Special Session Summary Loyalty By Design: Understanding Consumers= Reluctance to Shop When Buying Online

Kyle B. Murray, University of Alberta
[ to cite ]:
Kyle B. Murray (2002) ,"Special Session Summary Loyalty By Design: Understanding Consumers= Reluctance to Shop When Buying Online", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 8-10.

Advances in Consumer Research Volume 29, 2002     Pages 8-10



Kyle B. Murray, University of Alberta

In general, economic models of search predict that increased search leads to lower prices and reduced consumer loyalty. The ease with which an on-line shopper can travel from one internet retailer to the next has raised the spectre of internet retailing as a medium of frictionless commerce or Bertrand competition. Given the potential for such hyper-competition and the rapid growth of on-line shopping, it is of significant interest to marketers to better understand the character and the extent of consumer loyalty in electronic marketplaces.

The purpose of this special session is to present and explain an alternative view: despite the potential for search, on-line shoppers tend to be extremely loyal and reluctant to search. The Johnson, Moe, Fader, Bellman and Lohse paper examines search across competing electronic commerce sites by analyzing panel data from over 10,000 households, and demonstrates that on-line shoppers are not inclined to search. The Murray and HSubl paper looks at the interaction between consumer experience and website design in order to investigate the notion that the development of interface specific user-skills can lock shoppers in to a particular site. The Zauberman paper develops a framework for understanding how the information environment and time preferences affect search and switching behavior over time.

The papers that were presented in this session demonstrate the phenomena of online lock-in and examine the mechanisms underlying the emerging notion that on-line consumers are susceptible to becoming locked-in to a particular retailer, even though switching between retailers has never been easier. In total, these papers make an important contribution to consumer research by enhancing our understanding of consumer loyalty, extending the literature on lock-in to electronic environments, and by substantiating a phenomenon that is counter-intuitive and contradicts a large body of existing search theory.



Eric Johnson, Columbia University

Wendy Moe, University of Pennsylvania

Peter Fader, University of Pennsylvania

Steven Bellman, University of Western Australia

Jerry Lohse, TomorrowLab, McKinsey & Company

We directly examine search across competing electronic commerce sites. Recent developments in data collection allow us to examine consumer search directly, by looking at the shopping patterns of a large panel of Internet users over time. We use data collected by Media Metrix, Inc., a firm that records every URL visited by families that are members of its panel using a small program, the PCMeter, which runs continuously as a background application on the family’s home computer. By analyzing panel data from over 10,000 internet households and three commodity-like products, books, CDs and travel, we show that the amount of search is actually quite limited. On average, households visit only 1.1 book sites, 1.2 CD sites, and 1.8 travel sites. Using probabilistic models, we characterize consumer search behavior at the individual level in terms of (1) depth of search, (2) dynamics of search, and (3) activity of search.

By modeling an individual’s tendency to search as a beta-geometric process, we find that consumers search across very few sites in a given shopping month. We extend the beta-geometric model of search to allow for any time-varying dynamics which may exist causing the consumer to evolve and, perhaps, learn to search over time. We find that for two of the three product categories studied, search propensity does not change from month-to-month. However, in the third product category, we find mild evidence of time-varying dynamics, where search decreases over time from already low levels. Finally, we model the level of a household’s shopping activity and include it in our model of search. Interestingly, the results suggest that more active online shoppers also tend to search across more sites. This consumer characteristic largely drives the dynamics of search often mistaken as time-varying effects.

Like Adamic and Huberman, (1999) our data suggest that people visit few stores online despite the fact that consumers are "just a mouse click away" from other stores. Browsing behavior varied by product category and level of activity but showed no increase with experience. In understanding the significance of our results, it is important to realize that we examined product classes and time periods which are very similar to those used by researchers who have examined changes in price and price dispersion on the web. The attraction of books, compact disks, and travel for examining these questions is their apparent commodity-like status which should provoke higher levels of search. And our data have been collected during time frames that substantially overlap time periods used in prior studies of prices for CD’s and books (Brynjolfsson and Smith 2000b) and airline travel (Clemons et al. 1999), minimizing the potential for historical differences and suggesting that the lack of search is one reason we see price dispersion.


Adamic, L.A., and Huberman, B.A. (1999). "The Nature of Markets on the World Wide Web", Xerox Palo Alto Ressearch Center, Palo Alto, CA.

Brynjolfsson, E., and Smith, M. (2000a). "The great equalizer? Customer Choice Behavior at Internet Shopbots", Sloan School, MIT, Cambridge, MA.

