Somewhere in your ERP there is a number, sometimes called a reorder point, sometimes an order point, that decides when a purchase order gets triggered for every item you stock. Someone set that number once, usually when the item was first loaded into the system. Most operations never revisit it after that. The trouble is that the two things the number depends on, how fast the item sells and how long it takes to get more, keep changing quietly in the background, while the number that assumes they haven’t stays exactly where it was set.
What is your reorder point actually supposed to represent?
Hand-waving the arithmetic, a reorder point is answering one question: how much will we sell between the moment we place an order and the moment it arrives, plus a buffer for when that period runs longer or sales run higher than usual. In plain terms: reorder point equals normal demand during your lead time, plus a safety margin, and once stock falls to that level a new order should go out. Both halves of that number, the demand estimate and the lead time, are measurements of the world as it was when someone last calculated them. Neither one is a law of physics.
Why does that number go stale without anyone deciding it should?
Two ordinary things happen to every item over time, and either one is enough to make the old reorder point wrong.
Demand moves. A product gets a new institutional customer, loses a competitor’s shelf space, goes into a slow season, or simply drifts up or down the way most demand does. None of that requires anyone to make a decision. It just happens, and the reorder point calculated a year or two ago keeps assuming the old rate.
Lead time moves too, often by more and in the direction you feel the most. A supplier changes factories, a shipping lane gets congested, a new customs rule adds a week, or your old supplier quietly slips from four weeks to six. Nobody re-runs the reorder point math when that happens, because the lead time field in the ERP is usually treated as a fixed setting rather than something that needs watching. Lead time behaves in one particular way that makes this worse: it has a floor, production plus transit cannot go below some minimum, but no ceiling. A bad month can add weeks and nothing ever gives them back. When a reorder point is quietly built on last year’s average lead time, the years it is most wrong are exactly the years a delay actually hits, and the effect does not stay contained to one item. The original paper on this documented that order variance can run larger than actual sales variance, and that the distortion tends to increase the further upstream you look (Lee, Padmanabhan & Whang, “Information Distortion in a Supply Chain: The Bullwhip Effect,” Management Science 43(4), 1997, pp. 546-558). A stale reorder point, reacting late to a real change in demand or lead time, is exactly the kind of local rule that feeds a distorted signal into that chain.
Which of your items are most exposed to this?
Not all of them equally. Three patterns are worth checking first:
- Imported items with long or inconsistent freight. The lead time input is the one most likely to have quietly changed and the least likely to have been re-entered.
- Items whose demand pattern has genuinely shifted, a new key account, a channel that grew or shrank, a seasonal pattern that used to be flat.
- Anything nobody has touched since it was set up. If the reorder point field has the same value it had two years ago, that is not evidence it is still right. It is evidence nobody has checked.
What can you check on your own numbers this week?
An ERP export and a spreadsheet, nothing more.
- Pull the reorder point and its last-updated date for your top items by revenue. Anything untouched for over a year is due a look, regardless of how it has performed so far.
- Compare the lead time your ERP has on file against what actually happened on your last handful of purchase orders for that supplier. A gap of more than a few days on an imported item is common, and it means every reorder point built on the old number is off by that much.
- Check actual sales for the last two quarters against the demand rate the reorder point assumes. If they have drifted apart, the buffer is now sized for a business you no longer run.
- Start with the items that have stocked out or overstocked in the last year. That is not a coincidence list, it is where a stale assumption already cost you something.
This is a maintenance habit, not a project. The items worth rechecking are usually a short list, not your whole catalog, and finding them is the first thing our per-unit inventory work does before recommending anything be rebuilt.