Skip to content

How it works

How do you build an inventory system for one-of-one parts?

Every inventory system you can buy assumes you can order more. When quantity is permanently one, the reorder point, the safety stock and the forecast all become meaningless - and what replaces them is not obvious.

5 min read

Every inventory system in general circulation was built on one assumption: when you run low, you order more.

That assumption produces the reorder point, safety stock, lead time, economic order quantity and demand forecasting. It is the entire vocabulary of inventory management, and for a salvage yard every word of it is inapplicable.

Why does quantity-one break the model?

  • A reorder point tells you when to buy more. There is no more - the alternator came off one car.
  • Safety stock is a buffer against demand variance. You cannot hold a buffer of a thing that exists once.
  • Demand forecasting predicts how many you will sell. The answer is one or zero, decided by whether a specific buyer with a specific car finds you.

None of that is a limitation of the software. It is the software correctly solving a different problem. Bolt a salvage yard onto it and every item is permanently at critical stock level, and every alert is meaningless.

What replaces quantity?

Instead ofYou trackBecause
QuantityIdentityThe unit is a physical object with a history, not a line with a number
Reorder pointProvenanceThe donor vehicle is your fitment, your condition claim and your dispute defence
Safety stockLocationA part nobody can find is functionally sold to nobody
ForecastStateAvailable, listed, reserved, sold, ended - the transitions are where money is lost

The one rule that matters more than the rest

Exactly one record owns the count. If the number that says "this alternator exists" lives in your inventory AND the eBay listing AND the storefront AND a spreadsheet, you have four numbers and no fact. They agree the moment you reconcile them and drift immediately afterwards.

For stock with quantities, drift is an annoyance - you oversell by one and order more. For quantity one, drift is a cancellation, a refund, a defect, and a buyer who will not come back.

Channels read from one record. The clay path is the sale ending the others - an event, not a scheduled sync.

Once the count lives in one place, ending listings stops being a synchronisation task and becomes a consequence. The sale IS the event. The difference between those two designs is the difference between a window that is small and a window that is zero - on a five-minute sync you have a five-minute window, all day, every day.

What to do today, in a spreadsheet

  1. One sheet owns the count, and every channel is treated as a copy of it. Not two sheets. Not a sheet and your memory. When they disagree, the sheet is right and the channel is stale.
  2. A column for the donor vehicle, filled in at teardown. It costs seconds then and is unrecoverable later. The single highest-value column in your whole system.
  3. End the other listing before you pack the box, not after. The gap between sold and ended is the entire risk, and packing takes long enough for a second buyer to arrive.

None of that is a system. It is the habits a system would enforce, and they work until the day you are busiest - which is also the day two buyers are most likely to collide.

A used part is one physical thing on one shelf. A system that cannot say that in its data model will keep telling you that you have two.

Stop typing the same listing three times. One email when it opens. Nothing else, ever.