitive advantages, both tactically and strategically. Proposition 1 suggests that competitive advantage is a function of how one firm applies its operant resources to meet the needs of the customer relative to how another firm applies its own operant resources. S-D logic suggests that a firm’s competitiveness depends on its collaborative abilities and collection of competences and operant resources, which the firm can continually renew, create, and transform. Emphasizing the need for collaboration among firms, Proposition 2 argues that collaborative competence is a primary determinant of a firm’s ability to acquire knowledge and establish a competitive advantage. Within and throughout the value network or constellation, knowledge and information are highly dispersed, and the firm needs to find a way to integrate knowledge and information resources. Proposition 3 highlights the influence of information technology on the market, suggesting that the continued ascendance of advancing technologies, with the associated decrease in communication and computation costs, provides firms opportunities for increasing competitive advantage through
Source: Lusch, Vargo, and O’Brien 2007, p. 8.
8a. The value network member that is the prime integrator is in a stronger competitive position. 8b. The retailer is generally in the best position to become the prime integrator. 9. Firms that treat their employees as operant resources will be able to develop more innovative knowledge and skills and thus gain competitive advantage.
5. Understanding how the customer uniquely integrates and experiences service-related resources (both private and public) is a source of competitive advantage through innovation. 6. Providing service co-production opportunities and resources consistent with the customer’s desired level of involvement leads to improved competitive advantage through enhanced customer experience. 7. Firms can compete more effectively through the adoption of collaboratively developed, risk-based pricing value propositions.
3. The continued ascendance of information technology, with associated decrease in communication and computation costs, provides firms opportunities for increased competitive advantage through innovative collaboration. 4. Firms gain competitive advantage by engaging customers and value network partners in co-creation and co-production activities.
1. Competitive advantage is a function of how one firm applies its operant resources to meet the needs of the customer relative to how another firm applies its operant resources. 2. Collaborative competency is a primary determinant of a firm’s acquiring the knowledge for competitive advantage.
Proposition
Summary of Derivative Propositions and Rationale
Table 6.3
Rationale
Since competitive advantage comes from the knowledge and skills (FP4) of the employees, it can be enhanced by servant leadership and continual renewal.
The ability to integrate (FP9) operant resources (FP4) between organizations increases ability to gain competitive advantage through innovation. Reduced barriers to technology utilization combined with the trends of open standards, specialization, connectivity, and network ubiquity increase the likelihood of collaboration with firms and customers (FP6, FP8). Because the customer is always a co-creator of value (FP6) and the firm is a resource integrator (FP9), competitive advantage is enhanced by proactively engaging both customers and value-network partners. Since value is co-created (FP6), comprehending how customers combine resources (FP8, FP9) provides insight into competitive advantage. Expertise, control, physical capital, risk taking, psychic benefits, and economic benefits influence customers’ motivation, desire, and amount of participation (FP6, FP9) in service provision through collaboration (FP8). Appropriately shifting the economic risk of either firm or customer through co-created (FP6) value propositions (FP7) increases competitive advantage. The ability to effectively combine micro-specialized competences into complex services (FP9) provides knowledge (FP1) for increased competitive advantage (FP4).
Since applied operant resources are what are exchanged in the market (FP1), they are the source of competitive advantage (FP4).
158
Service-Dominant Logic: A Review and Assessment
159
innovative collaboration. In line with this notion of collaboration, Proposition 4 argues that firms can gain a competitive advantage by involving customers, employees, and other network partners in both the co-production and the co-creation of value. Proposition 5 states that understanding how the customer uniquely integrates and experiences service-related resources (both private and public) is an additional source of competitive advantage. Concentrating on engaging the customer with the firm, Proposition 6 says that providing service co-production opportunities and resources consistent with the customer’s desired level of involvement leads to improved competitive advantage through enhanced customer experience. While it is generally understood that organizations should proactively link co-production and pricing strategies, S-D logic implies a price co-production (Lusch and Vargo 2006b) link to the firm’s value proposition. A co-produced value proposition can make the price contingent upon the quality of service experience or other agreed-upon application of service. Here, the value-inexchange (price) is tied to the value realized by the customer. Consequently, if both buyer and seller have something at risk and something to gain, then collaboration will be more fruitful. Hence Proposition 7 states that firms compete more effectively through the adoption of collaboratively developed, risk-based pricing value propositions. Proposition 8 (a and b) considers the different organizations in the market and suggests that the prime integrator in a system or market is in a stronger competitive position, and that often this prime integrator takes the form of a retailer. While this proposition suggests that there may be optimal positions in the market for integrating resources, it is important to keep in mind that all social and economic actors are resource integrators (FP10) and that all organizations must work with other firms, employees, customers, and other stakeholders to co-create value. Proposition 9 emphasizes the importance of developing employee competences and argues that firms that treat their employees as operant resources will be able to develop more innovative knowledge and skills and thus gain competitive advantage. These propositions, derived from S-D logic’s foundational premises, focus on the firm’s ability to generate internal operant resources and integrate external operant resources, through collaboration and value co-creation. Each proposition provides a positive service-centered perspective of the market, which points toward one or more social and economic managerial implications. The underlying position of this perspective of competition is that firms gain competitive advantage by adopting a business philosophy based on the recognition that all entities collaboratively co-create value by serving each other. Competing through service is ultimately about grasping and applying this understanding of value-creation processes better than the competition. Advancing Marketing Science The formalization of an S-D logic lexicon is necessary for advancing the science of marketing through S-D logic, but this initial effort marks only the beginning of the transition. Marketing science, as is true of all sciences, can be viewed as a meta-competency that is co-created by members of a value network or constellation, in which the various actors are linked via value propositions connecting internal and external service subsystems. In order to advance the scientific approach of an S-D logic for marketing, S-D logic must be co-created by the value network of marketing and marketing-related scholars. The knowledge integration portion of this review presented the early discussions, elaborations, and extensions surrounding the foundational bases of S-D logic. In addition to the development of the core concepts of service-centered marketing, methodological issues in S-D logic have been raised (see Vargo 2007b). Connections have also been made beyond the discipline of marketing to the emerging industry-led, university-supported discipline of service science (Maglio and
160
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Spohrer 2008), in which the S-D logic meaning of service is used and emphasized in the study of interactions within and among service systems. In addition, S-D logic has made its way into the discussion of new approaches toward innovation (Flint 2006; Michel, Brown, and Gallan 2008), as well as management education (Ford and Bowen 2008). The acceptance of an S-D logic for marketing is in the hands of the academics and practitioners who have built up and will continue to advance the discipline. The development of a servicecentered approach to marketing remains in its infancy, and there is much work to be done. It is our hope that scholars from marketing and related social and economic disciplines will continue to participate in this development and evolution of marketing thought. In order to further S-D logic, several fundamental questions about its integration and application in marketing science need to be addressed. We encourage others to join us in exploring several central issues in understanding and developing an S-D logic for marketing science. The first major area of development that requires attention addresses the contents and processes of value networks or constellations in marketing science. The exploration of questions such as “Who comprises the value-network or constellation for marketing science?” and “What are distinguishing characteristics of internal and external service systems of the discipline?” will provide insight into the nature and evolution of knowledge generation in marketing science. This undertaking presents a variation on stakeholder theory with a new vantage point that all systems and subsystems are involved in the exchange of competences, which both create and integrate resources. To date, most knowledge advances in marketing, at least recognized by the public domain, arise in marketing departments and universities. It is important to understand the value propositions that marketing scientists are making to their departments, universities, and students, as well as to the public and society and the implications made by changes in these relationships. The second major issue deals with S-D logic’s notion of value co-creation and asks, “How do we begin to study social and economic actors without separating them or assuming one is a ‘consumer’ and one is a ‘producer,’ but that both are creators of value?” In the development of marketing science, the exchange of operant resources and co-creation of value is undeniable, and collaboration and competition are well known as the driving forces of innovation and evolution. However, the questions remain: “Can we co-develop methods and theories that focus more on interaction than the actions of separate actors?” and “How do we shift the focus of analysis from a unit of output to something more amorphous, such as a system?” Perhaps by understanding how value is co-created within marketing science, a general understanding of value co-creation and the value that is derived through interactions among actors can be developed further. A third area of research that needs further exploration focuses on the primacy of operant resources and S-D logic’s view that resources have no value without human appraisal. Value is only derived through and determined by human ingenuity and judgment. This implies that service-exchange systems are constantly filled with surprise and uncertainty because resources are continuously created and integrated throughout the dynamic network of exchange partners. Thus, linear model fitting and tightly controlled experiments are unlikely to capture this unfolding, as controlling or attempting to control a system creates a different system. The current limitations on measuring systems of exchange increase the need for methods that focus on the logic of discovery, particularly those that can capture emergent phenomena. Research Frontiers Although the evolution of S-D logic challenges some central tenets of marketing thought and has the potential to broaden and deepen the scope of marketing science, it is careful to protect certain
Service-Dominant Logic: A Review and Assessment
161
aspects that are central to the continued advancement of marketing. It offers marketing a different and, arguably, more robust model of exchange that is also arguably more societally friendly. S-D logic also maintains marketing’s managerial relevance and does not reject current marketing knowledge as much as it transcends it. Consider the following examples: 1. Moving from viewing the customer as someone who is exogenous to the firm, to be targeted and captured, to endogenous, a potential collaborator for the co-production of the firm’s offerings and co-creator of value. S-D logic’s consideration of marketing “with,” rather than “to,” customers provides opportunities for exploring these phenomena. 2. Moving from being constrained by external environments to drawing upon them and integrating them to develop new resources. Little attention has been paid to how firms can integrate competition, public policy, culture, and technological and ecological environments as sources of support rather than treating them as resistances or constraints. 3. Moving from the marketing mix and strategy being firm designed to being co-created by the customer (and other members of the value network). Although some initial work has been done, a deeper exploration of the customer’s involvement in innovation, co-production of value propositions, and dialogue and conversation with the firm is needed. Each of these examples represents a potentially deep research frontier for marketing science. Although newcomers to the discipline are encouraged to address these topics, it is likely that established scholars in consumer behavior, pricing, promotion, product innovation, placement and distribution, and segmentation will provide very valuable insights on the evolution of S-D logic in marketing science. It is our hope that those who have helped to develop the discipline of marketing will be a part of the community that further develops S-D logic by providing expertise and insight in their respective areas of research. We have become increasingly aware of the importance of understanding the aggregate marketing system (see Gummesson 2006; Layton 2007; Meade and Nason 1991; Shultz 2007; Wilkie and Moore 2006) and other large-scale marketing systems. We believe the emerging field of service science (Maglio and Spohrer 2008), when coupled with marketing science, using an S-D logic framework, can contribute to this often-neglected topic. This implies something like a fractal model of the market. It views all systems from small systems, such as an individual or family, to larger systems, such as firms and countries, to the global system, as fundamentally comprising resources, resistances, and needs. Each system, micro and macro, strives to meet its needs through reducing resistances and integrating resources using exchange systems. Some may approach the topic of marketing systems using a different framework, but we believe this systematic, service-centered approach is vital to the advancement of marketing science. Marketing does not occur in the isolation of economic exchange; it is integrated with other elements of society and should be explored from a positive viewpoint. As noted, a theory of markets is central to the development of normative marketing theory and understanding marketing phenomena. A fractal model of marketing systems, focused on service-for-service exchange, potentially provides the fundamental framework needed for a positive exploration of the market and a deeper and broader foundation for the advancement of marketing science. Concluding Comments Service-dominant logic is a work-in-progress; in fact, it is more incomplete than it is complete. It also is not fully captured by the work of Vargo, Lusch, and coauthors. As noted, S-D logic
162
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
represents the convergence of a host of diverse models, conceptualizations, and research streams, with the common feature of responding to the inadequacies of traditional, goods-centered logic. Importantly, the development of S-D logic is increasingly involving a growing, worldwide community of scholars who have joined to help co-create, elaborate, and extend this “new dominant logic” for marketing. Further, the logic is increasingly being extended not only to marketing but also to the market—the central exchange institution upon which most, if not all, societies are built. Thus, S-D logic has implications in social as well as economic exchange. There are a host of continuing challenges associated with the further growth and development of S-D logic, but most central is the need to further develop a language and lexicon to describe the market and marketing consistent with a service-centered mindset. Despite the need to continue to refine the lexicon of S-D logic, we believe that now is the time to begin empirical study using a multitude of research methods. Although S-D logic was initially not developed as a theory with testable elements, we believe that propositions like those derived (see Table 6.3) from the initial foundational premises are empirically testable. Also, the revised and expanded foundational premises (Table 6.1) can be used to further derive testable propositions or hypotheses. These premises can be studied in several domains, such as household resource integration, enterprise value-proposing, and customer co-creation, as well as a multitude of others. However, care must be taken to ensure that propositions derived from S-D logic are tested with S-D logic–compatible metrics (see Vargo 2007b). Finally, we encourage work that deals with multiple levels of aggregation and/or analysis and, in this regard, work that simultaneously deals with the market and marketing. Importantly, this research framework may help to break down the separation between micro-marketing and macromarketing. All micro-marketing unfolds into larger macro-systems, and all aggregate marketing systems influence the micro-actors in the system. Seeking to understand the market and marketing from a holistic perspective is what S-D logic uniquely offers the marketing discipline (and others). If we neglect to take up this grand challenge of unification, we will have shortchanged both the marketing discipline and society. Note 1. Because we are consolidating our prior writing on S-D logic, this chapter is a condensed form and, in selected passages, unabridged material, from prior joint publications of Lusch and Vargo or vice versa, as well as other coauthors. References Abela, Andrew V., and Patrick E. Murphy. 2008. “Marketing with Integrity: Ethics and the Service Dominant Logic for Marketing.” Journal of the Academy of Marketing Science 36 (1), 39–53. Achrol, Ravi S. 1999. “Marketing in the Network Economy.” Journal of Marketing 63, 146–164. Achrol, Ravi S., and Philip Kotler. 2006. “The Service-Dominant Logic for Marketing: A Critique.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 320–333. Armonk, NY: M.E. Sharpe. Amaldoss, Wilfred, and Amnon Rapoport. 2005. “Collaborative Product and Market Development: Theoretical Implications and Experimental Evidence.” Marketing Science 24 (3), 396–414. Ambler, Tim. 2006. “The New Dominant Logic of Marketing: Views of the Elephant.” In The ServiceDominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 286–295. Armonk, NY: M.E. Sharpe. Anderson, Eugene W. 2006. “Linking Service and Finance.” Marketing Science 25 (6), 587–589. Arndt, Johan. 1985. “On Making Marketing Science More Scientific: Role Orientations, Paradigms, Metaphors, and Puzzle Solving.” Journal of Marketing 49 (Summer), 11–23.
