Reducing Lead-Time Variability: The Cheapest Way to Cut Safety Stock

When we ask supply chain leaders why safety stock is high, the answer is almost always 'inaccurate demand forecasts'. But for many items, especially imported ones or those depending on foreign suppliers, the larger part of safety stock comes from lead-time variability, not demand variability. This variability is measured less and therefore managed less.

The math is simple: safety stock grows with the square root of two components — demand variability over the lead time, and the square of average demand times lead-time variability. When average demand is high and lead time unstable, the second component dominates. The practical result is that cutting the standard deviation of lead time from, say, 10 days to 5 can reduce safety stock more than any forecast-improvement project.

The first step is measurement. For each important supplier, record the gap between order date and actual receipt in the warehouse for the last 12 to 24 orders and compute the mean and standard deviation. Most organizations only have the contractual lead time in the system, which differs from reality; the difference between the two is itself an important finding.

After measuring, separate the sources of variability. Lead time consists of several parts: order processing, production or preparation at the supplier, transport, customs clearance and incoming inspection. The variability is usually concentrated in one or two, such as customs clearance or waiting for quality inspection at the warehouse. Fixing that one part is far more effective than a general effort to 'improve the supplier'.

Common solutions: a fixed, defined shipping schedule instead of on-demand shipping; regular ordering instead of large irregular orders (which reduces the supplier's load swings); contracts with a lead-time commitment and a follow-up mechanism; running clearance or inspection in parallel rather than in sequence; and, for items where foreign lead-time variability is too costly, a local supplier option.

A common mistake is solving the problem by inflating the lead time assumed in the system ('safer to put 45 days'). This does not remove variability; it only raises pipeline inventory and safety stock and hides where the problem originates. Every protective number in the system is capital paid for not knowing.

A practical starting point: choose the ten items with the highest safety-stock capital, measure their actual lead-time variability, and compute each component's (demand and lead time) share of safety stock. If lead time has the larger share, this list is your first work plan for cutting inventory without raising stockout risk.