A small business builds relationships with its customers by noticing their needs, remembering their preferences, and learning from past interactions how to serve them better in the future. How can a large enterprise accomplish some thing similar when most company employees may never interact personally with customers? Even where there is customer interaction, it is likely to be with a different sales clerk or anonymous call-center employee each time, so how can the enterprise notice, remember, and learn from these interactions? What can replace the creative intuition of the sole proprietor who recognizes cus tomers by name, face, and voice, and remembers their habits and preferences?
In a word, nothing. But that does not mean that we cannot try. Through the clever application of information technology, even the largest enterprise can come surprisingly close. In large commercial enterprises, the first step—noticing what the customer does—has already largely been automated. Transaction pro cessing systems are everywhere, collecting data on seemingly everything. The records generated by automatic teller machines, telephone switches, Web servers, point-of-sale scanners, and the like are the raw material for data mining. These days, we all go through life generating a constant stream of transac tion records. When you pick up the phone to order a canoe paddle from L.L.