Service-Dominant Logic: A Review and Assessment
163
Arnould, Eric J. 2005. “Animating the Big Middle.” Journal of Retailing 81 (2), 89–96. ———. 2008. “Service-Dominant Logic and Resource Theory.” Journal of the Academy of Marketing Science 36 (1), 21–24. Arnould, Eric J., Linda L. Price, and Avinash Malshe. 2006. “Toward a Cultural Resource-Based Theory of the Customer.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 91–103. Armonk, NY: M.E. Sharpe. Arnould, Eric J., and Craig J. Thompson. 2005. “Consumer Culture Theory (CCT): Twenty Years of Research.” Journal of Consumer Research 31 (4), 868–882. Ballantyne, David, and Richard J. Varey. 2006. “Introducing a Dialogical Orientation to the Service Dominant Logic of Marketing.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 224–235. Armonk, NY: M.E. Sharpe. Bastiat, Fredric. 1860. Harmonies of Political Economy, trans. Patrick S. Sterling. London: J. Murray. ———. 1848/1964. Selected Essays on Political Economy. Semour Cain, trans., George B. de Huszar, ed. Reprint, Princeton, NJ: D. Van Nordstrand. Berthon, Pierre, and Joby John. 2006. “From Entities to Interfaces.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 196–207. Armonk, NY: M.E. Sharpe. Berry, Leonard L. 2000. “Cultivating Service Brand Equity.” Journal of the Academy of Marketing Science 28 (1), 128–137. Bettencourt, Lance A., Stephen W. Brown, and Scott B. Mackenzie. 2005. “Customer-Oriented Boundary-Spanning Behaviors: Test of a Social Exchange Model of Antecedents.” Journal of Retailing 81 (2), 141–157. Bolton, Lisa, and Joseph W. Alba. 2006. “Price Fairness: Good-Service Differences and the Role of Vendor Costs.” Journal of Consumer Research 33 (2), 258–265. Bolton, Ruth N., George S. Day, John Deighton, Das Narayandas, Evert Gummesson, Shelby D. Hunt, C.K. Prahalad, Roland T. Rust, and Steven M. Shugan. 2004. “Invited Commentaries on “‘Evolving to a New Dominant Logic for Marketing.’” Journal of Marketing 68 (1), 18–27. Bradford, Kevin D., Anne Stringfellow, and Barton A. Weitz. 2004. “Managing Conflict to Improve the Effectiveness of Retail Networks.” Journal of Retailing 80 (3), 181–195. Brodie, Roderick J., Jaqueline Pels, and Michael Saren. 2006. “From Goods- Toward Service-Centered Marketing: Dangerous Dichotomy or an Emerging Dominant Logic?” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 307–319. Armonk, NY: M.E. Sharpe. Brown, Stephen W., and Mary Jo Bitner. 2006. “Mandating a Services Revolution for Marketing.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 393–405. Armonk, NY: M.E. Sharpe. Chimhanzi, Jacqueline. 2004. “The Impact of Integration Mechanisms on Marketing/HR Dynamics.” Journal of Marketing Management 20 (7/8), 713–740. Constantin, James A., and Robert F. Lusch. 1994. Understanding Resource Management. Oxford, OH: The Planning Forum. Constantinides, E. 2006. “The Marketing Mix Revisited: Towards the 21st Century Marketing.” Journal of Marketing Management 22 (3/4), 407–438. Cova, Bernard, and Veronique Cova. 2001. “Tribal Aspects of Postmodern Consumption Research: The Case of French In-Line Roller Skates.” Journal of Consumer Behavior 1 (June), 67–76. Day, George S. 1994. “The Capabilities of Market-Driven Organization.” Journal of Marketing 58 (October), 37–52. ———. 2006. “Achieving Advantage with A Service-Dominant Logic.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 85–90. Armonk, NY: M.E. Sharpe. Delaunay, Jean-Claude, and Jean Gadrey. 1992. Services in Economic Thought: Three Centuries of Debat. Boston, MA: Kluwer Academic Publishers. Denegri-Knott, Janice, Detlev Zwick, and Jonathan E. Schroeder. 2006. “Mapping Consumer Power: An Integrative Framework for Marketing and Consumer Research.” European Journal of Marketing 40 (9/10), 950–971. Donavan, D. Todd, Tom J. Brown, and John C. Mowen. 2004. “Internal Benefits of Service-Worker Customer Orientation: Job Satisfaction, Commitment, and Organizational Citizenship Behaviors.” Journal of Marketing 68 (1), 128–146.
164
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Duncan, Tim, and Sandra Moriarty. 2006. “How Integrated Marketing Communication’s ’Touchpoints’ Can Operationalize the Service-Dominant Logic.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 236–243. Armonk, NY: M.E. Sharpe. Edvardsson, Bo, Anders Gustafsson, and Inger Roos. 2005. “Service Portraits in Service Research: A Critical Review.” International Journal of Service Industry Management 16 (1), 107–121. Ekeledo, Ikechi, and K. Sivakumar. 2004. “The Impact of E-Commerce on Entry-Mode Strategies of Service Firms: A Conceptual Framework and Research Propositions.” Journal of International Marketing 12 (4), 46–70. Etgar, Michael. 2006. “Co-Production of Services: A Managerial Extension.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 128–138. Armonk, NY: M.E. Sharpe. ———. 2008. “A descriptive model of the consumer co-production process.” Journal of the Academy of Marketing Science 36 (1), 97–108. Flint, Daniel J. 2006. “Innovation, Symbolic Interaction and Customer Valuing: Thoughts Stemming from a Service-Dominant Logic of Marketing.” Marketing Theory, 6(3), 349–362. Flint, Daniel J., and John T. Mentzer. 2006. “Striving for Integrated Value Chain Management Given a Service-Dominant Logic for Marketing.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 139–149. Armonk, NY: M.E. Sharpe. Ford, Robert, and David Bowen. 2008. “A Service-Dominant Logic for Management Education: It’s Time.” Academy of Management Learning and Education 2 (2), 224–273. Foucault, M. 1994. “The Subject and Power.” In Power. Essential Works of Foucault 1954–1984, ed. J. Faubian. London: Penguin. Giddens, Anthony. 1979. Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis. London: Macmillan. Gronroos, Christian. 1994. “From Marketing Mix to Relationship Marketing: Towards a Paradigm Shift in Marketing.” Asia-Australia Marketing Journal 2 (August), 9–29. ———. 2006. “What Can a Service Logic Offer Marketing Theory?” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 354–364. Armonk, NY: M.E. Sharpe. Gummesson, Evert. 1995. “Relationship Marketing: Its Role in the Service Economy.” In Understanding Services Management, ed. William J. Glynn and James G. Barnes, 244–268. New York: John Wiley & Sons. ———. 2006. “Many-to-Many Marketing as Grand Theory.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 339–353. Armonk, NY: M.E. Sharpe. ———. 2008. “Extending the Service-Dominant Logic: From Customer Centricity to Balanced Centricity.” Journal of the Academy of Marketing Science 36 (1), 15–17. Hakansson, Hakan, and Ivan Snehota. 1995. Developing Relationships in Business Networks. London: Routledge. Holbrook, Morris B. 2006. “ROSEPEKICECIVECI versus CCV: The Resource-Operant, Skills-Exchanging, Performance-Experiencing, Knowledge-Informed, Competence-Enacting, Co-producer-Involved, ValueEmerging, Customer-Interactive View of Marketing Versus the Concept of Customer Value: ‘I Can Get It for You Wholesale.’” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 208–223. Armonk, NY: M.E. Sharpe. Hunt, Shelby. 1991. Modern Marketing Theory: Critical Issues in the Philosophy of Marketing Science, Cincinnati: Southwest Publishing. ———. 2000. A General Theory of Competition: Resources, Competences, Productivity, Economic Growth. Thousand Oaks, CA: Sage Publications. ———. 2002. Foundations of Marketing Theory: Toward a General Theory of Marketing. Armonk, NY: M.E. Sharpe. Hunt, Shelby, and Sreedhar Madhavaram. 2006. “The Service-Dominant Logic of Marketing: Theoretical Foundations, Pedagogy, and Resources-Advantage Theory.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 67–84. Armonk, NY: M.E. Sharpe. Hunt, Shelby, and Robert M. Morgan. 1995. “The Comparative Advantage Theory of Competition.” Journal of Marketing 59 (April), 1–15. Iacobucci, Dawn. 2006. “Three Thoughts on Services.” Marketing Science 25 (6), 581–583.
Service-Dominant Logic: A Review and Assessment
165
Jaworski, Bernie, and Ajay K. Kohli. 2006. “Co-Creating the Voice of the Customer.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 109–117. Armonk, NY: M.E. Sharpe. Kalaignanam, Kartik, and Rajan Varadarajan. 2006. “Customers as Co-Producers: Implications for Marketing Strategy Effectiveness and Marketing Operations Efficiency.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 166–180. Armonk, NY: M.E. Sharpe. Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Laczniak, Gene R. 2006. “Some Societal and Ethical Dimensions of the Service-Dominant Logic Perspective of Marketing.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 279–287. Armonk, NY: M.E. Sharpe. Laine, Teemu, Jari Paranko, Mikko Varila, and Petri Suomala. 2005. “Challenges in the Diffusion Process of New Industrial Services in the Network of an OEM.” In Frontiers of e-Business Research, 496–507. Lambert, Douglas M., ed. 2006. Supply Chain Management: Partnership, Process, Performance. Sarasota, FL: Supply Chain Management Institute. Lambert, Douglas M., and Sebastian J. Garcia-Dastugue. 2006. “Cross-Functional Business Processes for the Implementation of Service-Dominant Logic.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 150–165. Armonk, NY: M.E. Sharpe. Lambert, Douglas M., Martha C. Cooper, Janus D. Pagh. 1998. “Supply Chain Management: Implementation Issues and Research Opportunities,” International Journal of Logistics Management, 9(2), pp. 1–20. Layton, Roger. 2007. “Marketing Systems: A Core Macromarketing Concept.” Journal of Macromarketing 27 (3), 227–242. Lehmann, Donald R. 2006. “More Dominant Logics for Marketing: Productivity and Growth.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 296–301. Armonk, NY: M.E. Sharpe. Levy, Sidney J. 2006. “How New, How Dominant?” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 57–64. Armonk, NY: M.E. Sharpe. Lovelock, Christopher, and Evert Gummesson. 2004. “Whither Services Marketing? In Search of a New Paradigm and Fresh Perspectives.” Journal of Service Research 7 (1), 20–41. Lusch, Robert F., and Stephen L. Vargo, eds. 2006a. The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions. Armonk, NY: M.E. Sharpe. ———. 2006b. “Service-Dominant Logic as a Foundation for a General Theory.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 406–420. Armonk, NY: M.E. Sharpe. ———. 2006c. “Service-Dominant Logic: Reactions, Reflections and Refinements.” Marketing Theory 6 (3), 281–288. Lusch, Robert F., Stephen L. Vargo, and Alan J. Malter. 2006. “Marketing as Service-Exchange: Taking a Leadership Role in Global Marketing Management.” Organizational Dynamics 35 (3), 264–278. Lusch, Robert F., Stephen L. Vargo, and Matthew O’Brien. 2007. “Competing Through Service: Insights from Service-Dominant Logic.” Journal of Retailing 83 (1), 5–18. Lusch, Robert F., Stephen L. Vargo, and Mohan Tanniru. 2010 (forthcoming). “Service, Value Networks and Learning.” Journal of the Academy of Marketing Science. Lusch, Robert F., Stephen L. Vargo, and Gunter Wessels. 2008. “Toward a Conceptual Foundation for Service Science: Contributions from Service-Dominant Logic.” IBM Systems Journal 47 (1), 5–14. Madhavaram, Sreedhar, and Shelby D. Hunt. 2008. “The Service-Dominant Logic and Hierarchy of Operant Resources: Developing Masterful Operant Resources and Implications for Marketing Strategy.” Journal of the Academy of Marketing Science 36 (1), 67–82. Maglio, Paul P., and Jim Spohrer. 2008. “Fundamentals of Service Science.” Journal of the Academy of Marketing Science 36 (1), 18–20. McCarthy, E. Jerome. 1960. Basic Marketing: A Managerial Approach. Homewood, IL: Richard D. Irwin. Meade, William K., and Robert W. Nason. 1991. “Toward a Unified Theory of Macromarketing: A Systems Theoretic Approach.” Journal of Macromarketing 11 (Fall), 72–82. Michel, Stefan, Stephen W. Brown, and Andrew S. Gallan. 2008. “An Expanded and Strategic View of Discontinuous Innovations: Deploying a Service-Dominant Logic.” Journal of the Academy of Marketing Science 36 (1), 54–66.
