Abstract:

When we analyse a series of sales data, the first problem posed is to find out the meaning of each figure which makes up the series; identification of the possible links which each figure has with the preceding one can be the starting point for constructing the basis of forecasts of the succeeding figures which continue the series. It would be ideal, when preparing forecasts, to be able to quantify exactly each variable which has to be considered when preparing the sales figures, so that in this case everything is reduced to a simple mathematical calculation, completely eliminating the risk of subjective estimates. This situation, however, is unrealistic and the best level which one can hope to reach is the limitation of these estimates to a well-defined margin. The scope of our brief comments is to initiate the study of sales forecasts as parts of a whole and to seek a method of solving each element. On the one hand, there exist sales figures which, projected into the future, give us the basis of theoretical forecasts; on the other, a mass of information (inserted into a single element of the forecasting pattern) which permits calibration of theoretical data in the "real world" of the market.