A spare part is never the real expense. The expense is the production line that stops while a maintenance technician waits for a component worth a fraction of an hour of lost output. That asymmetry is what makes spare-parts logistics a design problem rather than a purchasing problem: the goal is not to hold the cheapest inventory, it is to make the right part reach the right machine in the shortest time that the business can justify paying for.
Two failure modes sit at opposite ends of the same decision. Hold too much and capital, shelf space, and obsolescence risk accumulate. Hold too little and every failure becomes an expedited freight event and an unplanned stoppage. The way out is not a single stocking level applied across the catalogue; it is classification, followed by a different policy for each class.
Price tells you what a part costs. Criticality tells you what its absence costs. The two are unrelated, which is exactly why a criticality classification is worth doing before any stocking number is fixed.
| Class | Typical examples | Reasonable policy |
|---|---|---|
| Line-stopping consumables | Belts, filters, wear pads, seals, photo-eyes, fuses | Stock on site, reorder on consumption, never allowed to reach zero |
| Long-lead critical parts | Gearboxes, servo drives, specialised rollers, controllers | Hold one unit as insurance or agree a guaranteed exchange with the supplier |
| Interchangeable standard parts | Bearings, fasteners, cylinders, standard motors | Small local stock plus a distributor agreement; the distributor holds depth |
| Slow-moving repairables | Motors, pumps, and drives that can be rebuilt | Rotate a repaired unit back into stock instead of buying new every time |
| Obsolescent and end-of-life parts | Discontinued boards, obsolete drive models | Decide deliberately: final buy, last-time-buy, or plan the machine retrofit |
The classification only works if it is reviewed. Machines get replaced, suppliers discontinue lines, and a part that was a consumable on the old line can become unavailable on the new one. A review tied to the annual maintenance plan is enough.
There is no single right location; there are four, and most operations need more than one.
Machine side. Small, cheap, frequently replaced items stored at the line cut walking time to seconds. The cost is space and the risk of unrecorded consumption, so the cabinet has to be counted on a rhythm.
Central stores. Depth for common parts across the site, with a system that shows where each item is. A stores area that cannot be searched is functionally a warehouse with no inventory record.
The supplier. A distributor or manufacturer holding stock on your behalf. This is the cheapest place to keep slow-moving items, but only where the lead time is genuinely reliable enough to be planned against.
An outsourced fulfilment partner. For distributed networks, remote sites, or field-service fleets, storage and despatch can be handled off site against agreed service levels, with kitting and courier choice managed for you. That removes the equipment and staffing decision from the maintenance budget, which is what makes it attractive to operations that are not in the warehousing business.
Most of the time lost on a planned job is spent assembling what the job needs rather than doing the work. Kitting attacks that directly: pick every part, seal, fastener, and consumable a specific task requires into one container before the technician arrives, with the procedure attached. The technician then starts work instead of hunting.
Staging goes further and is worth it for repeated work. If a machine fails the same way twice a year, the kit should be defined once and rebuilt from stock whenever it is used, so the second repair is faster than the first. Staging also exposes the parts that are not in stock before the job begins, which is the only point at which the problem can still be solved cheaply.
Reverse logistics is where spare-parts operations usually leak value, because the outbound process is designed and the inbound process is not. Three flows need a defined route: failed parts that go to repair, parts that go back to a supplier for credit or core refund, and parts that arrive wrong and must be returned without holding up the repair.
Two policies make the difference. A repairable part is only valuable if something actually sends it for repair, so the return path has to be part of the technician's job rather than an afterthought. And a return needs a decision deadline: a part that sits undecided in a corner for a month has effectively been paid for twice.
A partner can only commit to what you can describe. Give them the part list with dimensions and storage requirements, the expected order profile by class, the service level each class needs measured from order to despatch, the sites or technicians that have to be served, and the return and repair flows. Add how orders will arrive, since a system that assumes manual email orders will not scale when the volume doubles.
Ask what happens on a stock-out, how inventory accuracy is reported and how often, who absorbs the cost of a mis-pick, and what the exit looks like if you change provider. Those questions are more informative than the rate card.
Hold them yourself where the part is cheap, moves often, and the machine cannot wait. Outsource the slow-moving tail, the geographically dispersed stock, and the despatch and returns work, because that is where the fixed cost of doing it in house stops being justified.
When the value tied up in slow-moving items exceeds the cost of the downtime those items prevent, the balance has tipped. That comparison is possible once parts are classified by criticality; it is impossible while they are treated as one undifferentiated catalogue.
No, but you need a record that is trusted and searchable, and a discipline for updating it. The system choice matters far less than whether consumption is recorded at the moment it happens, because every downstream decision depends on that record being true.
Spare-part availability is part of the machine specification, not an afterthought once the equipment is installed. At ZH Industrial Machinery we plan service access and the critical spares list with the machine, and where a customer would rather not run storage and despatch themselves, movement of those parts can be handled by a fulfilment and warehousing partner against agreed service levels.
The operating model described here is third-party logistics: a provider holding and moving stock on a client's behalf. National forwarder and 3PL practice is coordinated internationally through FIATA, and the applied research on how logistics performance differs by market is published by the MIT Center for Transportation and Logistics.