166
Stephen L. Vargo, Robert F. Lusch, Melissa Archpru Akaka, and Yi He
Michel, Stefan, Stephen L. Vargo, and Robert F. Lusch. 2008. “Reconfiguration of the Conceptual Landscape: A Tribute to the Service Logic of Richard Normann.” Journal of the Academy of Marketing Science 36 (1), 152–155. Mill, John Stuart. 1848/1929. Principles of the Political Economy. London: Longmans, Green. Moller, Kristian. 2006. “Role of Competences in Creating Customer Value: A Value-Creation Logic Approach.” Industrial Marketing Management 35 (8), 913–924. Mouzas, Stefanos. 2006. “Marketing Action in Networks.” European Journal of Marketing 40 (11/12), 1271–1291. Normann, Richard. 2001. Reframing Business: When the Map Changes the Landscape. New York: John Wiley & Sons. Normann, Richard, and Rafael Ramirez. 1993. “From Value Chain to Value Constellation: Designing Interactive Strategy.” Harvard Business Review 71 (July–August), 65–77. Ottenbacher, Michael, Juergen Gnoth, and Peter Jones. 2006. “Identifying Determinants of Success in Development of New High-Contact Services: Insights from the Hospitality Industry.” International Journal of Service Industry Management 17 (4), 344–363. Palmatier, Robert W., Rajiv P. Dant, Dhruv Grewal, and Kenneth Evans. 2006. “Factors Influencing the Effectiveness of Relationship Marketing: A Meta-Analysis.” Journal of Marketing 70 (4), 136–153. Payne, Adrian F., Kaj Storbacka, Pennie Frow. 2008. “Managing the Co-creation of Value.” Journal of the Academy of Marketing Science 36 (1), 83–96. Peng, Leong Yow, and Qing Wang. 2006. “Impact of Relationship Marketing Tactics (RMTs) on Switchers and Stayers in a Competitive Service Industry.” Journal of Marketing Management 22 (1/2), 25–59. Prahalad, C.K. 2004. “The Cocreation of Value.” Journal of Marketing 68 (January), 23. Prahalad, C.K., and Gary Hamel. 1990. “The Core Competence of the Corporation.” Harvard Business Review 68 (May–June), 79–91. Prahalad, C.K., and Venkatram Ramaswamy. 2000. “Co-opting Customer Competence.” Harvard Business Review 78 (January–February), 79–87. Roos, Inger, Anders Gustafsson, and Bo Edvardsson. 2006. “Defining Relationship Quality for CustomerDriven Business Development: A Housing-Mortgage Company Case.” International Journal of Service Industry Management 17 (2), 207–223. Ross, William T., and Diana C. Robertson. 2007. “Compound Relationships Between Firms.” Journal of Marketing 71 (3), 108–123. Rust, Roland T. 2004. “If Everything Is Service, Why Is This Happening Now, and What Difference Does It Make?” Journal of Marketing 68 (January), 23–24. ———. 2006. “From the Editor: The Maturation of Marketing as an Academic Discipline.” Journal of Marketing 70 (3), 1–2. Rust, Roland T., and Chung Tuck Siong. 2006. “Marketing Models of Service and Relationships.” Marketing Science 25 (6), 560–580. Rust, Roland T., and Debora V. Thompson. 2006. “How Does Marketing Strategy Change in a Service-Based World?: Implications and Directions for Research.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 381–392. Armonk, NY: M.E. Sharpe. Sampson, Scott E., and Craig M. Froehle. 2006. “Foundations and Implications of a Proposed Unified Services Theory.” Production and Operations Management 15 (2), 329–343. Sawhney, Mohanbir. 2006. “Going Beyond the Product: Defining, Designing, and Delivering Customer Solutions.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 365–380. Armonk, NY: M.E. Sharpe. Say, Jean B. 1821. A Treatise on the Political Economy. Boston: Wells and Lilly. Schembri, S. 2006. “Rationalizing Service Logic, or Understanding a Service Experience?” Marketing Theory 6 (3), 381–392. Schlesinger, Leonard A., and James L. Heskett. 1991. “The Service-Driven Company.” Harvard Business Review 69 (September/October), 71–81. Shostack, G. Lynn. 1977. “Breaking Free from Product Marketing.” Journal of Marketing 41 (April), 73–80. Shugan, Steven. 2004. “Finance, Operations and Marketing Conflicts in Service Firms.” Journal of Marketing 68 (January), 24–26. Shultz, Clifford J. 2007. “Marketing as Constructive Engagement.” Journal of Public Policy & Marketing 26(2), 293–301.
Service-Dominant Logic: A Review and Assessment
167
Singh, Jagdip, Jean E. Kilgore, Rama K. Jayanti, Kokil Agarwal, and Ramadesikan R. Gandarvakottai. 2005. “What Goes Around Comes Around: Understanding Trust—Value Dilemmas of Market Relationships.” Journal of Public Policy & Marketing 24 (1), 38–62. Smith, Adam. 1776/1904. An Inquiry into the Nature and Causes of the Wealth of Nations. London: W. Strahan and T. Cadell. Srinivasan, Raji, Gary L. Lilien, and Arvind Rangaswamy. 2004. “First In, First Out? The Effects of Network Externalities on Pioneer Survival.” Journal of Marketing 68 (1), 41–58. Sweeney, Jillian C. 2007. “Moving Towards the Service-Dominant Logic: A Comment.” Australasian Marketing Journal 15 (1), 97–104. Vargo, Stephen L. 2007a. “On a Theory of Markets and Marketing: From Positively Normative to Normatively Positive.” Australasian Marketing Journal 15 (1), 56–63. ———. 2007b. “Paradigms, Pluralisms, and Peripheries on the Assessment of the S-D Logic.” Australasian Marketing Journal 15 (1), 108–111. ———. 2009. “Toward a Transcending Conceptualization of Relationship: A Service-Dominant Logic Perspective.” Journal of Business and Industrial Marketing (in press). Vargo, Stephen L., and Robert F. Lusch. 2004a. “Evolving to a New Dominant Logic for Marketing.” Journal of Marketing 68 (1), 1–17. ———. 2004b. “The Four Service Marketing Myths: Remnants of a Goods-Based, Manufacturing Model.” Journal of Service Research 6 (4), 324–335. ———. 2006. “Service-Dominant Logic: What It Is, What It Is Not, What It Might Be.” In The ServiceDominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 43–56. Armonk, NY: M.E. Sharpe. ———. 2008a. “Service-Dominant Logic: Continuing the Evolution.” Journal of the Academy of Marketing Science 36 (1), 1–10. ———. 2008b. “Why ‘Service’?” Journal of the Academy of Marketing Science 36 (1), 25–38. ———. 2008c. “From Goods to Service(s): Divergences and Convergences of Logics.” Industrial Marketing Management 37 (3), 254–259. Vargo, Stephen L., Paul P. Maglio, and Melissa Archpru Akaka. 2008. “On Value and Value Co-Creation: A Service Systems and Service Logic Perspective.” European Management Journal 26 (3), 145–152. Vargo, Stephen L., and Fred W. Morgan. 2005. “Services in Society and Academic Thought: An Historical Analysis.” Journal of Macromarketing 25 (1), 42–53. Venkatesh, Alladi, Lisa Penaloza, and A. Fuat Firat. 2006. “The Marketing as a Sign System and the Logic of the Market.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 251–265. Armonk, NY: M.E. Sharpe. Wilkie, William L. and Elizabeth S. Moore. 2006. “Examining Marketing Scholarship and the ServiceDominant Logic.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 266–278. Armonk, NY: M.E. Sharpe. Winklhofer, Heidi, Adrian A. Palmer, and Roderick J. Brodie. 2007. “Researching the Service-Dominant Logic: Normative Perspective Versus Practice.” Australasian Marketing Journal 15 (1), 76–83. Womack, J.P., and D. T. Jones. 2005. “Lean Consumption.” Harvard Business Review 38 (3), 58–68. Woodruff, Robert B., and Daniel J. Flint. 2006. “Marketing’s Service-Dominant Logic and Customer Value.” In The Service-Dominant Logic of Marketing: Dialog, Debate, and Directions, ed. Robert F. Lusch and Stephen L. Vargo, 183–195. Armonk, NY: M.E. Sharpe. Xie, Chunyan, Richard P. Bagozzi, Sigurd V. Troye. 2008. “Trying to Prosume: Toward a Theory of Consumers and Co-creators of Value.” Journal of the Academy of Marketing Science 36 (1), 109–122. Yaprak, Attila, S. Tamer Cavusgil, and Destan Kandemir. 2006. “Alliance Orientation: Conceptualization, Measurement, and Impact on Market Performance.” Journal of the Academy of Marketing Science 34 (3), 324–340. Zeithaml, Valarie A., A. Parasuraman, and Leonard L. Berry. 1985. “Problems and Strategies in Services Marketing.” Journal of Marketing 49 (Spring), 33–46. Zimmermann, Erich W. 1951. World Resources and Industries. New York: Harper and Row. Zuboff, Shoshana, and James Maxmin. 2002. The Support Economy. New York: Penguin.
