For companies proposing a large offer of products and services, it is important to identify the products for which the forecast quality is critical. The objective of this paper is to propose a strategic analysis of the products in order to determine priorities in the demand forecasting process: (1) a high priority class including the products for which forecast accuracy has a strong impact on the logistic performance of the company and where judgement should be integrated in the forecasting process; (2) a low priority class with the remaining products for which forecast accuracy is less important and where forecasts could be established in an automated way. The classification process takes into account different criteria chosen on the basis of cost efficiency reasoning.