Abstract: A system and method of selecting products and services for an audience likely to purchase such products and services utilizes a collection of data reflecting, for individual consumers, a history of purchases, and retention or return of such purchases. For individual consumers, subsets of consumers are created whose tastes correlate positively and negatively with the individual consumer. Lists are created of products for a target consumer based on the likes of the subset positively correlating and the dislikes of the subset negatively correlating. Recommended products are offered with a financial incentive such as a money back guarantee which serves the dual purpose of inducing the purchase as well as providing the purchase and return data needed in the data collection.