For many years IBM has been conducting a long-range investigation into the future of retail data processing. It was and still is our aim to design methods and techniques which will allow an automatic replenishment of the items carried in department stores, chain stores and food chains. The system is known under the name Retail-IMPACT. It has been clear from the beginning of our studies that not all types of merchandise would respond to the same system. We believe that for replenishment and inventory management purposes the merchandise can be divided into four major classes: Big ticket, fashion, staple and grocery. However, it is not possible to draw a straight line between these classes. For inventory management, big ticket items or dry grocery may well be treated like staple items. On the other hand, the decision if an item is fashion or staple may be very hard to make in today's volatile market environment.