Goal: Provide understandable predictions in your reporting system and their usage to automate subsequent processes.
- outlets sales forecasting - risk model for quick loans/factoring
- POS newspapers distribution optimization
- POS lottery tickets distribution optimization
- demand forecasting for supply chain optimization
We use high-performance, interpretable machine learning models (e.g. TFT) which consider product relations (cannibalization), seasonality, promotion, new products and POS.
Input data sources
Most of the data are ingested from SAP S/4 or ECC.
- Historical sales (transactions – Products, Outlets)
- Product data (Attributes, descriptions)
- Marketing campaigns & advertising (SalesForce Marketing cloud data)
- POS information (Location, setup..,)
- Competition (Pricing of competition)
- Public data (Weather, lockdowns…)