Machine Learning
Demand Forecasting & Inventory Optimization
Client: Retailer - 400 stores
The project
Development of a 90-day demand forecasting model for 400 stores and 12,000 product references, accounting for seasonality, promotions, weather and local events.
Architecture & implementation
Coupling of the forecasting model with an inventory optimization algorithm to minimize stockouts and overstock, factoring in logistics constraints, supplier lead times and storage costs.
Monitoring & adoption
Dashboard deployed for 400 store managers: personalized replenishment recommendations, imminent stockout alerts and promotion suggestions to clear excess inventory.
Results & production rollout
Results over 12 months: 32% fewer stockouts, 24% less overstock and 11% higher gross margin. The project was rolled out across the entire network and has become a daily tool for the teams.
Screenshots
