… for a store Jeans brand using ERP + external data
- A retailer wanted to forecast demand over nine monthly horizons for core store SKUs in Jeans to drive various business decisions (e.g., Fabric buy – 9 months prior to season, cut-orders – 4 months prior)
- Cross-sectional time-series solution was found to have highest forecast accuracy (after testing several alternative models)
- Store aggregate data, pricing and macro-economic data were used as inputs
- The solution was based on 5 cross-sections. Variable construction is depicted below.