Forecast of the purchasing volume

An automated, data-driven ECV forecast reduces deviations from the actual, creates transparency and increases planning reliability.

Challenge

  • The aim is to develop a forecasting model that makes the purchasing volume per business area transparent and homogeneous.

  • Previous decentralised forecasts of the FCA by business areas with partly considerable deviations from the actual situation

Solution

  • Linking of purchase orders, material master, general ledger accounts, planning data (TM1) and aggregating the data

  • Mapping of the historical ECV at different granularities per month level

  • Exploratory correlation analysis between actual and planned data

  • Creation of a forecast model using ML methods on the basis of historical ECV data

Result

  • Through an automated, data-driven EKV forecast for product groups per business area, deviations from the actual are reduced, transparency is created and planning reliability is increased.

Are you interested in your own use cases?