The Perceived Effectiveness of Virtual Shopping Agents For Search Vs. Experience Goods

EXTENDED ABSTRACT - Internet commerce has boomed in recent years. The U.S. Department of Commerce recently estimated 2001 e-commerce sales at $32.6 billion. In addition to the convenience of shopping online, consumers also enjoy the depth of product assortment that many virtual stores provide. Not only are the product choices numerous, the amount of information available on the Internet about these products is astounding as well. In fact, in most cases there is so much information available that consumers may feel overwhelmed by its abundance and experience an "information overload." To help handle such a large volume of information, several online retailers and independent web sites now offer customers online "shopping agents." These agents help customers make a variety of choice decisions through "dynamic customization." Research shows that the use of shopping agents improves the quality as well as the efficiency of purchase decisions.

Citation:

Praveen Aggarwal and Rajiv Vaidyanathan (2003) ,"The Perceived Effectiveness of Virtual Shopping Agents For Search Vs. Experience Goods", in NA - Advances in Consumer Research Volume 30, eds. Punam Anand Keller and Dennis W. Rook, Valdosta, GA : Association for Consumer Research, Pages: 347-348.

Advances in Consumer Research Volume 30, 2003     Pages 347-348

THE PERCEIVED EFFECTIVENESS OF VIRTUAL SHOPPING AGENTS FOR SEARCH VS. EXPERIENCE GOODS

Praveen Aggarwal, University of Minnesota, Duluth

Rajiv Vaidyanathan, University of Minnesota, Duluth

[The authors would like to thank Jeff Skubic for his help at various stages of this project.]

EXTENDED ABSTRACT -

Internet commerce has boomed in recent years. The U.S. Department of Commerce recently estimated 2001 e-commerce sales at $32.6 billion. In addition to the convenience of shopping online, consumers also enjoy the depth of product assortment that many virtual stores provide. Not only are the product choices numerous, the amount of information available on the Internet about these products is astounding as well. In fact, in most cases there is so much information available that consumers may feel overwhelmed by its abundance and experience an "information overload." To help handle such a large volume of information, several online retailers and independent web sites now offer customers online "shopping agents." These agents help customers make a variety of choice decisions through "dynamic customization." Research shows that the use of shopping agents improves the quality as well as the efficiency of purchase decisions.

In this paper, we examine the perceived effectiveness of shopping agents in making purchase recommendations. Two types of recommendation agents are examined in this study: rule-based filtering (RF) agents and collaborative filtering (CF) agents. RF agents base their recommendations on customer-stated preferences for product attributes whereas CF agents recommend on the basis of what other buyers with similar tastes and preferences chose. RF agents typically ask buyers their self-explicated preferences for product features and the attribute level trade-offs they are willing to make. Using a conjoint-analysis type modeling, these agents recommend products that maximize utility for a given set of attribute ratings. On the other hand, CF agents match users with buyers who have similar profiles and preferences, and make recommendations based on shared likes and dislikes.

This study also examines perceived agent effectiveness for different types of products. The two types included in this study are: search goods and experience goods. For search goods, it is easy to verify and inspect product attributes before making a purchase. For experience goods, it is infeasible to verify or inspect the attributes without purchasing and consuming the product. We hypothesize that recommendation agents will be more effective for search goods than experience goods because it is easier to define, observe, and evaluate product characteristics for search goods. We also hypothesize that for search goods recommendations, an RF-based process will be perceived as more effective, whereas for experience goods, a CF-based process will work better. We measure perceived effectiveness on three dimensions: perceived quality of recommendation, satisfaction with the recommendation, and intent to follow-up on a recommendation.

A mixed model, computer-aided experimental design was used to test model hypotheses, where recommendation development process (RF or CF) was a between subject factor and product type (search or experience) was a within subject factor. Subjects were 109 undergraduate business students at a Midwestern university. Subjects in the RF group were told that the recommendation would be based on matching their stated attribute preferences with a vast database of products. Subjects in the CF group were told that their recommendation would be made by comparing their preferences with those of "thousands of others" in its database. After collecting information related to product attributes and personal preferences, the software made a recommendation. Once subjects got the recommendation, they were asked a number of questions regarding their perceived effectiveness of the agent.

Data were first tested to ensure appropriate scale properties. The first hypothesis, that agents are likely to be more effective for search goods (than for experience goods) was strongly supported. There were statistically significant differences on all three dimensions: perceived quality of recommendation, satisfaction with the recommendation, and intent to follow-up on a recommendation. Second, as hypothesized for search goods, the RF process was perceived as more effective than the CF process. Third, although the CF process did not turn out to be more effective than RF process for experience goods, the superiority of the RF process that was witnessed in the case of search goods, disappeared in case of experience goods.