Chapter 7
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT Shantanu Dutta, Mark E. Bergen, and Sourav Ray
Abstract We suggest that consideration of costs of price adjustment in marketing offers a promising research direction. These costs can have substantial implications for research in pricing—from determining the magnitude and frequency of price changes to asymmetric pricing, pass-through in channels, and price synchronization. Our understanding of the nature and scope of these costs has been undergoing an evolution recently, from simple menu costs to richer decision-making, organizational, and customer-based costs. In this chapter, we review the literature in marketing and economics to summarize what we know about the nature, magnitude, and broad impact of these costs. We then identify some areas of potential interest to both researchers and practitioners in marketing, where consideration of price adjustment costs is likely to yield greater insights into marketing decisions. Introduction Pricing is one of the central decisions studied in marketing, and there is a wealth of research that has made substantial contributions to our understanding of this activity—from price sensitivity to competitive implications for pricing to pricing in channels of distribution to a variety of interesting pricing phenomena across a wide variety of markets.1 Most of this pricing research has assumed that organizations are endowed with the ability to adjust prices costlessly in response to changes in the environment—allowing prices to adjust flexibly, and allowing firms, industries, markets, and economies to function in the ways developed in classical economic theory. Indeed, this assumption is so deeply ingrained in our thinking that most of the existing literature in marketing, business, and strategy takes this ability for granted, assuming it as kind of inalienably inherent in organizations. There is, however, another perspective that assumes there are limits to the organization’s ability to adjust prices. Marketers have known about these limits for quite some time. These issues are raised in the first volumes of the Journal of Marketing, and the 1939 April volume has a section on “recent contribution to prices and price policies,” where Rufus S. Tucker (1939, p. 329) discusses the claim that “the workings of price are obscured by custom—meaning not only the conventions of accounting and business practice but especially the habits of and social standards of customers.” These limits are currently studied in the literature on the “costs of price adjustment,” which suggests that pricing can be a complex and costly organizational problem. Caplin and Leahy (1995) 168
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
169
emphasize that price adjustment is a “very difficult, costly and time-consuming process”; Levy and his coauthors (1997) suggest that changing prices “is a complex process, requiring dozens of steps and a non-trivial amount of resources”; Dutta, Zbaracki, and Bergen (2003) describe the “extraordinary complexity of the price setting process”; Zbaracki and colleagues (2004, p. 518) state that “the price change process reveals a series of managerial activities of vast scope and complexity.” Moreover, our understanding of the nature and scope of these costs has recently been undergoing an evolution in this literature. Some of the recent work has provided the first direct estimates of the magnitude of these costs, and a deeper exploration of the sources of these adjustment costs and their implications. This has moved the literature from simple menu costs to richer decisionmaking, organizational, and customer-based costs. This evolution is increasing the value of “costs of price adjustment” literature for the marketing and economics disciplines, and bringing these costs back to the roots that Tucker suggested seventy years ago in the Journal of Marketing. A key impact of these costs is price rigidity—the propensity of prices to remain unchanged in response to changes in the environment. Again, marketers have long been engaged in understanding issues related to price rigidity, such as pricing thresholds (Kalyanaram and Little 1994), price inertia (Srinivasan, Pauwels, and Nijs 2008), deal frequency (Krishna 1994), and price pass-through (Moorthy 2005). These costs can also have substantial macroeconomic implications. According to Blinder and his coauthors (1998, p. 21), costs of price adjustment have become “one of the main strands of New Keynesian theorizing,” as many predictions of traditional Keynesian and more recent New Keynesian models crucially depend on the existence of some form of price rigidity. Price rigidity is also central to microeconomic theory, ranging from theories of the firm (Coase 1937; Cyert and March 1963; Williamson 1979) to industrial organization (Carlton and Perloff 1990). According to Carlton (1986), sources of price rigidity fundamentally alter the outcomes of models in microeconomics and industrial organization—the very models that lie at the heart of our work in marketing, including such areas as pricing, channels of distribution, competitive strategy, and so forth. For marketers, this makes understanding the costs of price adjustment of fundamental interest. Despite the importance of these costs and their implications, the broader literature in marketing has long tended to ignore the role of these price adjustment costs and their relation to pricing strategy. It is only in recent years that marketing has seen a rising interest in the domain. Some recent papers that consider these costs are: asymmetric pricing (Chen et al. 2008; Ray et al. 2006), impact of item pricing laws (IPLs) on price levels (Bergen et al. 2008), price inertia (Srinivasan, Pauwels, and Nijs 2008), past price dependence (Nijs, Srinivasan, and Pauwels 2007), and price pass-through (Dutta, Bergen, and Levy 2002; Levy, Dutta, and Bergen 2002; Müller and Ray 2007; Ray et al. 2006). We believe that consideration of costly price adjustment, and the subsequent rigidity it suggests, offers a promising interdisciplinary research opportunity for the field of marketing. Empirically, marketing sits in the ideal position to study the sources and consequences of these costs. The kind of data required to make headway in our understanding of these costs requires a richer range of data sources and techniques that lie at the heart of the marketing discipline—from transaction data to surveys, interviews, fieldwork, and experiments. Theoretically, the marketing discipline has a wealth of managerial and behavioral theories that can be drawn upon to better understand these emerging organizational and customer-based sources of price adjustment and their consequences. We believe that some of the most promising future work on the costs of price adjustment and their implications can occur in the field of marketing—allowing us to inform related fields such as economics, strategy, and business. The marketing discipline itself can benefit from consider-
170
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
ation of these costs of price adjustment—to better understand patterns of price adjustment and market dynamics, as well as increasing our understanding of pricing phenomena such as pricing thresholds, asymmetric pricing, price pass-through, price points, promotion frequencies, pricing formats (such as EDLP or HI-LO), and the links between regulation, macroeconomic policy, and market prices. In this chapter we first review the relevant literature in economics and marketing about what is known about the costs of price adjustment and its impact on different pricing decisions. We then relate some issues of key interest to marketers and go on to discuss the implications for the marketing literature and future research in marketing, economics, and the interdisciplinary boundaries between the two. Literature Review We start with a broad overview of the literature pertaining to the costs of price adjustments. We first discuss the nature and scope of these costs and then move on to a discussion of the implications of these costs, primarily from a marketing perspective. For the first part, much of the literature is drawn from the “new-Keynesian” literature in economics.2 For the second part, we draw upon the economics as well as the relevant marketing literature. The Nature of Price Adjustment Costs: Sources, Magnitudes, Forms Until recently, little was actually known about what the sources and magnitudes of these costs really are. This led to doubts about the usefulness of “costs of price adjustment” in economics. For example, Prescott (1987, p. 113) suggested that theories of price adjustment costs “will never be taken seriously” until we know how to measure these costs directly. Similarly Blinder (1991, p. 90) emphasizes the importance of measuring these costs, stating that in the absence of an ability to measure them, theories of price adjustment costs “can be tested at best indirectly, at worst not at all.” And Kashyap (1995, p. 269) points out the weaknesses of the existing models of price adjustment because these models “do not explain why these [price adjustment] costs exist in the first place.” So what are these “costs” really? How do we even measure these costs? A natural and direct way to understand their size is to go to the source, organizations adjusting prices, and observe how they do this. Interestingly, early marketers also had an appreciation for the value in studying how pricing was done by managers. In fact, they were very descriptive in their methods. The earliest authors in the Journal of Marketing often described in rich detail how prices were set in practice, with an aim to inform and improve our understanding in marketing. For example, in the first volume of the Journal of Marketing, Taylor (1936) describes work being done by H.E. Agnew on the “Fundamentals of Price Making,” where the purpose of the study was to record all the influences that go into practical price making. E.T. Grether’s work (1939) eventually became a book on price control under fair trade legislation; that of P.D. Converse (1938) became an article in the Journal of Marketing that studied pricing patterns in local retail stores; and work by H.F. Taggart (1936) studying minimum prices under the National Recovery Administration (NRA). Rao (1984 S. 54) echoed a similar appreciation of the benefits of this kind of study: “the benefits of knowing more about decision processes of how industry managers go about determining (and changing) prices for their products are quite apparent.” In economics, there is early work by Hall and Hitch (1939) on how firms set and adjust prices. More recently, Blinder et al. (1998) interviewed hundreds of managers across a wide variety of
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
171
industries and asked them about possible sources of price rigidity. They found nearly half of all firms face non-trivial adjustment costs, that these varied by industry, and that these adjustment costs were perceived to be larger for price increases than price decreases. They also found that these adjustment costs “involve things other than printing new price lists, putting new price tags on goods, and so on,” and that concern with “antagonizing customers” is another important dimension of these adjustment costs.3 In a similar vein, following interviews with over 300 managers, labor leaders, and professional recruiters, Bewley (1999) finds that concerns with employee morale and willingness to support organizational objectives are important sources of wage rigidity. Other researchers have also explored price adjustment costs in different contexts. Since the 1990s, several of them have estimated these costs of adjustments by observing the pricing process at firms. Levy et al. (1997) use a detailed time and motion study of the price change processes in six major chains in the United States to document the first direct measure of the physical costs of price adjustment (the so-called menu costs) in the retail grocery industry. They estimate these costs to be over $100,000 per store per year. This translates to an average of $0.39 per price change, which is about 0.53 percent of revenues at the supermarket stores. Using the same methodology, Dutta and colleagues (1999) found the average annual menu costs per store to be almost $25,000 for the chain drugstore they studied. This corresponds to very similar marginal figures as those in the supermarket study, with estimated costs of $0.33 per price change comprising 0.59 percent of revenues. Zbaracki and his coauthors (2004) used a similar approach by deploying an ethnographic methodology to directly estimate the costs of price adjustments at a large industrial manufacturer. Using a combination of interviews, nonparticipant observations, and detailed study of internal documents, they estimated the menu costs to be almost $44,000 per year. They further estimated the marginal cost of a price change to be in the range of $0.80–$4.34, accounting for 0.71 percent of net margins. While the estimates of costs referred to in the earlier paragraphs relate to the direct physical costs of price adjustment, there have been speculations that there are managerial and customer dimensions of these costs that are possibly large components of price adjustment. As a case in point, Blinder and his coauthors (1998, ch. 13) present descriptive evidence about the significance of these elements. Ball and Mankiw (1994b, p. 142) “suspect that the most important costs of price adjustment are the time and attention required of managers to gather the relevant information and to make and implement decisions.” Blinder and colleagues (1998, pp. 313–314) suggest that “first, firms often told us—in a variety of contexts—that they are loath to change prices because this would ‘antagonize’ their customers. This imprecise thought does not fit neatly into any economists’ standard theoretical boxes, although it may be consistent with several. But it comes up so often that figuring out precisely what it means should be a high priority item on any future research agenda.” Nevertheless, direct measures of the managerial and customer dimensions of these costs are rare. To the best of our knowledge, the study by Zbaracki and his coauthors (2004) remains the only one to calibrate these. Using the ethnographic methodology referred to earlier, they estimate the “managerial costs” of price adjustments in the industrial manufacturer to be more than $280,000 annually. In terms of cost per price change, they estimate a range of $5.19–$28.05, accounting for 4.61 percent of the net margin. Interestingly, their estimate of the “customer costs” was significantly greater, at almost $900,000 annually. The marginal cost of a price change was estimated to be in the range $16.53–$121.64, for a marginal impact of 14.70 percent of net margin. When the managerial and customer costs were combined with the menu costs, the annual cost of price adjustments came to more than $1.2 million for the manufacturer, the total cost per price change being $22.52–$121.64, for a marginal impact of an astounding 20.03 percent of net margins.
172
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
While the above papers used more direct measures of price adjustment costs, another promising direction to understand these costs comes from analysis of patterns of actual price adjustments made by retailers. Slade (1998) used this approach to estimate these costs indirectly, using a structural modeling approach with discrete-choice dynamic programming. Using scanner panel data on four brands of saltine crackers across four stores, she estimated the average price adjustment costs to be about $2.72 per price change. Based on these field studies, the price adjustment costs can be seen as evolving from physical costs to a richer set of organizational and customer-based costs of price adjustment: 1. The physical (menu) costs: The direct costs of changing prices physically. For instance, in a supermarket context, Levy and his coauthors (1997) conservatively characterize the direct physical costs as being comprised of (a) labor costs of changing shelf prices, (b) the costs of printing and delivering new price tags, (c) the costs of mistakes made during the price change process, and (d) the costs of in-store supervision of the price-change process. 2. The managerial costs: This includes the managerial time and effort required to decide and implement a price change. For instance, in the industrial marketing context, Zbaracki and his colleagues (2004) found that the managerial costs comprised (a) informationgathering costs, (b) decision-making costs, and (c) communication costs. 3. The customer costs: This involves the opportunity costs of lost goodwill when customers are presented with a price change (cf. Blinder et al. 1998, p. 313). Zbaracki and coauthors (2004) characterize these as comprising of the costs associated with (a) communication and (b) negotiation with customers in the event of a price change. Even as the debate around the sources and magnitude of these costs is being addressed, there are questions about the form of these costs as well—specifically whether they are convex or fixed. Convexity in this context refers to whether price adjustment costs are a function of the magnitude of price changes. For example, Rotemberg (1982) models these costs as a quadratic function of the price change. Others like Barro (1972), Sheshinski and Weiss (1977), and Caplin and Spulber (1987) consider these to be a fixed cost. The interest in the form of these costs is not merely a definitional issue. As Blinder and his coauthors (1998) point out, convex costs could anchor multiple small price changes as opposed to larger but infrequent price changes anchored by the fixed form (p. 229). In the retail grocery context, Slade (1998) estimates the fixed components of the costs to be of much greater magnitudes (94 percent) than the convex ones (6 percent). In an industrial context, on the other hand, Zbaracki and his colleagues (2004) find that the convex components of these costs are of significantly greater magnitude than the fixed components. The Impact of Price Adjustment Costs on Prices The most direct implication of costs of price adjustment is price rigidity. As mentioned earlier, price rigidity refers to the propensity of prices to remain unchanged in response to changes in market conditions. Some of the seminal work in this area was done on patterns of price rigidity in business-to-business markets, which were carefully documented and analyzed by authors such as Stigler and Kindahl (1973) and Carlton (1986, 1989). There was also work on price rigidity for administered prices in the 1930s (Means 1935), which has been revisited over the years. For example, McRae and Tapon (1979) studied price rigidity with respect to administered pricing, which was seen to be “inflexible relative to other prices, tending to decrease less during recession and to rise less during recovery” (p. 410).
