Sizing safety stock on lead time, not demand.

Rajat Gupta · 6 min read

Ask most planners how much safety stock to hold and they will reach for demand variability — how much sales bounce around a forecast. In an import-dependent operation, that is the wrong variable. Demand you can usually see coming. What ambushes you is the supply side: the shipment that takes 62 days instead of 38, the vessel that skips a port call, the consignment held at the terminal for documentation. Your buffer exists to absorb variance, and the variance that matters is in the lead time.

Size the buffer on the variability of arrival, not the variability of demand. Stock-outs are caused by late trucks more often than by surprise orders.

What the data usually shows

Pull the last twenty purchase orders for any imported A-class material and compute the actual door-to-door lead time for each. The spread is routinely two to three times wider than the spread in monthly consumption for the same item. A material consumed at a steady 100 tonnes a month with arrivals ranging from five to eleven weeks needs its buffer built around those six weeks of arrival uncertainty — not around a demand curve that barely moves.

The working method

• For A-class materials, drive requirements from the production forecast through the bill of materials, and size safety stock on measured lead-time variance — mean arrival time plus a service-level multiple of its standard deviation.
• For C-class items, skip the ceremony: simple min-max on consumption history is cheaper than the analysis.
• Re-measure lead times quarterly. Ports, shipping lines and clearing agents drift; a buffer set on last year’s lanes protects against a world that no longer exists.

The constraint nobody models

In many markets the binding constraint is not the warehouse but the bank: letters of credit, currency availability, the funding window for the next order. When foreign-exchange access is uneven, the reorder decision is a treasury decision, and safety stock is partly a hedge against your own funding pipeline. Plan it jointly — procurement and treasury on the same calendar — and the buffer stops being a guess and becomes a policy.