Model developed for an industrial company, it compared the forecast of sales specialists and the statistical models of time series applying the Smoothing, ARIMA and ARIMA-X techniques.
The great expectation of the best model was in the ARIMA-X technique, which aims to correlate and adjust the forecast with economic indicators in the sales forecast.
In this sales forecast, the indicators of Inflation, GDP, Industrial Production Index, Dollar and Price Adjustments were evaluated.
In generating the models, 2 of these indicators showed the best model accuracy results (before evaluating the actual results). After 9 months of monitoring and comparison with the real data, the experts’ model generated an error of 18% in relation to the one performed, and the ARIMA-X model with one of the selected indicators reduced this error to 0.5%.
The forecast model also worked with other indicators, such as Production Volume and Price. By crossing Revenue, Production Volume and Price, it was possible to evaluate the trend of each product line to reassess commercial strategies.
Sales Forecast with Economic Indicators correlation