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
173
More recently, several authors (Bils and Klenow 2004; Cecchetti 1986; Dutta et al. 1999, 2002; Kashyap 1995; Levy et al. 2002; Warner and Barsky 1995) have studied more micro-level price rigidity, using even more disaggregate data at the firm and product level. For example, Cecchetti (1986) documented price rigidity in the newsstand prices of magazines, and Kashyap (1995) found price rigidity in goods sold through retail catalogs. Bils and Klenow (2004) used more disaggregated data from the Bureau of Labor Statistics to document wide-ranging levels of price rigidity across over 350 categories of goods and services. This micro-level rigidity tends to be of particular interest to marketers. In this section, we discuss four forms of price rigidity that have been addressed in the marketing literature—(1) rigidity in terms of magnitude (that is, the relative/absolute change in price levels), (2) rigidity in terms of frequency (the length of time between price changes), (3) rigidity in terms of pass-through (the proportion of upstream cost changes that is passed through as downstream price changes in a distribution channel), and (4) rigidity in terms of price synchronization (the degree to which the inter-temporal nature of rigidities is correlated across different products). As will be clear soon, despite the interest in rigidity, few papers in marketing directly relate rigidity to costs of price adjustments. Moreover, even in economics, price adjustment costs are not the only explanation of price rigidity (Blinder et al. 1998; Peltzman 2000; Bils and Klenow 2004). Therefore, we first review the relevant literature on rigidity in marketing and economics and then how the costs of price adjustment impact each of these four forms of price rigidity. Magnitudes of Price Changes: Thresholds, Asymmetric Rigidities, and Price Levels One major form of price rigidity that marketers have uncovered is the existence of price change thresholds that elicit no consumer reaction. For example, some authors (e.g., Della Bitta and Monroe 1981; Gupta and Cooper 1992) suggest that any price decrease less than 15 percent would be ignored by customers and hence not result in the desired sales bump. In other words, prices should be rigid in domains of small price changes. In a direct approach, DeSarbo et al. (1987) put forward a descriptive model of price change that incorporates such regions of rigidity and studies price-change behavior. He calibrates his “friction” model and concludes that pricing patterns are consistent with the notion that price changes occur only if the composite indices that determine prices exhibit movements beyond some high or low thresholds. And the strength of these movements, as stated in the article, has a positive relationship with the magnitude of the price change. An additional complexity is in the asymmetric rigidities implied by the inherent nature of these thresholds—both their asymmetric magnitudes (e.g., Kalyanaram and Little 1994) as well as asymmetric elasticity beyond the thresholds (Greenleaf 1995; Kalyanaram and Winer 1995; Krishnamurthi, Mazumdar, and Raj 1992). Mela, Gupta, and Lehmann (1997) discuss an additional dimension of these asymmetric thresholds—where customers may behave strategically and lie in wait for a better price promotion. The exact nature of the rigidities depends on the context. Pauwels, Srinivasan, and Franses (2007), for example, discuss the relative roles of “latitude of price acceptance” effects and “saturation” effects and how they might lead to contrasting rigidity outcomes. The implications for rigidity notwithstanding, the literature above does not explicitly consider the role of price adjustment costs. If at all, the roles of such costs are only implicit. For example, DeSarbo et al. (1987) refers to the “inert areas” in prices being possibly caused by “(the) actual implementation costs or management effort” (p. 300). A natural question, therefore, is whether these costs play any direct role in the magnitude of price changes and, especially, for such asymmetries. Chen and his coauthors (2008) find that there are far more “small” price increases than there are “small” price decreases. That is, the magnitudes of prices are far more rigid downward than
174
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
upward, but only for “small” price changes. However, such asymmetric rigidity vanishes for “large” price changes. They consider several possible sources of rigidity and speculate that this could be due to the rational inattention of consumers. If the cost-benefit tradeoff of processing the information pertaining to small price changes is unattractive, customers may ignore small price changes. This creates an incentive for the retailer to increase prices “in the small.” The existence of (asymmetric) thresholds has also been shown to exist in pricing models with costs of price adjustments. Sheshinski and Weiss (1977), for example, explain how during inflation, in the presence of costly price adjustment, firms might wait for real prices to fall to a level s, before adjusting price back up to the target price level S. Ball and Mankiw (1994a) take a similar perspective and show how inflation in the presence of these costs may lead to asymmetric price adjustments. Slade (1999) included competition in her model to illustrate and empirically demonstrate that the (s, S) thresholds may be asymmetric. Instead of inflation, however, she uses stocks of goodwill to motivate the reasons for such types of price rigidity. Similarly, Ray and colleagues (2006) show how these costs in a vertical channel of distribution may result in asymmetric price adjustments. Müller and Ray (2007) also report empirical evidence that suggests asymmetric rigidities may be consistent with a cost of price adjustment perspective. Nevertheless, there is still much to be learned about the role of price adjustment costs in asymmetric price adjustments (Peltzman 2000). Beyond the magnitude of price changes themselves, another intriguing question is whether higher costs of price adjustments result in higher prices for consumers. Bergen and his coauthors (2008) report that these costs may indeed result in higher prices for consumers. In a natural experiment, they find that retailers that are subject to item pricing laws consistently charge prices that are almost 10 percent higher than prices at retailers that are not subject to IPLs.4 This result holds even after controlling for demographic, store, and category factors. Since IPLs increase the retailer’s costs of price adjustments, this is among the few documented pieces of evidence of the link between price adjustment costs and price levels. They also find that prices in stores that invest in electronic shelf labels (ESLs) are consistently lower than prices at stores that do not. Since ESLs reduce the retailers’ marginal costs of price adjustments, this result offers additional support for the effect of these costs on prices. Frequency of Price Changes: Consumer Perceptions, Strategic Considerations, Price Formats The literature on frequency of price changes in marketing looks at consumer perceptions of frequent price changes (and the consequences for firm’s pricing) as well as the behavior of prices per se. For example, Krishna (1991) shows how consumers’ expectations of future deal frequency are positively correlated with their perceptions of past deal frequency. Consumers would therefore delay their purchases if they expect a deal in the near future. Indeed, a number of subsequent papers document the negative impact of frequent price reductions on future purchases (Krishna 1994; Mela, Gupta, and Lehmann 1997; Kopalle, Mela, and Marsh 1999), offering a logic for greater rigidity for price reductions. Other literature attempts to calibrate the inter-temporal variation of price rigidity by studying the behavior of market prices. There are several perspectives to interpret such variation. One is articulated by DeSarbo et al. (1987), who interprets price rigidities as belonging to three types—(1) univariate time-series models, which calibrate rigidities as past-price dependence, (2) experience curve effects, which calibrate rigidities to product life cycle, and (3) informational cost effects, which calibrate rigidities to the complexities of managerial decisions and efforts. Other perspectives offer more strategic interpretations. For example, in Varian’s (1980) model, firms randomly choose between high and low prices as a means to discriminate between informed
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
175
and uninformed customers. Villas-Boas (1995) found empirical support for this perspective in the coffee and saltine crackers markets. These offered no predictable patterns of rigidity. In contrast, Sobel’s (1984) model predicted greater rigidity—long periods of high prices followed by occasional deep discounts. Pesendorfer’s (2002) results find similar patterns for the ketchup category. In a further elaboration, Conlisk, Gerstner, and Sobel (1984) and Sobel (1991) offer models where sellers keep their price high early on and gradually lower prices over time. Similar to Varian’s model, this discriminates between high-valuation customers who buy early and the low-valuation customers who buy later. None of the above literature explicitly accounts for costs of price adjustments. Nevertheless, that costs of price adjustment may lead to greater inter-temporal rigidity is well acknowledged. Such costs would most certainly alter the frequencies of price changes considered above. Levy and his coauthors (1997), for example, show that supermarkets subject to IPL (hence with higher costs of price adjustment) indeed change prices less frequently than those that are not subject to IPL. The frequency of price change and the depth of price increase or decrease also depend on whether the costs of price adjustments are convex or fixed. Convexity refers to whether price adjustment costs are a function of the magnitude of price changes—in other words, the costs of changing price increases with the magnitude of price increase. When price adjustment costs are convex, it is easier for managers to undertake multiple small price changes as opposed to larger but infrequent price changes anchored by the fixed form (Blinder et al. 1998, p. 229). The empirical evidence is unclear and is likely to be context dependent. At one end of the spectrum, in the retail grocery context, Slade (1998) estimates the fixed components of the costs to be of much greater magnitudes than the convex ones. In an industrial context, on the other hand, Zbaracki and his colleagues (2004) find that the convex components of these costs are of much greater magnitude than the fixed components. This latter work offers particular insights into the contextual nature of these adjustment costs. Zbaracki and his coauthors’ (2004) paper supports the view that the organizational costs of price adjustment are convex in settings where the price adjustment process has a substantial managerial and organization component, that is, these costs are likely to increase with the size of the price change (Rotemberg 1987). The processes are likely to entail more resources when the price changes are larger—more analysis, more discussions, and more iterations. The data suggest that when the price changes necessitated by changes in market conditions were small, the firm did not have to devote too many resources—in terms of both organization members’ time and effort—to the price adjustment decisions. That is because more organization members were willing to give in and compromise when it came to relatively small price changes, even if they disagreed with the initiative. When it came to large changes, however, the company found it very costly to deal with them. The costs of the disputes, the debates, the arguments, and the disagreements that the organization was incurring under such conditions were enormous. These disagreements and disputes manifested themselves not only in various functional group meetings but also in informal settings such as during lunchtimes, in chats and conversations in the corridors and the hallways, and even in the complaints and the frustrations the various organization members would frequently take home with them. The consideration of costly price adjustment is also implicit in the choice of retail formats that guide the frequency of price changes. The Every Day Low Price (EDLP) stores position themselves as offering steady prices (greater rigidity) so that consumers can get good value regardless of when they shop. The HI-LO stores, on the other hand, offer frequent promotions (low rigidity) as a means to offer value to consumers that are willing to wait. As Hoch, Drèze, and Purk (1994)
176
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
point out, among the rationales for adopting an EDLP format are considerations of lower managerial costs “because it is easy to implement by simply matching or beating the most aggressive local competition,” as well as lower customer costs because “[its simplicity and consistency] may be easier to communicate to consumers” (p. 16). Shankar and Bolton (2004) document that EDLP stores indeed have higher rigidity in their promotional prices than the HI-LO stores. Partial and Asymmetric Pass-Through in Distribution Channels There is a large literature in marketing that deals with pass-through—the proportion of upstream cost changes that is passed through the channel in terms of downstream price changes (Moorthy 2005; Tyagi 1999). As Moorthy (2005) enumerates, such upstream cost changes can span a wide spectrum, including trade promotions, changes in regular wholesale prices, idiosyncratic changes like those in inventory positions, or even a systemwide change like changes in currency rates. In a retail context, if the retailer immediately passes through all of such cost changes as equivalent retail price changes, there is 100 percent pass-through and no rigidity. If the retailer does not change its prices immediately—or even if it does, changes by only a fraction of the cost changes—there is less than 100 percent pass-through, and prices may be said to exhibit some form of rigidity. In a bilateral monopoly (the simplest case), this happens because of double marginalization (see Tirole 1988). In analytical models of channels with a vertical information structure, incomplete pass-through is a well-known result, robust to most specifications of well-behaved demand functions. Different information structures and vertical contracts have been suggested to address this inherent source of rigidity (Ingene and Parry 1995; Jeuland and Shugan 1983; Moorthy 1988). However, none of this research explicitly assumes costs of price changes. The question then is, What happens to pass-through when there are costs of price adjustment? Nijs, Srinivasan, and Pauwels (2007) find that retail prices exhibit a fair amount of past price dependence even in the presence of wholesale price fluctuations (that is, rigidity in pass-through). They suggest that one reason for this rigidity could be related to the managerial inability to “deal with multiple objectives in the face of limited information” (p. 481), similar to the managerial costs of price adjustments discussed earlier. They go on to explore other reasons for price rigidity in retail prices. Another outcome of such costs of adjustments is argued to be asymmetric pass-through. This is defined in general as a phenomenon where cost increases are more likely to be passed through than cost decreases. Peltzman (2000) offers a summary of the literature and detailed empirical analyses to conclude (a) the evidence of asymmetric pass-through is unclear and (b) there is no definitive evidence that costs of price adjustment play a significant role in determining asymmetric pass-through. Using a more disaggregate approach however, Müller and Ray (2007) document not only evidence of asymmetric pass-through in the same data but also tentative evidence that costs of price adjustment may be playing a role. In perhaps the most direct treatment of costs of price adjustment in channel in marketing, Ray and colleagues (2006) show how the presence of price adjustment costs at the downstream retail level may create incentives of asymmetric pass-through at the upstream manufacturer level. This is because, with costs of price adjustments, retailers may have no incentive to engage in small price changes. This would reduce manufacturers’ incentives to pass through small wholesale price decreases, because they would not see any demand effect due to the rigidity of retail prices. On the other hand, it may enhance their incentive to pass through small wholesale price increases, because they invite no demand penalties due to the same rigidity of retail prices. Ray and colleagues derive equilibrium conditions and present empirical evidence of such asymmetric adjustments being limited to only “small” price changes.
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
177
Price Synchronization Across Different Product Lines Coordination of prices across the different products carried by a multiproduct firm has been of great interest to marketers. With the growing interest in category management, these have become of even greater importance (see Basuroy, Mantrala, and Walters 2001; Moorthy 2005; Zenor 1994 for more on the subject). At issue often is whether prices of different products will change at the same time (synchronized) or will change at different times (staggered). Shankar and Bolton (2004) use the term “price-promotion coordination” to empirically investigate this issue. They find that the level of synchronization across the retailer’s brands depends on a number of factors spanning competition, as well as store and product factors. While the marketing literature above does not consider the role of price adjustment costs in the synchronization of prices, it tends to be of great interest to economists because of the impact synchronization may have on aggregation of inter-temporal price rigidities. Sheshinski and Weiss (1992), for example, consider a model of price coordination under inflation and price adjustment costs. They illustrate how the extent to which prices are staggered and synchronized depends not only on the degrees of demand complementarities but also on the nature of the price adjustment costs, especially whether there are increasing returns in the adjustment costs. They point to the critical role the form of such costs can play as well—that is, whether the costs are mostly the physical costs or have a large component of “decision costs” similar to the managerial costs discussed earlier. Discussion We hope this chapter helps to serve as a call to the marketing discipline to focus more attention on price adjustment costs and their implications in the price-setting process. The time is right, as this literature evolves from physical (menu) costs to richer organizational and behavioral costs, and broadens its horizons to consider richer sources of micro-pricing data. These shifts suggest that marketing may have a great deal to offer, as well as learn from, this developing literature on the costs of price adjustments. Among the business disciplines, marketing is perhaps best situated to make headway on the critical issues resting at the heart of this literature. The marketing discipline has a broad range of data sources, empirical tools, and models that can be used to contribute to the work on price adjustment costs. Empirically this ranges from a long history in ethnographic methods, to direct access to managers and firms, to survey methods to develop the scale and scope of price rigidity and its sources, to large pricing data sets and emerging empirical techniques that could study price rigidity and its antecedents. Theoretically, this ranges from models of consumers managers, channels, organizations, and strategy that may be effective ways to model these sources of price rigidity and their implications for markets. Given the central role of managers in price adjustment that is arising in the literature, our marketing understanding of organizations and how they make business decisions offers an opportunity to make fundamental contributions to price theory and all the disciplines that rely on it. Marketing also has a lot to learn from the literature on the costs of price adjustment and price rigidity. Incorporating these costs of adjustment and price rigidities into our thinking and models enlarges our toolkit for understanding the role and limitations of pricing in the marketplace, and improves our guidance on difficult managerial problems. Given marketers’ keen interest in price changes, it presents a particularly fertile area of research with potentially significant payoffs for theory as well as for practice. We identify some of these areas and discuss them in the following paragraphs.
