A Reliable 21 Days Beats a Volatile 10

Ask a buyer which vendor is better: one who ships in 10 days, or one who ships in 21? Everyone picks the 10. Now add the fine print — the 10-day vendor actually delivers anywhere from 6 to 30 days depending on the month, while the 21-day vendor hits 20–22 like a metronome.

The 21-day vendor is better. It's not close. And most purchasing organizations are set up to reward the wrong one.

Length is a planning input. Variance is a risk.

A long lead time you can trust is just arithmetic: order earlier, carry a bit more pipeline stock, done. It costs something, but it costs a known something you can price into the reorder point once.

A variable lead time can't be planned around — only buffered against. Every safety stock formula worth using takes lead-time variance as a direct input, and it's brutal about it: uncertainty in when the order arrives inflates the buffer faster than uncertainty in demand for many SKUs. You end up carrying stock not because customers might buy more, but because your vendor might show up whenever.

You can schedule around slow. You can only insure against unpredictable — and the premium is inventory.

Why the volatile vendor wins anyway

Because of how vendors get measured. The stated lead time lives in the item master and gets negotiated at the annual review. Variability lives nowhere. Most ERPs happily store one lead-time number per vendor or item and call it a day — so the vendor who quotes 10 days looks better on every screen than the one who quotes 21, even while the 10-day quote fails a third of the time.

Meanwhile the damage from variance shows up downstream, disguised as other problems: expedite fees, stockouts on items with "correct" reorder points, buyers padding mins by gut feel because they've been burned. The padding is safety stock too — it's just unmanaged, invisible safety stock that nobody sized deliberately.

What to actually track

Three numbers per vendor, from PO date to receipt date, rolling twelve months:

• Median actual lead time — not the quote. The quote is marketing.
• The spread — a simple standard deviation, or even just the 90th percentile minus the median. This is the number that sets your buffer.
• Trend — a vendor drifting from 12±2 to 14±6 is telling you about their own supply problems before they tell you.

Put the spread next to the quote at the vendor review. "You quote 10 days; you deliver between 6 and 30; here's what that range costs us in safety stock across your line" is a very different negotiation than arguing the quote down to 9 — and it points at the fix that actually helps. A vendor who commits to 15 days reliably is usually offering you more inventory reduction than one who promises 8 and means "sometimes."

The takeaway

Lead time length costs you pipeline inventory once. Lead time variance costs you safety stock forever, plus every expedite and every stockout the buffer fails to catch. Negotiate reliability first, duration second — and if your system can only store one lead-time number, at least make it the honest one.