EBIT Forecasting

Challenge

EBIT is one of the most crucial KPIs for guiding future business success. In times of increasing economic upheaval, our client, an international automotive corporation, faced a laborious and time-consuming manual process for determining EBIT. The project's goal was to support manual planning with an automated, model-based planning system as a decision support tool, thereby reducing planning uncertainties.

Approach

To forecast the future monthly EBIT, various potential influencing factors were consolidated and prepared from the client's controlling systems. These included various internal planning metrics, inventory and sales figures, seasonal effects, and external economic indicators. Due to the limited historical data, methods for selecting the most suitable influencing factors (feature selection) were conducted. The identified relevant factors were combined in a machine learning model. The trained and optimized model was then implemented into the client's IT infrastructure. A forecast for the next one to three months is now generated automatically every month and made available to decision-makers via a dashboard.

Result

The EBIT forecasts from our model provide a supportive component in the client's monthly planning process and regularly outperform the client's manual planning. The automation of the process ensures that forecasts are available before manual planning, allowing for early, proactive action.