An interval cumulative sales forecasting method for consumer goods promotions

Abstract

Forecasting and monitoring promotional sales evolution is a managerial concern. In order to optimize opportunities and resources, promotional sales forecasting can be accomplished using cumulative methods. This paper proposes a interval sales forecasting method based on partially accumulated data, using a diffusion pattern among several promotions. The method was applied on a consumer goods industry, in order to forecast the results of nine promotions of three weeks length. The main conclusions remarks the ease of implementation of the method and accuracy considered acceptable by the manager. Results show that accuracy is related to the accumulation of data, and that although statistically the ex ante variation is not the same as the ex post variation, it is possible to use interval forecasts with reasonable success by the managerial point of view.