A sales forecasting model for food retailing

This paper presents a forecasting model for the daily sales of a food retailing chain developed on the basis of real data from hypermarkets. Sales variability is high as a result of marked seasonality and intensive marketing activity. The model is applicable to daily sales of individual items, only parameters vary with the type of article and shop. To forecast the daily sales the model takes into consideration the position of that day in the week, and year and the proximity of the day to an holiday. A classification of the ABC type provides the basis for an appropriate forecasting and replenishment method for each one of the items in the range. The flexible structure adopted will allow the incorporation of the price variation effect on sales for class A items and class B items, when promoted.