A DISTRIBUTED SOFTWARE FOR COLLABORATIVE SALES FORECASTING Kary FRÄMLING * , Jan HOLMSTRÖM ** * ARCADA Polytechnic, Skogsmansgränden 3, 02130 Espoo, FINLAND kary.framling@arcada.fi , http://www.sit.fi/~kary ** TAI institute, Helsinki University of Technology, Otakaari 1, 02150 Espoo, FINLAND jan.holmstrom@hut.fi Abstract: Main advantages of distributed software are that computers connected over a network may access information stored anywhere in the network at any time and get automatically notified about information changes. This paper describes how a distributed software is built for collaborative demand forecasting. Demand forecasts of both providers and clients can be accessed and used for getting a better idea about the marketplace and thereby do more accurate demand forecasting. Keywords: computer-integrated enterprises, co-operation, decision support systems, information systems, data privacy. 1. INTRODUCTION In a competitive and changing business environment, reliable forecasts are needed both for planning long- term business strategies and ensuring that the supply chain operates effectively on a day-to-day basis. In the consumer goods business, delivery times to the trade are short and the product ranges are wide. Product ranges are wide because a widely used strategy to increase demand is to increase product variety (Quelch and Kenny, 1994). Few producers are able to allocate capacity to produce according to daily consumer demand. A high delivery service is maintained by supplying from stock. To improve the supply chain efficiency producers strive to forecast sales. If the attempt is successful it is possible to reduce safety stock levels while maintaining the service level. Suppliers, wholesalers and retailers are today increasingly working together to improve supply chain performance and consumer value (ECR Best Practices Operating Committee, 1995). There are obvious advantages from the co-operation in forecasting and supply chain control. A major advantage is that feedback on consumer demand changes can be quickly communicated upstream to suppliers. Similarly, the trade can benefit from being involved in new product launches and promotion planning. This paper presents a distributed program whose purpose is to facilitate and accelerate co-operation between the members of the supply chain. Section 2 describes how the members of the supply chain connect together, how the distributed program is set up and managed, as well as the technical solutions behind it. Section 3 describes the currently used method for doing assortment forecasting (Holmström, 1998). It also explains the kind of data that is moved around in the distributed program. Section 4 concerns data security both from the network point of view and from the data accessibility point of view. 2. DISTRIBUTED SUPPLY CHAIN MANAGEMENT PROGRAM A supply chain usually consists of suppliers, wholesalers and retailers. Some of these are just clients to others, others are just providers, while some are both clients and providers to other members of the supply chain (Fig. 1). From a software point of view, the client and provider roles are more useful than the supplier, wholesaler and retailer roles. Treating them all as potential clients and/or providers makes it possible to model each member of the supply chain in a uniform way.