The application of prediction capability and adaptive target control to the newspaper business

A newspaper wholesaler determines the optimal supply of newspapers, which minimises costs or maximises profit when demand is uncertain. Targets concerning the mean number of unsold newspapers and the probability of not stocking out are commonly used in the absence of known costs. These targets help to strike a balance between excessive wastage (overage) and increased stock-outs (underage). Wholesalers also use various methods of predicting demand, some of which achieve one of the targets while others achieve both. We investigate the relationship between these targets and the prediction method. We say that a prediction method is ‘capable’ if both targets can be achieved when using that method. We introduce a technique for adapting targets, which renders the prediction method capable. We show how optimality is achieved while preserving prediction capability. We use a case study drawn from the newspaper business to illustrate the automatic allocation of newspapers.