Optimizing spare parts stock means balancing stockout cost (the price of not having a critical part when a machine fails) against holding cost (the price of keeping that part on a shelf), then stocking each item at the level where the two forces cancel out. Every spare you buy is insurance against downtime, but insurance is not free. Carry too much and cash, warehouse space, and obsolescence eat your budget. Carry too little and a single missing bearing can idle a production line for days. The goal is not zero stockouts or zero inventory: it is the lowest total cost across both.
Stockout cost is the financial damage that occurs when a needed spare is unavailable at the moment of failure. It is far more than the price of the part expedited overnight. A realistic figure captures:
Because downtime scales with how critical the machine is, the same part can carry a very different stockout cost depending on where it sits. A gearbox on a constraint machine (see the theory of constraints) is not the same as an identical gearbox on a redundant backup line. Pairing your stockout estimate with real reliability metrics like MTBF and MTTR tells you both how often a failure happens and how long the machine stays down waiting for the part.
Holding cost, also called carrying cost, is the annual expense of keeping a part in stock, usually expressed as a percentage of the part's value. Typical components:
A common working range for annual holding cost is 15 percent to 30 percent of part value. So a spare worth 4,000 euros held for a year can quietly cost 600 to 1,200 euros just to sit there. Tracking inventory turnover exposes the spares that never move and are pure carrying cost.
The two costs pull in opposite directions. As you raise stock and service level, holding cost climbs steadily while expected stockout cost falls. Total cost is the sum of the two, and it forms a U-shaped curve with a clear minimum. The optimal stock sits at the bottom of that U, not at either extreme.
For a single expensive critical spare, the decision is often binary: hold one or hold none. The clean way to decide is to compare the annual cost of holding one unit against the annual expected cost of not holding it:
For high-volume, lower-value consumables (filters, seals, fasteners), a continuous model fits better. There you set a reorder point, size replenishment with the economic order quantity, and buffer demand variability with calculated safety stock rather than a single hold-or-not call.
A packaging line runs on a specialized drive motor. The numbers:
Holding cost of stocking one motor: 6,000 euros multiplied by 0.25 equals 1,500 euros per year.
Downtime if the part is not stocked: a 5-day wait at 24 hours is 120 hours. At 900 euros per hour, one stockout event costs 108,000 euros. Even assuming half the failures happen when a repair window absorbs part of the delay, call it a conservative 60,000 euros per event.
Expected annual stockout cost: 25 percent probability multiplied by 60,000 euros equals 15,000 euros per year.
Verdict: 15,000 euros of expected stockout cost dwarfs 1,500 euros of holding cost by ten to one. Stock the motor. Now flip one variable: if the same motor sat on a non-bottleneck line with a hot standby, downtime cost might be near zero, expected stockout cost collapses, and the 1,500 euro holding cost is no longer justified. Same part, opposite decision, because criticality changed the stockout term.
You cannot run a full trade-off calculation on ten thousand line items, so segment first. Run an ABC analysis to separate the vital few from the trivial many, and use a Pareto analysis to confirm that a small share of parts drives most of the risk and value. Then apply the right policy per tier:
Criticality should also be informed by failure analysis. An FMEA flags which failure modes carry the highest severity, and understanding where each asset sits on the bathtub curve tells you whether failures are random, wear-driven, or early-life, which changes how much buffer you need.
A good trade-off model is only as accurate as the data feeding it, and that data usually lives in three disconnected places: production, maintenance, and the storeroom. Fabrico brings them together as a real-time data foundation. As a field-ready CMMS, it manages work orders, assets, preventive schedules, and spare parts in one place, so every part consumption is logged against the asset and failure that triggered it. Its real-time OEE and production monitoring quantify what an hour of downtime on a given line actually costs, which is the number your stockout calculation lives or dies on. Fabrico's computer vision can even monitor machines that have no PLC, extending accurate data capture to older assets. Because it is EU-built with EU data residency, that operational history stays under European governance. Fabrico does not automatically reorder parts for you, but it gives you the verified consumption rates, failure frequencies, and downtime costs that make the holding-versus-stockout math trustworthy. Explore the CMMS solution overview to see how spare parts and work orders connect.
Yes. For cheap parts on non-critical or redundant equipment, the downtime cost is so low that carrying inventory, especially slow-moving obsolescence-prone stock, costs more than an occasional wait for delivery. The trade-off model makes that call explicit: when expected stockout cost sits below holding cost, planning to reorder on demand is the rational, lower-total-cost choice.
Start with the lost production margin: units the machine would have produced per hour multiplied by the contribution margin per unit, but only if that output cannot be recovered elsewhere. Add idled labor and any penalty or scrap costs. Machines that are bottlenecks or feed the whole line carry the highest figures, which is why the same part warrants very different stocking decisions in different positions.
Lead time is a multiplier on stockout cost. A longer supplier lead time means more downtime hours per event, which pushes the expected stockout cost up and makes holding the spare more attractive. Shortening lead time through a local supplier or a vendor-managed arrangement (VMI) can flip a hold decision back to reorder-on-demand and free up carrying cost.
Ready to base your spare parts decisions on real downtime costs and verified failure data instead of guesswork? Book a Fabrico demo and see how connected OEE and CMMS data sharpens every holding-versus-stockout call.
Login
Blog