TOWARDS A DYNAMIC VIEW OF ONLINE SHOPPING
By Ed Colet
The rapid growth of electronic commerce on the Internet has led to numerous research studies and surveys geared towards understanding this phenomenon. How are people shopping on the web? What patterns characterize their shopping behavior? What can retailers and technologists learn from such studies? An example of some interesting results from a recent study is discussed here - but taken at face value such results may be of limited value. But by taking into account data mining considerations and trends in personalization and one-to-one marketing, then a better understanding of Ecommerce and online shopping can be gained.
The Consumer Electronics Manufacturers Association (CEMA) conducted an email survey of 2725 online shoppers and concluded that online shoppers generally fall in to four different types of categories. CEMA is a part of the Electronic Industries Alliance; a trade organization representing various facets of electronics manufacturing such as audio, video, accessories, mobile electronics, communication, information technologies, and multimedia products that sold through consumer channels. The full survey can be bought at http://www.eBrain.org. The highlights presented here are based on a report available via news services.
The CEMA study identified 4 categories of online shoppers:
(1) Convenience lovers: 65% of respondents were classified into this category. The profile of such as shopper is a baby boomer, with a high household income, likely to make impulse purchases. This category has a high percentage of females. Convenience is a driving factor for purchases.
(2) Money Savers: Generation X'ers, mostly male, with high household incomes, likely to shop at retail stores and then buy online. They represent the largest potential for growth of online shopping.
(3) Smart Shoppers: Baby boomers with middle household incomes, willing to research and likely to buy at retail or online based on research results.
(4) Selection shoppers: Have a lower income household, often go directly to a web site to make purchases and may have limited traditional retail selections where they live.
Based on the above shopping type, the study identifies several web sites that fit those shopper characteristics. Some examples are http://www.eBay.com for "Money Savers", http://www.ConsumerReports.com for "Smart Shoppers", http://www.amazon.com for "Selection Shoppers", and http://www.911gifts.com for "Convenience Lovers".
But their assumption of a "fixed model", i.e., the assumption that both shopper characteristics and web sites are stable over time may limit this study (and other similar ones). But in this age of dynamic change the assumption of fixed shopper characteristics and stable web sites may be short sighted and of limited utility. I propose that what is more likely to exist is a dynamic model, i.e., where shopper characteristics and/or web sites change.
To elaborate, there are the 4 possible scenarios, and their likely evolution:
(1) Stable shoppers and stable web sites - Assumed by the fixed model described above.
(2) Shoppers' change but web sites are stable. This is probably true today to some extent. For example, based on my own purchasing habits and interests, in certain situations I'm interested in lowest cost, in others will undertake extensive research, in some cases an impulse purchase is made, and at other times I go directly to a web site. Thus based on my habits I seem to touch on all four categories identified by the study. It does appear that the web sites I visit remain relatively stable (other than their content being updated).
(3) Web sites change but shoppers don't. This would mean that web sites change dynamically, not just their content, but their navigation paths, their look and feel, etc. A changing web site will have content uniquely geared to the individual shopper whose characteristics are thought to be stable. It's the current view of personalized and one-one marketing - an individually customized web site for a stable user. Identifying the shopper so that the site can "adapt" itself to him/her has been the difficulty in achieving this. Data mining of user-histories and web logs are current approaches that are utilized. It may be limited unless the fact that people change can be taken into account.
(4) Shoppers and web sites both change. This is the trend and direction things may evolve towards. In this scenario, both shopper characteristics and web sites are dynamic. This can be done via continued efforts and research at personalization and customization - much of which involves data mining. Data mining conducted at the level of an individual user - e.g. patterns about how I shop for certain products, and then being directed to appropriate web sites that are consistent with such shopping modes (money savers, research oriented, etc.). The information on the web sites is then dynamically customized and presented in a manner that is consistent with the type of shopper characteristics or modal pattern that best fits.
Is this last scenario a long way off? Because it's already suggested by efforts underway at developing personalized technology, and because data mining technology enables finer and finer grained analysis, it may be closer than we think.
---
Ed Colet is the Acting Director of Research at Virtual Gold Inc., responsible for developing analytical methods for data mining and for investigating human factors and usability issues of business intelligence systems. At present, he is in the final stage of completing a doctoral dissertation in the Cognition and Perception program at New York University's Department of Psychology. Ed has also worked for IBM Research at the T.J. Watson Research Center. At IBM, Ed was a member of the group that developed Advanced Scout, the data mining application for NBA teams. His research interests focus on statistical methods and human factors.
For more information, see http://www.virtualgold.com.