The findings of this study have important managerial implications. First, it appears that recommendation agents will be received much better in the context of search goods with clear and concrete criteria for their evaluation. As these criteria become more experiential in nature, the effectiveness of recommendation agents is likely to decline. Second, regardless of the "actual" recommendation quality, subjects’ assumption of the process by which the agent arrived at its recommendation makes a difference in their evaluation of the recommendation. For the search product in our study (camera), given identical final recommendations, subjects evaluated the recommendation based on individual attribute based (RF) process more favorably than the same recommendation based on the evaluations of similar others (CF process). Thus, online retailers should prefer an RF-based process if the agent is intended for making primarily search goods recommendations. We had hypothesized that subjects would prefer a CF-based process for experience goods recommendations. However, we did not see such a preference in our study. It is conjectured that "source credibility" may be an issue that moderates the perceived effectiveness of CF-based recommendations. Future studies should control for credibility to fully understand how the effectiveness of CF-based agents is determined.

REFERENCES

Aggarwal, Praveen (1998), "Deciding Not to Decide: Antecedents of Consumer Decision Delegation," Unpublished Doctoral Dissertation, Syracuse University, Syracuse, NY.

Alba, Joseph W., John Lynch, Bart Weitz, Chris Janisewski, Rich Lutz, Alan Sawyer, and Stacy Wood (1997), "Interactive Home Shopping: Consumer, Retailer, and Manufacturer Incentives to Participate in Electronic Marketplaces," Journal of Marketing, 61 (July), 38-53.

Ansari, Asim, Skander Essegaier, and Rajeev Kohli (2000), "Internet Recommendation Systems," Journal of Marketing Research, 37 (August), pp. 363-375.

Ariely, Dan, John G. Lynch, Jr. and Manuel Aparicio IV (2002), "Which Intelligent Agents Are Smarter? An Analysis of Relative Performance of Collaborative and Individual Based Recommendation Agents," working paper, Sloan School of Management, MIT.

CDNOW (2002), CDNOW "AlbumAdvisor" Website at http://www.cdnow.com/albumadvisor/.

Churchill, Gilbert A. (1979), "A Paradigm for Developing Better Marketing Constructs," Journal of Marketing Research, 16 (February), pp. 64-73.

CyberAtlas Staff (2002), "B2B E-Commerce Headed for Trillions," online report of March 6, 2002 at http://cyberatlas.internet.com/markets/b2b/print/0,,10091_986661,00.html.

Emons, Winand (1997), "Credence Goods and Fraudulent Experts," RAND Journal of Economics, 28 (Spring), 107-119.

Freiden, Jon B. and Ronald E. Goldsmith (1989), "Prepurchase Information Seeking for Professional Services," Journal of Services Marketing, 3 (4), 45-55.

Gilovich, Thomas and Victoria H. Medvec (1995), "The Experience of Regret: What, When, and Why," Psychological Review, 102 (April), pp. 379-395.

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

Malhotra, Naresh K. (1982), "Information Load and Consumer Decision-making," Journal of Consumer Research, 8 (March), pp. 419-430.

Moorman, Christine, Rohit Deshpande, and Gerald Zaltman (1993), "Factors Affecting Trust in Market Research Relationships," Journal of Marketing, 57 (January), 81-101.

Nelson, Phillip (1970), "Information and Consumer Behavior," Journal of Political Economy, 78 (March/April), 311-329.

Pastore, Michael (2002), "U.S. E-Commerce Spikes in Q4 2001," online report of February 20, 2002 at http://cyberatlas.internet.com/markets/retailing/print/0,,6061_977751,00.html.

Peterson, Robert A., S. Balasubramanian, and Bart J. Bronnenberg (1997), "Exploring the Implications of the Internet for Consumer Marketing," Journal of the Academy of Marketing Science, 25 (4), 329-346.

Taylor, Chris (1999), "Bot Till You Drop," Time, October 11, 1999,154 (15).

Venkatraman, N. and John C. Henderson (1998), "Real Strategies for Virtual Organizing," Sloan Management Review, 40 (Fall), pp. 33-48.

Vigoroso, Mark W. (2002), "Whatever Happened to E-Commerce Robots?" E-Commerce Times, February 28, 2002. Available online at http://www.ecommercetimes.com/perl/printer/16548/.

Wilder, Clinton (1998), "Agents Go Price Shopping," Internetweek, December 7, pp. 23-25.

----------------------------------------