178
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
Need to Know More About the Nature of These Costs and Their Implications Take first the basic issues surrounding the nature and scope of price adjustment costs. Given the nature of these costs per se, our ability to easily measure and categorize these costs has been quite limited. As future researchers attempt to bridge this knowledge gap, we see a crucial role for more descriptive studies (Zbaracki et al. 2004) alongside structural models like that of Slade (1998), and we encourage the use of a wide variety of marketing tools and techniques to gather more information about these costs and their consequences for price rigidity. This empirical research should focus on the nature of these costs of price adjustment. Despite the growing interest in the area and increasing evidence of the significance and importance of these costs, we are just beginning to gain an understanding of their sources, magnitudes, and forms. This is especially true of the emerging organizational and customer-based dimensions of these costs, which offer some of the most significant and interesting sources of price rigidity. This is also true of convexity of these adjustment costs. While a number of authors (e.g., Blinder et al. 1998; Rotemberg 1982; Sheshinski and Weiss 1977; Zbaracki et al. 2004) illustrate the important role of such convexity, only a few (e.g., Slade 1998; Zbaracki et al. 2004) even attempt to calibrate them. For marketers this is an area of interest. Perhaps more than any other marketing mix decision, magnitudes and frequencies of price changes frame the marketer’s response to the dynamics of consumer tastes and competition. Given the impact that convexity can have on the magnitude and frequency of price changes (Blinder et al. 1998), a greater understanding of the form of price adjustment costs is a promising area to explore. Marketing offers a wealth of behavioral theory and knowledge that can be used to unlock the nature of these organizational, customer, and convexity issues, as well as a rich array of empirical methodologies to uncover the empirical significance of these as sources of price rigidity. In terms of the scope of these costs, while there is some empirical work that estimates these costs in retail supermarkets (Levy et al. 1997; Slade 1998; Dutta et al. 1999), there is not much research in other contexts. Zbaracki and coauthors (2004) are among the few that study price adjustment costs in an industrial context. More empirical research into the nature of these costs in various contexts is necessary to create a more robust framework. In fact, there may be substantial variation in price rigidity patterns across product sizes, categories, brands, and channels (Gordon 1990). Different sectors vary in factors that can impact price adjustments. This could be not only on the basis of end-products (for example, manufacturing versus services) but also in the nature of channels—the diversity of members, number of different levels, degrees of decentralization, and so forth. Newer technological advances (such as ERP systems, RFID), practices (for example, category management and outsourcing), and economic realities (such as globalization) may also fundamentally impact the costs of price adjustments and hence the way prices are set in diverse sectors. As an example, Nijs, Srinivasan, and Pauwels (2007) explore variation in past price dependence and possible explanations for these patterns of price rigidity. There is much to be learned from documenting these variations in price rigidity and understanding the explanations for this rigidity (Bils and Klenow 2004; Dutta, Bergen, and Levy 2002; Nijs, Srinivasan, and Pauwels 2007; Stigler and Kindahl 1970). Therefore we call for a sustained effort to map the nature of these costs and their consequences in different contexts—such as B2B markets, service industries, international markets, throughout entire supply chains, and so on. Drawing Insights from Economic Modeling Price rigidity has been widely studied in the macroeconomics literature (Sheshinski and Weiss 1977; Gordon 1990; Ball and Mankiw 1994a, b; Carlton and Perloff 1990). This literature has been
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
179
developing to incorporate the evolving understanding of the costs of price adjustment, offering new ways to model price rigidity. For example, Mankiw and Reis (2002) and Ball, Mankiw, and Reis (2005) explore how modeling the price adjustment costs as the costs of managerial decision can yield a more plausible Phillips Curve relation. There is also a literature exploring the use of organizational rules of thumb and their implications for economic activity (Amato and Laubach 2003; Galí and Gertler 1999). Rotemberg (2005) explores the implications of customer costs of price adjustment for macroeconomic policy. These costs can also have implications for the debate on time-dependent versus state-dependent models of nominal price rigidity (Basu 2005).5 This organizational evolution also suggests exploring the span of control models (Lucas 1978; Prescott and Visscher 1980); recasting these toward the ability of firms to adjust prices (as well as manage production and other activities) may be a promising direction for marketers to explore. All of these literatures provide interesting theoretical insights into the complexities and implications of these adjustment costs—which can be of substantial benefit to the marketing literature as it explores these costs in the future and their implications for pass-through, price tiers, and other dynamic pricing issues. Managerially, one can also view the emerging literature on six sigma pricing and other process improvement methodologies being brought into the field of pricing as outgrowths of this deeper understanding of the costs involved in price adjustment. For example, the work of Sodhi and Sodhi (2008) articulates a view of pricing processes that is fraught with costs, and it helps managers apply six sigma and other statistical and process control techniques to improve the efficiency and effectiveness of their pricing processes. Small Versus Large Price Changes We can benefit from a greater understanding of how price adjustment costs affect not only large but also small price changes (Kashyap 1995). There is an intriguing juxtaposition of perspectives in the domain of small price changes. The presence of frequent small price changes in the presence of price adjustment costs has been puzzling to some economists like Kashyap (1995). The relation between small price changes and the convexity of price adjustment costs has been speculated upon as a possible explanation (Blinder et al. 1998; Kashyap 1995; Zbaracki et al. 2004). On the other hand, for marketers, small price changes are intuitively appealing given a long history of Just Noticeable Differences (JND) and related literature (Monroe 1973; Lichtenstein, Bloch, and Black 1988; Kalyanaram and Little 1994). Ray and colleagues (2006) and Lach and Tsiddon (2007) are among the few that apply a price adjustment perspective to small price adjustments. We believe greater attention to how small price changes and price adjustment costs relate to each other will yield a rich set of insights for both theory and practice. The discussion above illustrates that when changing prices, managers have to take into account the regions of price changes to which customers pay less attention. As price changes beyond those ranges, especially for price increases, this large price change can increase the cost that a firm has to incur to convince the customers that the new price still provides value to them. The firm therefore may have to pay more attention to the importance of customer costs. The price rigidities in these thresholds can then be explained by the costs that customers and managers incur in an effort to move prices beyond the thresholds. This is supported by evidence found in the industrial context, where the firm incurs significant managerial and customer costs associated with a pricing decision—especially if they have to be large price changes (Zbaracki et al. 2004). It also points to how marketers in certain contexts might deal with the competing concerns of costly price changes and the business prerogative of making a temporary price change. As documented in Levy and his coauthors (1997, 1998) and
180
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
Dutta and colleagues (1999), in the supermarkets and drugstores they study, a vast majority of the price changes were temporary promotions, not the more permanent changes in the list or regular prices. Presumably, changes in the list prices (given their more permanent nature) would hold more significant long-term implications for the firm and hence would require greater consideration and higher managerial costs of price adjustment. Frequency of Promotional Price Changes There are a number of perspectives that drive the frequency of price changes. One of the more prominent ones is store positioning, broadly categorized as EDLP or HI-LO. Despite the general interest in these forms (Bell and Lattin 1998; Hoch, Drèze, and Purk 1994), few papers exist to study the implications of price adjustment costs for such pricing formats. While EDLP stores engage in few promotional price changes, HI-LO stores undertake frequent promotions. As such, these formats present an opportunity to understand how price adjustment costs may operate in two very different scenarios. Deal sensitivity is yet another perspective that drives frequency of price changes (Kopalle, Mela, and Marsh 1999; Krishna 1994; Mela, Gupta, and Lehmann 1997). The idea here is that frequent price changes may lead consumers to buy only when there is a sufficiently attractive price promotion. This may not be good news for the retailer’s bottom line. Nevertheless, promotional pricing is often interpreted from a competition lens. What is not clear is how the existence of price-adjustment costs across different competitors may impact the outcomes regarding frequency of price changes. This is another area where marketing is uniquely poised to contribute to our understanding of how price adjustment costs affect prices. Managerially, adjustment costs offer a new set of variables to consider when setting broader pricing strategies such as an EDLP strategy, than the ones suggested in the existing marketing literature (e.g., Hoch, Drèze, and Purk 1994; Lal and Rao 1997; Bell and Lattin 1998; Bell, Ho, and Tang 1998; Ailawadi, Lehmann, and Neslin 2001). The existence of these costs of price adjustment suggests that in the short run any complex pricing scheme like bundling or usage-based pricing will have to take into account the firm’s existing costs to adjust prices, and consider the costs to change the strategy itself. Essentially this suggests another level of costs of price adjustment—adjustment in the form of pricing—as a promising research direction to explore. Asymmetric Price Adjustments Chen and his coauthors (2008) show there is an asymmetry between positive and negative retail price changes in the small—with small increases vastly outnumbering small price decreases. This asymmetry however, vanishes for large changes. Such asymmetric price adjustments have always been of great interest. Nevertheless, per Peltzman (2000), there are as yet no compelling explanations of the phenomenon. While it has always been part of the consideration set, the empirical evidence of the role of price adjustment costs on asymmetric adjustments are mixed in both the economics and marketing literature (Ball and Mankiw 1994a; Müller and Ray 2007; Peltzman 2000; Ray et al. 2006). Certainly, the outcomes would be intuitive if the costs themselves were asymmetric, that is, price decreases were more costly than price increases. However, even when asymmetric costs were referred to (for example, in Blinder et al. 1998; Zbaracki et al. 2004), they have usually been in different directions (price increases more costly than price decreases). Interestingly, some recent work shows that even if these costs per se are not asymmetric, strategic interaction between market players could still lead to asymmetric adjustments (see Ray et al. 2006). There are only a handful of papers in this area. Among them, Slade (1999) looks at how price adjust-
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
181
ment costs impact horizontal interactions (competition), and Basu (1995) looks at how price adjustment costs impact pricing and production outcomes in a vertical context (channels). A related literature from behavioral economics offers another lens to interpret asymmetric price adjustments, whose existence as a retail pricing practice is often puzzling from a consumer fairness perspective. Certainly, if the practice was considered patently unfair, the firm’s ability to profit from it would be limited. To this end, Kahneman, Knetsch, and Thaler (1986a, 1986b) propose the “dual-entitlement principle” to explain why such asymmetric pricing may still be considered fair by consumers and hence sustainable from the firm’s point of view. However, their perspective anchors on the manner in which prices respond to input costs but make no explicit allowance for the costs of price adjustments or the retailer’s ability to make price changes. Despite the clear managerial implications, there is little to no research that explores this area. Given the rich set of tools and perspectives available to marketers to study consumer, channel, and competitive interactions, we believe marketing is uniquely situated to yield valuable insights for both theory and practice in this domain. These insights could draw not only from formal economicoriented approaches but also from more cognitive and behaviorally oriented methodologies. Price Points and Price Rigidity A similar line of reasoning can also be applied to the interpretation of price point rigidities (Levy et al. 2007).The phenomenon of nine cent ending, for example, has been of continuing interest to both economists and marketers (see Basu 1997, 2006; Schindler and Kirby 1997; Stiving and Winer 1997; Twedt 1965). The traditional explanation of nine-cent endings suggests that retailers strategically leverage the limits of consumers’ cognitive abilities. Nevertheless, the widespread deployment of such price points may in fact point to a pricing routine at the retail level as opposed to a more deliberate strategic decision (see the work on “customary prices” [Ginzberg 1936]). This would also suggest that moving from nine-cent endings could have its own organizational costs. Hence we call for additional research to expand our understanding of price point rigidities by incorporating costs of price adjustments. Price Coordination Across Competition and Across Brands Price coordination has always been a major consideration for marketers. At an inter-retailer level, such coordination can range from explicit collusion to simple price-follower strategy, where a retailer merely readjusts its prices following the price leader’s adjustment. In another variation of this, a nonstrategic cost plus rule may see all retailers in the industry automatically adjust prices following an industry-wide cost shock (as is often observed in retail gasoline prices). It seems intuitive that to the extent price adjustment costs impact rigidity of prices, they would also impact the type of price coordination being referred to here. Nevertheless, it is not clear what would be the nature of such an impact. Certainly, if the price-change process is routinized to the extent that no new decisions are involved at the time of a price change, the outcome could be vastly different than if each change point was an occasion for a new decision. Research on price competition and competitive response in marketing has looked at the nature of competitive interactions using scanner data (e.g., Kadiyali, Vilcassim, and Chintagunta 1999; Leeflang and Wittink 1996) or game theoretic models (e.g., Moorthy 1985). The existence of these costs of price adjustment suggests more subtle modeling and empirical issues as to how they affect competition. To the best of our knowledge, there is little work in this area, and we believe that the marketing literature will greatly benefit from more research in this domain.
