Why is the shelf empty? Out-of-stock root-cause analysis across four warehouses
One year of daily stock, orders and deliveries – to measure the real service level, put a price on lost sales and find out what actually causes stock-outs.
Business question
Sales reported frequent out-of-stocks, while the warehouse reported high stock levels. Every department had its own explanation: wrong forecasts, late trucks, too little safety stock.
Goal: measure service level per warehouse and category, quantify lost sales and assign each stock-out to a root cause – so that actions target the real problem.
Data & method
Fill rate = units delivered from stock ÷ units demanded, per week, warehouse and category.
Unserved demand × selling price for every SKU-warehouse-week with a stock-out.
Open purchase order overdue → late delivery; delivered < ordered → short delivery; demand > 140% of forecast → demand spike; otherwise → safety stock too low.
Interactive dashboard
Weekly fill rate
Root causes of stock-outs
By warehouse
Fill rate by category
Top 10 critical SKUs
Key findings
of lost revenue is caused by late supplier deliveries – not by forecasting. Forecast-related spikes explain only 20%.
of lost revenue falls on A items: the flat one-week safety stock is too thin exactly where it matters most.
Beverages has the lowest fill rate – seasonal peaks plus long, unreliable lead times.
has the lowest fill rate (95.6%). Differences between warehouses are small – the problem is upstream, not in one site.
Recommendations
Tackle supplier reliability first
Start a supplier review with the four least reliable suppliers (see case study 04) – this addresses the biggest share of lost sales.
Raise safety stock for A items from risky suppliers
Include lead-time variability in the safety-stock formula instead of a flat one week of cover.
Weekly exception list for planners
Flag overdue purchase orders for A items every Monday so planners can expedite or reallocate stock between warehouses.
Agree promo volumes with Sales
Get promotion volumes at least one lead time in advance to remove the remaining demand-spike stock-outs.
Data
Every stock-out event with warehouse, supplier, lost units and root cause: