The challenge
A key retail client with more than 100,000 SKUs was losing sales to stock-outs in some categories while overstocking others.
International vendors had lead times running into months, so the client needed enough safety stock to cover variability in lead and review times.
What we did
- Built an automated SAS forecasting and inventory planning model at SKU-market level.
- Used the forecasts to set safety stock, order-up-to levels and order quantities.
- The model considered current inventory, vendor lead times, review times and forecast demand, and set order quantities every week.
The result
After rollout on one product category, the client saw a significant reduction in lost sales from stock-outs and cost savings from lower overstock.
How we would build it today
- Probabilistic demand forecasts feeding multi-echelon inventory optimisation.
- A planner web app that flags only the exceptions that need a human.
- An AI assistant that explains why each order quantity was recommended.
This describes our current toolkit for the same problem. It is not part of the original engagement.