Clemons, E., Hann, I.H., and Hitt, L. (1999). "The Nature of Competition in Electronic Markets: An Empirical Investigation of Online Travel Agencies Offerings", The Wharton School, University of Pennsylvania.



Kyle B. Murray, University of Alberta

Gerald HSubl, University of Alberta

Marketing researchers (see for example, de Figueiredo 2000; and Johnson, Lohse, and Mandel 2000) and the popular press (Lewis 1997) have argued that on-line shopping represents a dramatic shift in consumers’ behavior in relation to the competitive marketplaces within which they interact. Specifically, it has been argued that the dramatic reduction in search costs and the ease with which comparison-shopping can be accomplishedBoften referred to as a "frictionless" economyBwill lead to a substantial reduction in consumer loyalty (de Figueiredo 2000; and Lewis 1997).

However, contrary to popular opinion, on-line shopping has not lived up to its billing as a frictionless marketplace. Although the potential to reduce costs related to information search and product comparison clearly exists in electronic markets (HSubl and Trifts 2000), recent research suggests that consumers are engaging in only limited search and comparison shopping. In fact, Johnson, Moe, Fader, Bellman and Lohse (2000) find that the average number of on-line stores searched before a CD is purchased is 1.1, and that 70% of CD shoppers and 70% of book shoppers are loyal to a single site. Similarly, Brynjolfsson and Smith (1999) find that on-line consumers tend to make purchases from higher priced vendors even for homogeneous products such as CDs and books. The early evidence from on-line shopping seems to suggest that cyberspace is far from frictionless. An emerging stream of research describes this sort of on-line loyalty as cognitive lock-in (Johnson, Bellman, and Lohse 2000; Shapiro and Varian 1999; Zauberman 2001).

Cognitive lock-in refers to the idea that, even in situations where search costs are low, and searching results in consumers paying a lower price, consumers will not shop around. Instead, they will return to the site with which they are experienced. This type of loyalty has been explained as the result of the development of user skills (Stigler and Becker 1977; Wernerfelt 1985), which make incumbent sites more useful, even though, given the ame experience with another site, it would be equally useful (Johnson, Bellman, and Lohse 2000). The user skills explanation predicts that both the amount of exposure to the incumbent and the similarity of the two sites will have a significant impact on the subjects’ site preference. In addition, this explanation predicts that with greater exposure to the incumbent interface, the ease of switching and perceived similarity should decrease. The focus of this paper is on the sufficiency of the user skills theory of brand loyalty for explaining cognitive lock-in within electronic marketplaces.

In an experimental setting we examine consumers’ switching behavior under conditions in which we vary the amount of exposure to an incumbent interface versus the amount of exposure to a competitor interface, as well as the similarity between the two interfaces. We conceptualise non-functional similarity as the differences in appearance that effect the relative learning costs between two interfaces that perform equally well. We manipulated learning costs, by manipulating interface similarity, and measured its impact on the choice of interface. The core design of the experiment is a 2 (interface similarity) by 6 (number of exposures to the incumbent interface), between subjects design. Subjects are exposed to an incumbent interface, with which they must complete an on-line shopping task, from one to six times. Subjects who were assigned to less than 6 shopping trips were required to complete "filler" tasks, in which they were exposed to the same stimuli, but in a task that was unrelated to the shopping task and using an interface that was unrelated to the shopping task interface. They were then forced to complete a similar shopping task using a competitor interface, which appeared to be either very similar to the incumbent interface or quite different from the incumbent interface. Which interface served as the incumbent, and which served as the competitor, was counter-balanced between conditions. After exposure to the competitor interface, subjects were forced to choose which interface (the competitor or the incumbent) they wished to use for another set of shopping tasks. They were also asked to indicate the extent of their preference, on a ten-point scale, for the interface that they have chosen. At the end of the experiment a number of measures were taken related to how easy it was to switch from using the incumbent to the competitor and how similar the two interfaces appeared to be.

Analysis of the data provides some interesting results. We find that as exposure to the incumbent increases, transferable user skills increase for both the incumbent and the competitor. However, non-transferable skills, specific to the incumbent interface are also acquired, which results in a preference for the incumbent. Moreover, when the incumbent interface is different from the competitor the preference for the incumbent is intensified. In this paper we find strong evidence in favor of the notion that consumers can become locked-in to a particular interface as a result of experience with that interface, even when an equally useful competitor interface is available.


Brynjolfsson and Smith (1999), "Frictionless Commerce? A Comparison of Internet and Conventional Retailers", Management Science, 46(4), April 2000.

de Figueiredo, John M. Finding Sustainable Profitability in Electronic Commerce (2000). Sloan Management Review (Summer), 41-52.

