INFER: An expert system for automatic analysis of scanner data

Scanner technology has resulted in the problem of "too much" data. Marketing managers in packaged goods companies are inundated with 100 to 1,000 times more bits of data than what they had access to even a few years back. In this paper, we report our approach and experiences in developing a prototype expert system, called infer, that helps translate large quantities of marketing data into meaningful insights for managerial actions. The system automates data interpretation by employing a knowledge base of simple data interpretation techniques often used by analysts. The system generates a summary report useful for management decision making.