رفتار خرید مصرف کننده در اینترنت: یافته های حاصل از داده های پانل

Online retailing became big business in late 1998, as millions of people placed orders for holiday gifts online and retailers scrambled to upgrade their distribution networks to cope with the growth (Cyberatlas.com, 1999). Companies planning for the growth of online retailing need reliable estimates of the growth of online shopping. Data about online consumer purchasing behavior are also needed to help companies define their online retail strategies for Web site design, online advertising, market segmentation, product variety, inventory holding, and distribution. Forecasts are more likely to be reliable if they are based on the behavior of online consumers, rather than consumers’ stated intentions, or worse, the guesses offered by Web marketing “experts.”

Online retailing became big business in late 1998, as millions of people placed orders for holiday gifts online and retailers scrambled to upgrade their distribution networks to cope with the growth (Cyberatlas.com, 1999). Companies planning for the growth of online retailing need reliable estimates of the growth of online shopping. Data about online consumer purchasing behavior are also needed to help companies define their online retail strategies for Web site design, online advertising, market segmentation, product variety, inventory holding, and distribution. Forecasts are more likely to be reliable if they are based on the behavior of online consumers, rather than consumers’ stated intentions, or worse, the guesses offered by Web marketing “experts.”survey of WVTM panelists (hereinafter referred to as the WVTM2)—in particular the changes over time within individuals that only a panel such as the WVTM can reveal. We then close by modeling changes over time within individuals to forecast total online spending for the year 2000 using a Monte Carlo simulation.

The strength of panel data is its ability to provide information about changes in behavior over time for individual consumers. Using the WVTM, we found that over a 12-month period online consumers doubled the number of items bought online and spent nearly three quarters more on each purchase, two facts that together resulted in a tripling of the spending of the average online consumer over time. The size of the online retail market will be driven not only by an increase in the number of people who go online for the first time in the next few years, but also by increases (and, interestingly, decreases) in online shopping by individual consumers already online. The results of the WVTM2 show that the Internet population is already starting to look more and more like the general population, at least in the United States. Companies will have to plan their Web site design for an audience that is less Web savvy, less educated, earns less, and is possibly less tolerant of new technology. According to our projections, most of the money earned by online retailers in 2 years’ time will come from people who have yet to connect to the Internet. However, this research also shows that differences between new online consumers and more experienced online consumers are erased over time by the rapid acclimatization of consumers to this new medium of consumption. The opportunity to observe differences over time within subjects has also revealed some of the reasons why people change their online shopping behavior. Those people who bought online last year but “dropped out” of online shopping this year seem to have had some badexperiences with online retailers. These people also claim to be getting more spam than other people online. Online retailers will have to provide significant incentives to win these people back to buying online again (news of stockouts and bogus companies will not help). In the meantime, these Dropouts are returning to paper- based catalogs for convenient shopping from home. The people who have never bought are increasing the amount of time they are spending online, and may in time make at least one online purchase. The WVTM2 offers the chance to observe changes in time over two periods. With two observation points, it is possible to extrapolate a straight line with which to forecast the future. But the growth in Internet shopping seems so dynamic that it is hardly likely to be a simple linear phenomenon. To gauge whether changes over time within individuals are linear, or curvilinear (for example, a tripling each year of the annual spending of each individual), we will need to observe at least three points in time. Even more periods may allow estimates of the limits of this growth, which cannot be infinite, whereas linear and simple curvilinear projections have no asymptotes. With this in mind, the Wharton Forum on Electronic Commerce is collecting a third year of data from the WVTM.