Choosing electronic supermarket customers.

Electronic commerce setup costs can be expensive to the service provider. For this reason it is desirable that those who undergo the electronic commerce registration process do actually use the service. This current paper explores the notion of the use of a 'weighted application blank' (WAB) to identify individuals who will, or will not, use the electronic commerce service. Participants consisted of 174 individuals registered with an electronic supermarket. Participants completed a short attitudinal questionnaire assessing their attitudes to the electronic supermarket system. Use of the system was then monitored for six consecutive months. From the sample, 73 individuals never used the service in the six months after completion of the questionnaire and 40 used the service for at least five months. A linear discriminant analysis was then conducted to ascertain if a measure could be developed to discriminate between these two groups. The analysis identified that the two groups could be discriminated between, such that a 76.99% correct classification could be made based upon two questions. Next, the function derived from the non-shopper and consistent shopper groups was applied to the full 174 individuals, where the probability level of becoming a consistent shopper was calculated for each individual. False acceptance and false rejection rates were then calculated for each group after manipulating the cut-off source to illustrate how such a system could be used in the operational environment. Collectively, the results suggest that the WAB may be a useful tool in the selection of electronic commerce customers when the service provider incurs some setup costs. [ABSTRACT FROM AUTHOR]