182
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
The nature of price coordination may differ as well. At the inter-brand level, a single retailer may consider all the brands in a category collectively in its price adjustment decisions. The outcome could be synchronized price changes (multiple brand prices change at the same time) or staggered price changes (the price changes do not happen simultaneously). Such price coordination is considered a key part of category management (see Basuroy, Mantrala, and Walters 2001; Shankar and Bolton 2004). Nijs, Srinivasan, and Pauwels (2007) document the profit impact of a category management perspective. Costs of price adjustment have a particular resonance to category management. It is not clear whether the complexity of the decision making would be necessarily greater than a more brand-management focus. However, one would think that the nature of the costs would be a critical consideration. In particular, if there are such economies as increasing returns to scale to these costs, retailers would be more prone to coordinate price changes across multiple brands. At the same time, it is unclear if these economies would necessarily lead to price coordination in the sense of synchronization. For example, it may be possible for retailers to incur a fixed decision cost that would dictate whether the synchronized or the staggered patterns would be deployed. As Sheshinski and Weiss (1992) point out, price synchronization (or lack thereof) could very well depend on whether the nature of the costs is dominated by the physical or managerial components. Our perspective also suggests that in the short run, the ability of manufacturers or retailers to rely on pricing strategies to engender tacit collusion may depend on these costs. For instance, Lal (1990) suggests that trade promotions can be an instrument for tacit collusion across national brand manufacturers. This assumes that all the firms do not face substantial costs of price adjustment; otherwise this may not be feasible. Similarly, the mechanism of price matching guarantees to engender tacit collusion (Hess and Gerstner 1991; Chen, Narasimhan, and Zhang 2001) relies on minimal costs of adjustment to make this strategy feasible. Despite the apparent implications for pricing decisions, we know very little about the role of price adjustment costs in the price coordination decisions. We therefore call for more investigations— empirical research to map the descriptive parameters of the relationships, as well as theoretical research to create a framework to interpret the direct and indirect effects of these costs on price coordination outcomes. Interaction with Other Marketing Mix Adjustments In the preceding paragraphs, we have highlighted some of the key tactical domains where price adjustment costs may impact the marketer’s pricing decisions. But pricing is only a part of the broader marketing strategy and cannot be calibrated in isolation. The notion that the marketing mix variables are correlated is well accepted in marketing. Given the historically scant attention given to price adjustment costs in marketing, our knowledge of how price adjustment costs interact with the other marketing variables is also very limited. Specifically, how price adjustment costs impact rigidities in product introductions/innovations and/or channel structure is not explored. It stands to reason that if price is a main marketing tool in the value exchange process, any rigidity in prices would impact other marketing mix decisions. Examples of such interactions abound. When Virgin Mobile introduced their new product in America, the key value proposed was a noncomplex, no-contract pricing plan. The more established incumbents in the mobile market were not able to change their price format to compete with Virgin because of the managerial and customer costs they would have had to incur to make that change. Similarly, changes from HI-LO pricing to EDLP pricing may call for major changes in the channel arrangements (for example, EDLPP [Every Day Low Purchase Price]; see Hoch, Drèze, and Purk 1994). As Sears found out much to its grief, such a move is not just a matter of changing price forms.
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
183
Conclusion In this chapter, we review the literature in marketing and economics to summarize what we know about the nature, magnitude, and broad impact of price adjustment costs. We then identify some areas of pricing that are of particular interest to marketers where considerations of price adjustment costs are likely to yield insight. Our basic conclusion is that there are significant domains of pricing decisions that are under-researched from the perspective of price adjustment costs. We believe more explicit consideration of these costs will lead not only to greater understanding of pricing but also to better pricing decisions. We also believe that this offers an exciting interdisciplinary research opportunity of fundamental importance to the business and economics disciplines, where marketers can play a leading role. Therefore, we urge marketers to pay more attention to the evolving literature on the costs of price adjustment. At the same time, we do not mean to suggest that the study of these costs is a panacea or the only route to study pricing in the marketing literature. Many issues are better studied without consideration of these complexities. Moreover, while price rigidity is a key outcome of price adjustment costs, these costs are only one explanation for price rigidity in the literature. Our goal here is to generate consideration of these costs and their implications for researchers interested in studying the dynamics of pricing. We feel the research in marketing, economics, and many related business disciplines will benefit greatly from such an approach. Acknowledgments The authors would like to thank the anonymous participants in their research, whose generous cooperation made this study possible. Special thanks are due to Haipeng (Allan) Chen and Neil Bendle for their thoughtful feedback on the manuscript. The authors would also like to thank Mark Ritson, Daniel Levy, and Mark Zbaracki for their active participation in this line of research, as well as Padmasri Suriyakumar and Kulothungan Jeganathan of McMaster University, and Benjamin Lehrman of the University of Pittsburgh, for their able research assistance. The authors rotate order of coauthorship. All authors contributed equally. Notes 1. See Monroe 1973; Rao 1984; Moorthy 1985; Wilson, Weiss, and John 1990; Hoch Drèze, and Purk 1994; Wernerfelt 1994; Simester 1995; Lal and Rao 1997; Bell, Ho, and Tang 1998; Ailawadi, Lehmann, and Neslin 2001; Stremersch and Tellis 2002. 2. See Blinder and his coauthors (1998), Ball and Mankiw (1994a, 1994b), Rotemberg (1987), Gordon (1990), and Mankiw and Romer (1991) for a comprehensive discussion of this “New Keynesian” literature. 3. See Blinder and his coauthors (1998) for an excellent presentation of the importance of price rigidity for economics, the main theories of price rigidity, more results on costs of price adjustment, and a wealth of other findings on price rigidity. 4. In their most basic form, item pricing laws (IPLs) require a price tag on every item sold by a retailer. Currently, IPLs exist in nine U.S. states; in Quebec, Canada; in some European countries; and in Israel. 5. See also Sheshinski and Weiss (1977).
References Ailawadi, Kusum L., Donald R. Lehmann, and Scott A. Neslin. 2001. “Market Response to a Major Policy Change in the Marketing Mix: Learning from Procter & Gamble’s Value Pricing Strategy.” Journal of Marketing 65 (1), 44–61.
184
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
Amato, Jeffery D., and Thomas Laubach. 2003. “Rule-of-Thumb Behaviour and Monetary Policy.” European Economic Review 47 (5), 791–831. Ball, Laurence, and N.G. Mankiw. 1994a. “Asymmetric Price Adjustment and Economic Fluctuations.” Economic Journal 104, 247–261. ———. 1994b. “A Sticky-Price Manifesto.” Carnegie-Rochester Conference Series on Public Policy, 41, 127–151. Ball, Laurence, N. G. Mankiw, and Ricardo Reis. 2005. “Monetary Policy for Inattentive Economies.” Journal of Monetary Economics 52 (4), 703–725. Barro, Robert J. 1972. “A Theory of Monopolistic Price Adjustment.” Review of Economic Studies 39 (117), 17–26. Basu, Kaushik. 1997. “Why Are So Many Goods Priced to End in Nine? And Why This Practice Hurts the Producers.” Economics Letters 54 (1), 41–44. ———. 2006. “Consumer Cognition and Pricing in the Nines in Oligopolistic Markets.” Journal of Economics and Management Strategy 15 (1), 125–141. Basu, Susanto. 1995. “Intermediate Goods and Business Cycles: Implications for Productivity and Welfare.” American Economic Review 85 (3), 512–531. ———. 2005. “Comment on: ‘Implications of State-Dependent Pricing for Dynamic Macroeconomic Modeling.’” Journal of Monetary Economics 52 (1), 243–247. Basuroy, Suman, Murali K. Mantrala, and Rockney G. Walters. 2001. “The Impact of Category Management on Retailer Prices and Performance: Theory and Evidence.” Journal of Marketing 65 (4), 16–32. Bell, David R., Teck-Hua Ho, and Christopher S. Tang. 1998. “Determining Where to Shop: Fixed and Variable Costs of Shopping.” Journal of Marketing Research 35 (3), 352–369. Bell, David R., and James M. Lattin. 1998. “Shopping Behavior and Consumer Preference for Store Price Format: Why ‘Large Basket’ Shoppers Prefer EDLP.” Marketing Science 17, 66–88. Bergen, Mark, Daniel Levy, Sourav Ray, Paul Rubin, and Benjamin Zeliger. 2008. “When Little Things Mean a Lot: On the Inefficiency of Item Pricing Laws.” Journal of Law and Economics 51 (2), 209–250. Bewley, Truman F. 1999. Why Wages Don’t Fall During a Recession. Boston, MA: Harvard University Press. Bils, Mark, and Peter J. Klenow. 2004. “Some Evidence on the Importance of Sticky Prices.” Journal of Political Economy 112 (5), 947–985. Blinder, Alan S. 1991. “Why Are Prices Sticky? Preliminary Results from an Interview Study.” American Economic Review 81, 89–100. Blinder, Alan S., Elie R.D. Canetti, David E. Lebow, Jeremy B. Rudd. 1998. Asking About Prices: A New Approach to Understanding Price Stickiness. New York: Russell Sage Foundation. Caplin, Andrew, and John Leahy. 1995. “The Economics of Adjustment.” In The Natural Rate of Unemployment: Reflections on 25 Years of the Hypothesis, ed. Rod Cross, 77–89. New York: Cambridge University Press. Caplin, Andrew S., and Daniel F. Spulber. 1987. “Menu Costs and the Neutrality of Money.” Quarterly Journal of Economics 102, 703–725. Carlton, Dennis W. 1986. “The Rigidity of Prices.” American Economic Review 76, 637–658. ———. 1989. “The Theory and Facts About How Markets Clear: Is Industrial Organization Valuable for Understanding Macroeconomics?” Handbook of Industrial Organization 1, 909–946 (chapter 15). Carlton, Dennis W., and Jeffrey M. Perloff. 1990. Modern Industrial Organization. Glenville, IL: Scott, Foresman/Little Brown Higher Education. Cecchetti, Stephen G. 1986. “The Frequency of Price Adjustment: A Study of the Newsstand Prices of Magazines.” Journal of Econometrics 31 (3), 255–274. Chen, Haipeng (Allan), Daniel Levy, Sourav Ray, and Mark Bergen. 2008. “Asymmetric Price Adjustment in the Small.” Journal of Monetary Economics 55 (4), 728–737. Chen, Yuxin, Chakravarthi Narasimhan, and Z. J. Zhang. 2001. “Research Note: Consumer Heterogeneity and Competitive Price-Matching Guarantees.” Marketing Science 20 (3), 300. Coase, Ronald. 1937. “The Nature of the Firm.” Economica 4 (16), 386–405. Conlisk, John, Eitan Gerstner, and Joel Sobel. 1984. “Cyclic Pricing by a Durable Goods Monopolist.” Quarterly Journal of Economics 99(3), 489–505. Converse, S. 1938. “Prices and Services of Chain and Independent Stores in Champaign, Urbana, 1937.” Journal of Marketing 2 (3) (January), 193–200. Cyert, Richard M., and James G. March. 1963. A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall.
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
185
Della Bitta, Albert J., and Kent B. Monroe. 1981. “A Multivariate Analysis of the Perception of Value from Retail Price Advertisements.” Advances in Consumer Research 8 (1), 161–165. DeSarbo, Wayne S. Vithala R. Rao, Joel H. Steckel, Jerry Wind, Richard Colombo, 1987. “A Friction Model for Describing and Forecasting Price Changes.” Marketing Science 6(4), 299–319. Dutta, Shantanu, Mark Bergen, and Daniel Levy. 2002. “Price Flexibility in Channels of Distribution: Evidence from Scanner Data.” Journal of Economic Dynamics & Control 26 (11), 1845–1900. Dutta, Shantanu, Mark Bergen, Daniel Levy, and Robert Venable. 1999. “Menu Costs, Posted Prices, and Multiproduct Retailers.” Journal of Money, Credit & Banking 31 (4), 683–703. Dutta, Shantanu, Mark J. Zbaracki, and Mark Bergen. 2003. “Pricing Process as a Capability: A ResourceBased Perspective.” Strategic Management Journal 24 (7), 615–630. Gali, Jordi, and Mark Gertler. 1999. “Inflation Dynamics: A Structural Econometric Analysis.” Journal of Monetary Economics 44 (2), 195–222. Ginzberg, Eli. 1936. “Customary Prices.” American Economic Review 26 (2), 296–296. Gordon, Robert J. 1990. “What Is New-Keynesian Economics?” Journal of Economic Literature 28, 1115–1171. Greenleaf, Eric A. 1995. “The Impact of Reference Price Effects on the Profitability of Price Promotions.” Marketing Science 14, 82–104. Grether, Ewald T. 1939. Price Control Under Fair Trade Legislation. New York: Oxford University Press. Gupta, Sunil, and Lee G. Cooper. 1992. “The Discounting of Discounts and Promotion Thresholds.” Journal of Consumer Research 19 (3), 401–411. Hall, R. L., and C. J. Hitch. 1939. “Price Theory and Business Behavior.” In Oxford Economic Papers 2 (May). Oxford: Clarendon Press. Hess, James D., and Eitan Gerstner. 1991. “Price-Matching Policies: An Empirical Case.” Managerial & Decision Economics 12 (4), 305–315. Hoch, Stephen J., Xavier Drèze, and Mary E. Purk. 1994. “EDLP, Hi-Lo, and Margin Arithmetic.” Journal of Marketing 58 (4), 16–27. Ingene, Charles A., and Mark E. Parry. 1995. “Channel Coordination When Retailers Compete.” Marketing Science 14, 360–377. Jeuland, Abel P., and Steven M. Shugan. 1983. “Managing Channel Profits.” Marketing Science 2 (3), 239–272. Kadiyali, Vrinda, Naufel Vilcassim, and Pradeep Chintagunta. 1999. “Product Line Extensions and Competitive Market Interactions: An Empirical Analysis.” Journal of Econometrics 89 (1), 339–363. Kahneman, Daniel, Jack L. Knetsch, and Richard Thaler. 1986a. “Fairness as a Constraint on Profit Seeking: Entitlements in the Market.” American Economic Review 76 (4), 728–741. ———. 1986b. “Fairness and the Assumptions of Economics.” Journal of Business 59 (4), S285–300. Kalyanaram, Gurumurthy, and John D.C. Little. 1994. “An Empirical Analysis of Latitude of Price Acceptance in Consumer Package Goods.” Journal of Consumer Research 21 (3), 408–418. Kalyanaram, Gurumurthy, and Russell S. Winer. 1995. “Empirical Generalizations from Reference Price Research.” Marketing Science 14, 161–169. Kashyap, Anil K. 1995. “Sticky Prices: New Evidence from Retail Catalogs.” Quarterly Journal of Economics 110, 245–274. Kopalle, Praveen K., Carl F. Mela, and Lawrence Marsh. 1999. “The Dynamic Effect of Discounting on Sales: Empirical Analysis and Normative Pricing Implications.” Marketing Science 18 (3), 317–332. Krishna, Aradhna. 1991. “Effect of Dealing Patterns on Consumer Perceptions of Deal Frequency and Willingness to Pay.” Journal of Marketing Research 28, 441–451. ———. 1994. “The Impact of Dealing Patterns on Purchase Behavior.” Marketing Science 13, 351–373. Krishnamurthi, Lakshman, Tridib Mazumdar, and S.P. Raj. 1992. “Asymmetric Response to Price in Consumer Brand Choice and Purchase Quantity Decisions.” Journal of Consumer Research 19, 387–400. Lach, Saul, and Daniel Tsiddon. 2007. “Small Price Changes and Menu Costs.” Managerial & Decision Economics 28 (7), 649–656. Lal, Rajiv. 1990. “Price Promotions: Limiting Competitive Encroachment.” Marketing Science 9 (3), 247. Lal, Rajiv, and Ram Rao. 1997. “Supermarket Competition: The Case of Every Day Low Pricing.” Marketing Science 16 (1), 60. Leeflang, Peter S. H., and Dick R. Wittink. 1996. “Competitive Reaction Versus Consumer Response: Do Managers Overreact?” International Journal of Research in Marketing 13 (2), 103–119.
