Average lead time isn't the problem — variability is

"This supplier takes 30 days" is the sentence almost everyone uses to plan orders. The problem is that single number hides what actually matters: a supplier who reliably delivers between 28 and 32 days is a completely different risk than another one with the same 30-day average who sometimes delivers in 20 days and sometimes in 50.

Here we explain why lead time variability — not just the average — is what should set your safety buffer, and how to calculate it from your own order history.

What on-time delivery benchmarks say

Industry analyses put "best-in-class" on-time delivery (OTD) at 95% to 99%+ — typical of Tier 1 automotive suppliers, aerospace, and high-volume electronics manufacturing. A 90-94% rate is considered acceptable for mid-market manufacturers, and anything below 90% is read as a signal of a systemic scheduling, equipment-reliability, or supply-chain problem — not a one-off incident.

A widely cited Aberdeen Group study found that best-in-class companies with 95%+ on-time delivery achieve 15% to 20% higher customer retention than laggards below 85% — your suppliers' reliability flows through, with a lag, into your own reliability toward your customers.

The average hides the variability

Here's the key point almost nobody measures: a supplier with a 30-day average lead time can still create real risk if actual deliveries range from 20 to 50 days. Two suppliers with the exact same average may need completely different safety stock depending on how much their lead times actually swing order to order.

Reducing lead time variability — not just its average — is, together with lowering the cost of placing orders, one of the most direct levers for improving inventory management without carrying more stock.

The math: your lead time safety buffer

Instead of trusting the "quoted" lead time, calculate the standard deviation of your actual order-by-order lead times and apply a buffer proportional to the service level you want to guarantee:

Illustrative example (sample figures — swap in your own in the calculator below): a supplier with a 30-day average lead time but a real standard deviation of 8 days (calculated from your own order history, not their sales sheet) needs, for a 95% service level (z ≈ 1.65), a buffer of 1.65 × 8 ≈ 13.2 days. In other words: to avoid stocking out 95% of the time, you should plan around ≈ 43 days, not the "official" 30 — 13 days more than you're probably using in your reorder-point calculation today.

That buffer isn't conservative overkill — it's the difference between a reliable supplier with a long lead time and another supplier with the same average lead time who leaves you empty-handed once every twenty orders.

What to do with this

• Stop using the "quoted" lead time in your reorder-point calculations — use the real average lead time and its standard deviation, calculated from your own order history.
• If two suppliers have the same average lead time, compare them on variability too: the more predictable one, even with a somewhat longer average, may need less total safety stock.
• If a supplier's on-time delivery rate drops below 90%, treat it as a structural warning sign, not one-off bad luck — recalculate your safety buffer for that supplier right away.

Sources

• On-time delivery (OTD) benchmarks by performance tier: SourceDay, "On-Time Delivery (OTD): Definition, Formula & Benchmarks".
• Lead time variability concept and the 20–50 day range on a 30-day average example: SourceDay, "Lead Time Variability: How Manufacturers Reduce Supplier Risk".
• Customer retention tied to 95%+ on-time delivery (Aberdeen Group): MetricHQ, "On-time Delivery (OTD)".

The average lead time, standard deviation, and service level in the worked example are illustrative — built from sample numbers to explain the calculation, not a statistic about any real supplier. Swap in your own order history using the calculator.