Lead Time Variability:
the silent killer of inventory

Here is the equation that governs safety stock in most planning tools: SS = z × σ_LT × d, where σ_LT is the standard deviation of lead time and d is average daily demand. The average lead time does not appear in the numerator. It never did. It affects the formula only indirectly, through the demand term. The primary driver of safety stock cost is variability, not speed.

A supplier delivering in 30 days on average, with actual arrivals ranging from 15 to 45 days, forces you to hold roughly three times the safety stock of a supplier with the same 30-day average but a range of 27 to 33 days. Same cost on the purchase order. Same position on your lead time report. Completely different working capital profile. The planning team sees two identical suppliers. The balance sheet sees something else entirely.

The reason this stays invisible for so long is that most supplier scorecards measure On-Time Delivery (OTD) against a target window. A supplier hitting 92% OTD against a 5-day window looks reliable. But if the 8% misses are clustered in a specific season, or if the "on time" deliveries span the full 5-day window rather than arriving predictably on day 3, the variance is significant and it is not captured anywhere. OTD with a generous window is a compliance metric. Coefficient of Variation (CV = standard deviation divided by average lead time) is a planning metric. They measure different things, and confusing them is expensive.

Same average lead time (30 days). Different planning reality.

The safety stock formula is, at its core, a working capital formula. Every additional day of buffer stock has a carrying cost: capital tied up, warehouse space consumed, obsolescence risk accumulated. For a team managing €50M in inventory, the delta between planning with a low-variability supplier (CV=0.10) and a high-variability one (CV=0.33) at equivalent demand volumes can represent €80,000 to €120,000 in additional safety stock per supplier. Across a portfolio of 15 to 20 strategic suppliers, that is not a planning detail. That is a working capital line that belongs in the CFO conversation.

There is a second effect that rarely appears in the analysis: Days Inventory Outstanding (DIO). When goods arrive unpredictably, they do not align cleanly with demand peaks. Some arrive too early and sit in the warehouse. Some arrive late and force emergency replenishment or stockouts. Both scenarios increase DIO. The cash conversion cycle lengthens not because your payment terms changed, but because your arrival variability is silently expanding the holding period. The P&L does not show the cause. The planning model does not flag the mechanism. The connection stays invisible until someone looks at variability data and asks the right question.

The problem surfaced during a stockout review, not during planning. The supplier had solid OTD. The average lead time was within target. Everything looked fine on paper until we asked a different question: not when deliveries arrived on average, but how far they spread around that average. The answer made the safety stock miscalculation obvious in about ten minutes. We had been solving the right formula with the wrong input for months.

When I added the coefficient of variation to the supplier scorecard alongside OTD, three things happened. Procurement started selecting suppliers differently, because reliability now meant something more precise than "usually on time." Safety stock parameters came down by 12% across the affected lanes. And the planning team stopped firefighting every Monday morning, because the buffers were sized on the actual distribution rather than a comfortable average. That shift did not require new tools or a system change. It required measuring what the formula actually needs, not what was easiest to track.

The pattern I have seen consistently is that teams discover this problem during a stockout post-mortem, not during planning. The supplier has good OTD. The average lead time is within target. But the stockout happened because a delivery arrived four days later than the average, and the safety stock was sized for the average. At that point the team asks: why was our buffer not enough? The answer is sitting in the lead time distribution they never looked at.