Automating the detection of turning points: Inventory control at Computershop

Inventory control for a product catalogue of 3000 products is carried out by two managers at ComputerShop. While there is a substantial level of automation of product flow in this company, there is no analysis of inventory levels, nor of trends in demand for each product. Inventory management thus is an area that imposes a high workload on the managers and is characterized by the usual problem of inventory and demand being poorly balanced. In this paper we have studied two techniques that can be applied to detect turning points in a sequence of sales data. We have demonstrated that both of these techniques can be used to support the partial automation of decisions on inventory control.