Consumers’ annual spending on the Internet tripled from 1997 to 1998, according to an ongoing study of Internet users by the Wharton Forum on Electronic Commerce.

To help companies selling on-line to design their web site and plan their Internet-based marketing and distribution strategies, the Forum has developed the Wharton Virtual Test Market (WVTM), a consumer panel designed to represent the universe of World-Wide Web users. A panel is a research tool that tracks over time the attitudes and behavior of a group of participants, selected to be representative of a study population. TV ratings, for example, are a type of panel research. Panels can help project the behavior of consumers with high accuracy from a relatively small sample over selected periods of time.

The WVTM study recently reached a key step with the completion of a follow up survey, known as WVTM2. The findings of both WVTM1, undertaken in 1997, and WVTM2, completed in May, are summarized by Gerald L. Lohse, Steven Bellman and Eric Johnson in their paper, “Consumer Buying Behavior On The Internet: Findings From Panel Data.”

Lohse, Bellman and Johnson discovered that two major categories of variables predicted buying behavior: “time starvation” (people who had relatively little leisure time because they spend 50 or more hours a week at work); and a “wired” lifestyle (people who are heavy Internet users).

Time starvation exercised an important influence on buying on-line because, the researchers reasoned, people who work long hours are more likely to use the Net to shop, in order to save time. People who buy online also tend to be more “wired” than non-buyers. That is, they spend more time on the Internet, receive more e-mail, use the Internet at work more and have been on the Internet for years, as compared to months for non-buyers.

The researchers divided panelists into four segments:

1. “Never buys,” who did not buy on-line in either time periods (14% of the panelists).

2. “Dropouts,” who bought online in 1997 but not in 1998 (15%).

3. “Newbies,” who bought in 1998 but not in 1997 (31%).

4. “Steadfast buyers,” who bought in both years (40%).

There were striking differences among these groups. New and steadfast buyers had a higher household income ($55,447 and $52,968 a year, respectively) than never-buys and drop-outs ($42,800 and $44,618). They also worked more (47.7 and 49.7 hours per week versus 40.2 and 43.0), had been on the Internet for a longer time (32.3 and 34.2 months versus 26.4 and 27.1), and spent more time on the Internet (14.1 and 15.4 hours per week versus 11.9 and 12.3).

Interestingly, between 1997 and 1998, drop-outs decreased hours online from 14.7 hours each week to 12.3, received the most Spam (junk e-mail), increased their paper-based catalog orders by 20% and decreased use of the Internet for completing their work by 16%.

The more time an individual spent online, the greater the chance of buying. In 1997, 34% of the respondents who had been online for one month had made an online purchase. In 1998, 84% of those who had been online for five years had made a purchase. Also in 1998, buyers who had been using the Internet for 60 months had $300 in median annual purchases, compared to $180 for those using the Internet for 12 months.

On-line buying was also positively correlated to on-line product searching (the biggest predictor), the amount of e-mail received, use of printed catalogs and use of the Net for travel information, finance information, or news and current events. Males were slightly more likely to buy online than females, but they purchased only about 50 cents more annually.

The median on-line purchase increased from $49.53 in 1997 to $86.31 in 1998, while buyers increased their average number of purchases from 4.3 to 7.4. That’s a 74 percent increase in spending and a 72 percent jump in number of purchases, translating to an increase in annual expenditures from $213 to $639 within a single year. On-line spending increased by about 4 cents a year for every additional $10,000 in household income, the panel showed.

Based on an estimate of 55 million Internet users as of July 1998, WVTM data projects to $29.2 billion in on-line sales for the year 2000. Further assuming that the growth of new Internet users is 2.5% per month (as a number of sources have estimated), that 15% of online buyers would not purchase from the Web the next year and that new consumers’ buying matches that of existing online buyers, sales to new buyers will exceed old buyers by the year 2001.

WVTM1 and 2 show that a person’s tendency to buy and spend on the Internet can change rapidly over time. To find out just how fast these changes are accelerating, Wharton Forum researchers are already preparing WVTM3 for later this year. That should provide even more accurate data about the future size and shape of Internet shopping.