AI-supported sales forecast

Quick Check

Initial situation

NDW Draht und Stahl GmbH is a company in the steel industry that also offers services for affiliated companies. NDW is part of a family-run group of companies which, in addition to wire production, is also active in various related technology areas. In order to optimize processes and make data-based decisions, future sales are to be forecast. The aim of the quick check was to investigate the feasibility of forecasting future sales with the available data using machine learning methods.

Solution idea

In order to achieve a reliable sales forecast, future sales can be predicted by analyzing past sales. The forecast quality is improved by taking external influencing factors into account. The identification of various characteristic factors, such as seasonal influences, enables a better understanding of the data and a higher forecast quality.

Result

The data provided was analyzed as part of the Quick Check. The time series of sales was broken down into a seasonal and a trend component, on the basis of which the future course of sales can be forecast. The influence of external factors on sales was also examined. It was found that a direct correlation between the external data and the sales data can hardly be identified. As a result, a differentiated consideration of external variables, for example using methods such as SARIMAX, would appear to make sense.