DAX-40 Company : Accurate Revenue Forecasts through Data-Driven Forecasting with External Drivers

Many companies face the challenge of reliably predicting their revenues in global markets. A leading DAX company had previously relied on manually created forecasts within its divisions, which were heavily dependent on the knowledge of individual experts. This approach was time-consuming, slow to react to market changes, and only limitedly considered external influencing factors. To provide a more solid basis for decision-making, the existing forecast was to be supplemented by a data-driven second opinion – a solution that combines internal and external data and uses modern machine learning methods. The goal was to make the forecasting process more accurate, efficient, and flexible.

We developed a rolling forecast with a horizon of up to 18 months, delivering results updated weekly. To achieve this, we combined classical time series methods with modern machine learning approaches, supplemented by over 100 external drivers from various clusters. Through the ensemble approach with market- and product-specific models, both short-term and medium-term forecasts could be integrated. The solution was productively implemented on the customer’s infrastructure and now automatically delivers data-based forecasts for all divisions of the DAX-40 group. 

The ML-based forecasting model significantly increases accuracy compared to previous manual forecasts. Systematic backtesting demonstrated that the models clearly outperform the previous manual forecasts in several areas. Particularly relevant for users: the ML engine explains which drivers it identified as relevant for generating the forecast and therefore included. Business units can thus see which factors – whether micro, macroeconomic, or internal – most strongly influence revenue in the respective divisions. The application impresses with a user-friendly frontend and is actively used by the target users. Even the first forecasts served as a second opinion at the division level and helped the company to react more quickly, data-driven, and market-oriented.