186
Shantanu Dutta, Mark E. Bergen, and Sourav Ray
Levy, Daniel, Mark Bergen, Shantanu Dutta, and Robert Venable. 1997. “The Magnitude of Menu Costs: Direct Evidence from Large U.S. Supermarket Chains.” Quarterly Journal of Economics 112, 791–825. Levy, Daniel, Shantanu Dutta, and Mark Bergen. 2002. “Heterogeneity in Price Rigidity: Evidence from a Case Study Using Microlevel Data.” Journal of Money, Credit & Banking 34 (1), 197–200. Levy, Daniel, Shantanu Dutta, Mark Bergen, and Robert Venable. 1998. “Price Adjustment at Multiproduct Retailers.” Managerial & Decision Economics 19, 81–120. Levy, Daniel, Dongwon Lee, Haipeng Chen, Robert J. Kauffman, and Mark Bergen. 2007. “Price Points and Price Rigidity.” Working paper, Carlson School of Management, University of Minnesota. Lichtenstein, Donald R., Peter H. Bloch, and William C. Black. 1988. “Correlates of Price Acceptability.” Journal of Consumer Research 15 (2), 243–252. Lucas, Robert E. 1978. “On the Size Distribution of Business Firms.” Bell Journal of Economics 9 (Autumn), 508–523. Mankiw, N. G., and Ricardo Reis. 2002. “Sticky Information Versus Sticky Prices: A Proposal to Replace the New Keynesian Phillips Curve.” Quarterly Journal of Economics 117 (4), 1295–1328. Mankiw, N. G., and David Romer, eds. 1991. New Keynesian Economics. Vol. 1. Imperfect Competition and Sticky Prices. Cambridge, MA: MIT Press. McRae, James J., and Francis Tapon. 1979. “A New Test of the Administered Pricing Hypothesis with Canadian Data.” Journal of Business 52 (3), 409–427. Means, Gardiner C. 1935. “Industrial Prices and Their Relative Inflexibility.” U.S. Senate Document 13, 74th Congress, 1st Session, Washington, DC. Mela, Carl F., Sunil Gupta, and Donald R. Lehmann. 1997. “The Long-Term Impact of Promotion and Advertising on Consumer Brand Choice.” Journal of Marketing Research 34, 248–261. Monroe, Kent B. 1973. “Buyers’ Subjective Perceptions of Price.” Journal of Marketing Research 10 (1), 70–80. Moorthy, K.S. 1985. “Using Game Theory to Model Competition.” Journal of Marketing Research 22 (3), 262–282. ———. 1988. “Strategic Decentralization in Channels.” Marketing Science 7 (4), 335–355. Moorthy, Sridhar. 2005. “A General Theory of Pass-Through in Channels with Category Management and Retail Competition.” Marketing Science 24 (1), 110–122. Müller, Georg, and Sourav Ray. 2007. “Asymmetric Price Adjustment: Evidence from Weekly Product-Level Scanner Price Data.” Managerial & Decision Economics 28 (7), 723–736. Nijs, Vincent R., Shuba Srinivasan, and Koen Pauwels. 2007. “Retail-Price Drivers and Retailer Profits.” Marketing Science 26 (4), 473–487. Pauwels, Koen, Shuba Srinivasan, and Philip H. Franses. 2007. “When Do Price Thresholds Matter in Retail Categories?” Marketing Science 26 (1), 83–100. Peltzman, Sam. 2000. “Prices Rise Faster than They Fall.” Journal of Political Economy 108 (3), 466– 502. Pesendorfer, Martin. 2002. “Retail Sales: A Study of Pricing Behavior in Supermarkets.” Journal of Business 75 (1), 33–66. Prescott, Edward C. 1987. “The New Keynesian Microfoundations: Comment.” In NBER Macroeconomics Annual: 1987, ed. Stanley Fischer, 110–114. Cambridge, MA. and London: MIT Press. Prescott, Edward C., and Michael Visscher. 1980. “Organization Capital.” Journal of Political Economy 88 (3), 446–461. Rao, Vithala R. 1984. “Pricing Research in Marketing: The State of the Art.” Journal of Business 57(1), S39–60. Ray, Sourav, Haipeng Chen, Mark E. Bergen, and Daniel Levy. 2006. “Asymmetric Wholesale Pricing: Theory and Evidence.” Marketing Science 25(2), 131–154. Rotemberg, J.J. 1982. “Sticky Prices in the United States.” Journal of Political Economy 90, 1187–1211. Rotemberg, Julio J. 1987. “The New Keynesian Microfoundations.” In NBER Macroeconomics Annual: 1987, ed. Stanley Fischer, 69–104. Cambridge, MA: MIT Press. ———. 2005. “Customer Anger at Price Increases, Changes in the Frequency of Price Adjustment and Monetary Policy.” Journal of Monetary Economics 52 (4), 829–852. Schindler, Robert M., and Patrick N. Kirby. 1997. “Patterns of Rightmost Digits Used in Advertised Prices: Implications for Nine-Ending Effects.” Journal of Consumer Research 24, 192–201. Shankar, Venkatesh, and Ruth N. Bolton. 2004. “An Empirical Analysis of Determinants of Retailer Pricing Strategy.” Marketing Science 23 (1), 28–49.
MARKETING IN A WORLD WITH COSTS OF PRICE ADJUSTMENT
187
Sheshinski, Eytan, and Yoram Weiss. 1977. “Inflation and Costs of Price Adjustment.” Review of Economic Studies 44 (137), 287–303. ———. 1992. “Staggered and Synchronized Price Policies Under Inflation: The Multiproduct Monopoly Case.” Review of Economic Studies 59, 331–359. Simester, Duncan. 1995. “Signaling Price Image Using Advertised Prices.” Marketing Science 14 (2), 166. Slade, Margaret E. 1998. “Optimal Pricing with Costly Adjustment: Evidence from Retail-Grocery Prices.” Review of Economic Studies 65 (222), 87–88. ———. 1999. “Sticky Prices in a Dynamic Oligopoly: An Investigation of (s, S) Thresholds.” International Journal of Industrial Organization 17 (4), 477–511. Sobel, Joel. 1984. “The Timing of Sales.” Review of Economic Studies 51 (166), 353–368. ———. 1991. “Durable Goods Monopoly with Entry of New Consumers.” Econometrica 59, 1455–1485. Sodhi, Manmohan, and Navdeep Sodhi. 2008. Six Sigma Pricing. Upper Saddle River, NJ: FT Press. Srinivasan, Shuba, Koen Pauwels, and Vincent Nijs. 2008. “Demand-Based Pricing Versus Past-Price Dependence: A Cost-Benefit Analysis.” Journal of Marketing 72 (2), 15–27. ———. The Behavior of Industrial Prices, New York. Stigler, George J., and James K. Kindahl. 1970. “Industrial Prices, as Administered by Dr. Means.” American Economic Review 63 (4), 717–721. Stiving, Mark, and Russell S. Winer. 1997. “An Empirical Analysis of Price Endings with Scanner Data.” Journal of Consumer Research 24, 57–67. Stremersch, Stefan, and Gerard J. Tellis. 2002. “Strategic Bundling of Products and Prices: A New Synthesis for Marketing.” Journal of Marketing 66 (1), 55–72. Tagart, H. F., Minimum Prices Under the NRA, Michigan Business Studies, Vol. VII (3), viii, 307. Taylor, Malcolm D. 1936. “Progress in Marketing Research.” Journal of Marketing 1 (1) (July), 56–64. Tirole, Jean. 1988. The Theory of Industrial Organization. Cambridge, MA: MIT Press. Tucker, Rufus S. 1939. “Price and Price Policies by Walter Hamilton and Associates—A Review by R.S. Tucker.” Journal of Marketing 3 (1) (April), 326–330. Twedt, Dik W. 1965. “The Current Marketing Question.” Journal of Marketing 29 (4), 54–55. Tyagi, Rajeev K. 1999. “A Characterization of Retailer Response to Manufacturer Trade Deals.” Journal of Marketing Research 36 (4), 510–516. Varian, Hal R. 1980. “A Model of Sales.” American Economic Review 70, 651–659. Villas-Boas, J.M. 1995. “Models of Competitive Price Promotions: Some Empirical Evidence from the Coffee and Saltine Crackers Markets.” Journal of Economics and Management Strategy 4 (1), 85–107. Warner, Elizabeth J., and Robert B. Barsky. 1995. “The Timing and Magnitude of Retail Store Markdowns: Evidence from Weekends and Holidays.” Quarterly Journal of Economics 110 (2), 321–352. Wernerfelt, Birger. 1994. “Selling Formats for Search Goods.” Marketing Science 13 (3), 298–309. Williamson, Oliver E. 1979. “Transaction-Cost Economics: The Governance of Contractural Relations.” Journal of Law and Economics 22 (2), 233–261. Wilson, Lynn O., Allen M. Weiss, and George John. 1990. “Unbundling of Industrial Systems.” Journal of Marketing Research 27 (2), 123–138. Zbaracki, Mark J., Mark Ritson, Daniel Levy, Shantanu Dutta, and Mark Bergen. 2004. “Managerial and Customer Costs of Price Adjustment: Direct Evidence from Industrial Markets.” Review of Economics and Statistics 86 (2), 514–533. Zenor, Michael J. 1994. “The Profit Benefits of Category Management.” Journal of Marketing Research 31, 202–213.
ABOUT THE EDITOR AND CONTRIBUTORS Melissa Archpru Akaka is a doctoral student in marketing at the Shidler College of Business, University of Hawaii at Manoa. Hans Baumgartner is chair of the Department of Marketing, Smeal College of Business, Pennsylvania State University. Mark E. Bergen is the Carolyn I. Anderson Professor of Business Education Excellence at the Carlson School of Management, University of Minnesota. Pradeep K. Chintagunta is Robert Law Professor of Marketing at the Booth School of Business, University of Chicago. Shantanu Dutta is Vice Dean of Research Strategy and Advancement and the Dave and Jeanne Tappan Chair in Marketing at the Marshall School of Business, University of Southern California. Yi He is Assistant Professor at the Department of Marketing and Entrepreneurship, College of Business and Economics, California State University, East Bay. Dipak C. Jain is the Dean of the J. L. Kellogg School of Management at Northwestern University. Robert F. Lusch is Lisle and Roslyn Payne Professor in Marketing and the Department of Marketing Head at Eller College of Management, University of Arizona. Naresh K. Malhotra is Regents’ Professor in the Department of Marketing, College of Management, Georgia Institute of Technology. Matthew S. O’Hern is an Assistant Professor at Lundquist College of Business, University of Oregon. Koen Pauwels is Associate Professor of Business Administration at the Tuck School of Business, Dartmouth College. Sourav Ray is Associate Professor of Marketing at the DeGroote School of Business, McMaster University. Dave Reibstein is the William Stewart Woodside Professor and Professor of Marketing at The Wharton School, University of Pennsylvania. 189
190
ABOUT THE EDITOR AND CONTRIBUTORS
Aric Rindfleisch is Professor of Marketing and Associate Dean for Research and PhD Programs at the School of Business, University of Wisconsin–Madison. Siddharth S. Singh is Assistant Professor of Management at the Jesse H. Jones Graduate School of Management, Rice University. S. Sriram is Assistant Professor of Marketing at the Stephen M. Ross School of Business, University of Michigan. Stephen L. Vargo is Professor of Marketing and Shidler College Distinguished Associate Professor at the Shidler College of Business, University of Hawaii at Manoa.