Forecasts, stock and lead times become an order someone can approve
International purchasing involves long and highly variable lead times. Stock across multiple warehouses, outstanding deliveries, bills of materials and Amazon FBA requirements add further complexity. Spreadsheets could no longer keep these factors together.
Order Brain holds them in one model. Every suggestion shows its calculation: demand, stock cover, lead time and minimum quantity are visible to the person approving it. The return path to inventory management is part of the same system.
The starting point: blanket buffers without a daily plan
Safety stock was set using blanket rules, without a daily projection. Despite the buffers, some products ran out while others tied up capital on the shelf. Purchasing, production, the main warehouse and Amazon FBA were considered separately and maintained in separate spreadsheets. It was difficult to reconstruct why one order quantity differed from another.
The turning point: calculate, suggest and approve
Forecast demand
A weighted forecast based on historical sales provides the baseline. Days without stock distort that picture. We exclude them so that the model does not read a forced sales gap as a genuine drop in demand. An AI forecast with seasonal and calendar factors runs alongside it, and both models are compared.
Projected stock
Current stock, expected demand, outstanding deliveries and delivery calendars feed a daily simulation. It shows buffer and target stock cover and flags future shortages before they occur.
Generated suggestions
Order Brain determines ordering dates and quantities, assigns suppliers and ordering schedules, and respects pack sizes and minimum quantities. Profitability-based criticality classes establish priority, while unavailable products receive separate treatment. Bill-of-materials products generate production suggestions; Amazon FBA products generate transfer suggestions.
Exported approvals
Approved purchase orders and transfers return to inventory management as CSV files through JTL-Ameise. A safeguard prevents duplicate exports. The history records which plan was exported and when.
McFilter is a medium-sized e-commerce company with its own purchasing, product development, warehouse and shipping operations.
It sells across Europe through multiple marketplaces and its own shops. Unimatrix brings the operational data from these channels together in a shared data platform on which Order Brain runs.