HSubl, Gerald and Valerie Trifts (2000), "Consumer Decision Making in Online Shopping Environments: The Effects of Interactive Decision Aids", Marketing Science, 19(1), 4-21.

Johnson, Eric, Wendy Moe, Peter Fader, Steven Bellman and Gerald L. Lohse (2000), "On the Depth and Dynamics of Online Search Behavior", Working paper, The Wharton School. University of Pennsylvannia.

Johnson, Bellman, and Lohse (2000), "What Makes a Web Site 'Sticky’? Cognitive Lock-In and the Power Law of Practice", Working paper, The Wharton School. University of Pennsylvannia.

Lewis, T. G. (1997). The Friction-Free Economy: Strategies for Success in a Wired World. New York, NY: Harpercollins.

Shapiro, Carl and Hal R. Varian (1999), Information Rules, Boston, MA: Harvard Business School Press.

Stigler, George J. and Gary S. Becker (1977), "De Gustibus Non Est Disputandum", American Economic Review, 67, 76-90.

Wernerfelt, Birger (1985), "Brand Loyalty and User Skills", Journal of Economic Behavior and Organization, 6, 381-385.

Zauberman, Gal (2001), "Consumer Lock-In Over Time: The Impact of Time Preferences and Failure to Predict Switching Costs", Dissertation, Duke University.



Gal Zauberman, University of North Carolina, Chapel Hill

The emergence of the Internet has generated increased interest in the effect of information acquisition and processing costs on search and choice behavior. Economic theory as well as common belief generally assumes that lower search and evaluation costs will allow consumers greater mobility and the ability to find the best fitting option. Market behavior and initial research, however, do not support these claims, and suggest even greater consumers stickiness (i.e., reduced switching) on the Internet, compared to the brick and mortar retail environment. The effects of the Internet on consumer search and switching provide a timely and important instance of the general effects of the information cost structure on behavior over time.

Conceptually, the focus of this work is on the impact of the information environment and time preferences on consumer search patterns and choice behavior over time. A key element proposed in this work is that the information cost structure is conceptualized in terms of two temporally distinct information cost categories that are traded off: initial set-up costs and on-going usage costs. The overarching hypothesis is that he relative difference between information costs over time, coupled with consumers’ time preferences, can explain consumers’ reluctance to search extensively for better alternatives and ultimately switch. I argue that consumers are centered on short-term considerations and do not fully incorporate the impact of current investment on future behavior. Lock-in is induced by a preference for minimizing present costs and underpredicting switching costs.

The current theory presents two aspects in the lock-in process. The first aspect relates to the effect of the cost structure and time preferences on consumers’ initial selection, and the second aspect relates to the factors that affect search and choice over time once consumers have invested. The main points are that consumers do not fully incorporate the future into their decisions; instead, they tend to mispredict the impact of future switching costs. Once the future arrives, these switching costs exert effects more powerful than the consumer had originally predicted.

The findings of the four experiments provide support for the proposed account of the lock-in phenomenon. The results from experiment 1 and 2 demonstrate that consumers mispredict their propensity to search and switch at a later point. These results indicate that consumers do not incorporate the full impact of their decisions on later behavior. Experiments 3 and 4 demonstrate that the relative cost structure, whether manipulated or due to prior investment leads to lock-in. Due to the relative cost structure and the focus on the current decision, participants with prior investment base their subsequent decisions on their past selection and engage in little additional search.

The results of experiment 3 and 4 also point to several consequences of lock-in for consumer welfare. Both experiments indicate that lock-in reduced the quality of the selected option, but that, participants who were locked in expressed greater satisfaction with their chosen retailer and the entire shopping experience. The results from experiments 3 also showed that consumers that are locked in are more satisfied with their chosen (inferior) option. This suggests that the search experience is incorporated as part of the set-up costs that are psychologically bundled with the other aspects of the shopping experience.

Taken together, these four experiments support the theory and its conclusion that consumers’ time preferences and the relative costs they face are critical factors in determining initial selection of retailers as well as subsequent switching decisions. The results indicate that consumers select options that minimize upfront costs and mispredict the consequences of this decision. Consumer lock-in is a combined effect of reduced search, and thus reduced exposure to other alternatives, as well as a diminished tendency to switch when other options are considered.

In conclusion, the initial question was motivated by the low switching rates observed on the Internet; the theory I present, however, is more general. The theoretical framework developed in this article combines concepts from intertemporal choice, switching, and search behavior to provide a way to understand lock-in. In this article I argued that this framework is a useful perspective on consumers lock-in in the reduced search cost environment of e-commerce.