Spare Parts Inventory Management for Repair Shops

Overstock is the comfortable mistake. Buying ten screens because the unit price drops feels like good management, and it is — for the three references you genuinely turn. For everything else it converts cash into inventory that ages, and phone and laptop parts age quickly: a model leaves the repair cycle far faster than most shops plan for.

Understock is the visible mistake. A repair waits four days for a part, the device occupies a shelf, the customer calls twice, and the technician handles the same job three times instead of once. The cost never appears in the accounts as a line item, which is exactly why it goes unmanaged.

Most shops make both mistakes simultaneously, on different references. They are deep in stock on parts that used to move and short on the ones that move now. The fix is not to buy more or less overall — it is to change what the buying decision is based on.

Export twelve months of part consumption and sort by number of movements. You will almost certainly find that a small minority of references accounts for the large majority of the volume. That top group is where availability matters and where a stockout genuinely costs a job. Everything below it is a per-job order, whatever the volume discount says.

Set three simple tiers. Tier A: parts used weekly, always in stock, reordered automatically. Tier B: parts used monthly, one unit on the shelf, reordered on use. Tier C: everything else, ordered only against a confirmed job, ideally against a paid deposit. Three tiers are enough; a shop that invents seven will maintain none of them.

Review the tiers every quarter, not every year. Device models move between tiers faster than people expect, and the shift is usually noticed too late — a part stays in tier A for months after the model stopped arriving. A quarterly half hour with the consumption report prevents most of the dead stock a shop accumulates.

For dead stock already on the shelf, decide rather than wait. Use it as a warranty spare, sell it to another shop, list it, or write it off. Every month you keep hoping it will move, it is worth a little less, and it is occupying the slot a moving part needs.

The reorder point is average consumption during the supplier's lead time, plus a safety margin. If you use four of a screen per week and your supplier delivers in five working days, you consume roughly four units while waiting, so reordering at six or seven is sensible. That is the whole calculation, and it beats reordering when the shelf looks empty.

Use the supplier's real lead time, not the advertised one. Track what actually arrived and when, for a couple of months. Advertised 48-hour delivery that lands in five days on a third of orders is a five-day lead time for planning purposes, and treating it as 48 hours is what produces the stockouts nobody can explain.

Raise the safety margin around known peaks: back-to-school, holiday periods, the weeks after a new model launches, and anything seasonal in your trade — scooter and bike work in particular collapses and explodes with the weather. Lower it again afterwards, deliberately, because a temporary margin left in place quietly becomes overstock.

Keep a second source for your top twenty references, even if you rarely use it. A single supplier is a single point of failure, and the failure never happens at a convenient moment. Having an account already open elsewhere converts a two-week outage into a two-day inconvenience.

Judge suppliers on landed cost and reliability together. A part three euros cheaper that arrives four days later, or that has a higher defect rate, is more expensive once you count the second bench pass and the customer who was told the wrong date. Record which supplier a part came from, so that when a batch turns out to be poor you know precisely which jobs are exposed.

Record supplier references alongside your own naming. Every supplier calls the same screen something different, and a shop that only keeps its internal label spends real time re-identifying parts at reorder. Keeping both references on the same product record removes a small friction that occurs several times a week.

A part that leaves stock without being attached to a repair makes two things impossible: knowing what the job earned, and trusting your stock figure. Both failures compound quietly — the count drifts, and margins become guesses defended with conviction.

Once the link exists, real margin becomes visible, and it is often not what people assumed. Count the part, the consumables, the bench time, and the share of jobs that come back under warranty. A repair type with an attractive headline price and a five per cent return rate can earn less than a cheaper one that never comes back.

The same link gives you warranty traceability. When a batch of parts turns out to be defective, you want to list the affected repairs in seconds rather than reconstruct them from invoices. That is the difference between a controlled recall of twelve customers and a month of surprises.

Annual full counts are painful, disruptive and out of date within a fortnight. Cycle counting works better: count your tier A references every month, tier B every quarter, everything else once a year. The counting takes fifteen minutes a week and your figures stay usable all year rather than for one week in January.

Investigate discrepancies instead of just correcting them. A recurring gap on one reference has a cause — a part taken for a warranty job without recording it, a mis-scan at reception, a return that never came back into stock. Adjusting the number without finding the cause guarantees the same gap next month.

In SERVO, parts are attached to the repair that consumes them, incoming deliveries are received by scanning, and both your reference and the supplier's are kept on the product. That is not the only way to do it — a disciplined shop can run this on a spreadsheet — but the discipline is the requirement, not the tool. The wider workflow is covered in our guide to running a repair shop.