How do you set minimum stock levels for critical spares?
The empty shelf on breakdown day and keeping two of everything are both failures of policy. A minimum is built from lead time, real usage and what a stockout costs, and wear parts and failure parts are stocked on different logic.
A core pin shears on a Thursday and the toolroom goes to the cabinet to find the shelf empty. Someone fitted the last one weeks ago, nobody flagged it, and now a running job waits two days for a courier. The reaction is predictable and wrong. Somebody says keep two of everything from now on, and the shop swings from holding too little to tying up cash in shelves of parts it may never fit. Neither the empty shelf nor the overstuffed one is a stocking policy. A minimum stock level is a number you can defend, and it is not a gut feel.
What a minimum stock level is actually made of
Strip the guesswork out and a minimum is built from three plain quantities. The first is the replacement lead time, how long it genuinely takes to get another one in hand, not the supplier's hopeful quote but the real door-to-door time including your own receiving and any inspection. The second is the consumption rate, how many of this part you actually get through in that lead time, read off your own history rather than a feeling. Multiply those two and you have the stock that covers normal demand while a reorder is on its way. The third quantity is the one people skip, a safety buffer sized not by the part but by the consequence of being caught short. A cheap seal whose stockout parks a whole press line earns a bigger buffer than an expensive part you can work around for a day. Set the minimum below the sum of those three and you are betting the next failure lands on a good week.
Wear parts and failure parts are not stocked the same way
The single biggest improvement is to stop treating every spare with one rule and split them by how they die. Wear parts go predictably, the punches, guides, springs and seals that fatigue against the hits, and their consumption is a rate you can read straight off the history, so their minimum falls out of the burn rate almost mechanically. Failure parts are the opposite, the random breakages that give no warning, and for those a consumption rate is close to meaningless. You stock them against risk, how likely the failure is over a period against what it costs you when it lands, which is an insurance decision rather than an inventory average. Muddling the two is how shops both overstock the wear items they could reorder on a rhythm and understock the rare failure part that actually stops a line. Some of both can also sit as supplier-held consignment stock, on your shelf but not your money until you pull it, which lets you hold a deeper margin on the parts that would hurt most without carrying all the cash.
What to pull together before you set a number
- The real replacement lead time, measured door to door on the last few orders, not the quoted one.
- Your own consumption history for the part, so the usage rate is evidence and not a guess.
- What a stockout actually costs, a parked line and idle labour or a minor inconvenience, because that sets the buffer.
- Whether the part is a predictable wear item or a random failure item, because the two are stocked on different logic.
- Which die or line the part belongs to, so the minimum lives against the tooling it protects rather than floating in a general store.
The number is only as good as the movements behind it
A minimum written once and never revisited drifts out of date the first time a die runs harder or a supplier gets slower. What keeps it honest is recording every movement, each part booked in and each one fitted, against both the spare and the tooling it went into. Do that and the consumption rate corrects itself from real usage, the current count is always known, and a part sliding under its minimum becomes a signal the moment it happens instead of a discovery at the next breakdown. That is how VoraTool holds critical spares, a record per spare with its minimum on it and movements logged against the spare and the die, so the same notes that stop the shop running out are the ones that reset the minimum as reality changes. When a part drops below its floor it shows as needing attention rather than waiting to be noticed, and when an auditor asks whether your critical spares are controlled, the minimum, the movements and what each part was fitted to come straight off those records.
So the answer to how you set a minimum is to build it, not guess it. Take the real lead time and the real usage, add a buffer sized by what a stockout costs, stock the predictable wear parts on their rate and the random failure parts on their risk, and then let recorded movements keep the number true. A minimum you can explain in those terms is one you can defend to a plant manager asking why the cash is on the shelf and to an auditor asking whether your critical spares are actually under control.
Tooling, job cards and preventive maintenance with the history written as the work happens.
Talk to someone who understands manufacturing control and audit evidence. We do not do generic demos.