The Reorder Point That Was Set in 2019 and Never Questioned Again

Most operators treat a reorder point like a fact. Calculate it once, trust it forever. It doesn’t work that way. A reorder point is a snapshot of demand, lead time, and risk on the day you set it, and every one of those inputs drifts.

If the number on the screen was written in 2019 and nobody has touched it since, it isn’t a threshold anymore – it’s a fossil.

The frustrating part is that the fossil still fires alerts, still triggers purchase orders, and still looks like it’s doing its job. That’s how a stockout in 2026 gets blamed on the supplier when the real problem was set six years earlier.

The Number Was Correct When You Wrote It Down

Rewind to whenever the reorder point was first set. Someone pulled a year of usage, asked the supplier how long a replenishment took, added a safety cushion, and posted the number in the system. On that day, it was defensible.

Demand looked steady. The supplier was quoting a lead time everyone believed. Freight was predictable. The safety stock reflected a world where the biggest risk was a slow week, not a closed port.

That number carried a lot of hidden assumptions with it. Weekly usage would keep looking like last year’s weekly usage. The supplier would keep quoting the same lead time. Freight lanes would keep behaving.

None of those assumptions came with an expiration date printed on them, so nobody knew when to check.

Then the Inputs Quietly Moved

Between 2019 and now, every input under that formula shifted. Demand patterns changed as customers reordered on different cycles. Suppliers consolidated, moved plants, or added minimum order quantities. Lead times stretched, sometimes by weeks.

Post-pandemic research on global supply chains has documented that lead times from Chinese suppliers rose sharply after 2019, and that shift never fully reversed for a lot of categories.

The reorder point in the system doesn’t know any of that. It’s still solving yesterday’s problem with yesterday’s numbers. The alert still fires at the same on-hand quantity. The PO still goes out on the same day.

The part just shows up later. Or not at all.

The Symptoms Show Up Somewhere Else First

The tell is rarely a screaming alarm. It’s a pattern of small annoyances nobody connects back to the reorder point:

• Expedited freight. Air freight or hot-shot deliveries start showing up on parts that used to arrive on a truck. Nobody flags it because each one looks like a one-off.
• Rush POs at odd hours. Buyers spend Monday morning chasing parts that should have been ordered three weeks ago. That’s a lead-time assumption that broke and never got updated.
• Growing safety stock. Operators start hiding a little extra in a corner of the rack because they don’t trust the trigger anymore. The real reorder point becomes a bin nobody wants to talk about.
• Stockouts on steady movers. The most predictable parts run out first, because those are the ones nobody thought to review.

Rebuilding the Number Is Not a Project, It’s a Habit

The fix is less dramatic than most people expect. You don’t need a platform overhaul. You need to treat the reorder point as a living calculation and put it on a review cadence.

Post-2020 supply chains have settled into structurally longer lead times, and supplier collaboration has become part of the recalibration, not a separate task. That means calling the supplier once a quarter and asking what a real replenishment looks like today, not what the contract says.

The mechanics are boring, and that’s the point. Pull the last six to twelve months of usage. Confirm the current lead time with the supplier, in writing. Reset safety stock against the variability you’ve actually seen, not the variability you remember.

Update the number. Move on. This is also where modern replenishment tools earn their keep, since the whole loop, from usage data to supplier lead time to trigger point, can be recalculated on a schedule instead of a heroic once-a-year sweep.

A Healthy Reorder Point Looks Boring on Purpose

The reorder points that work are the ones that get quarterly attention from someone who understands the part. A short review, a small adjustment, and a note in the record explaining why the number moved. Over a year, that produces triggers that reflect the current world, not a snapshot from six years ago.

The 2019 number was not wrong when it was written. It’s wrong now because nobody asked it to prove itself again. Ask it. If it can’t defend itself against current usage and current lead times, replace it and put the next one on the calendar.