Statistical demand forecasting

Switzerland, Italy, Brazil

For Holcim Switzerland forecasting is a central need in order to set up logistics & production processes in the right manner. A good planning has a direct impact on optimization and cost reduction.

The planning process as we know it from decades, rely on the information sales employees provide manually. Every over or under forecasting has a negative impact on our business. In order to increase forecast accuracy and reduce uncertainty, Statistical Demand Forecasting has been implemented.

The Statistical Demand Forecasting (SDF) is an integrated machine learning solution to support optimization and the planning process. The SDF is a forecasting process of making predictions of the future demand based on past and present data. As a main difference to the manual process, the SDF can provide weekly/monthly automated planning and includes parameter like net working days and weather forecast.

The pilot project has successfully proven that SDF provides a higher accuracy with lower manual effort. Today 5 cement plants make use of the SDF as a main data source for short term planning (weekly planning).


40% reduction of unnatural moves

Increased forecast accuracy by +8%


Project currently in global scale-up phase

LATAM roll-out ongoing (Brazil, Mexico & Argentina are live)

9 new countries in the pipeline

Related case studies