Key takeaways
See EAM vs CMMS for tracking spare parts.
The typical storeroom holds two populations of parts that look similar on the shelf and are very different in purpose. Operating spares, bearings, belts, seals, sensors, get consumed regularly. They have a measurable failure rate, a min/max level that follows it, and a reorder pattern driven by consumption. The framework that handles them is the one covered in our broader spare-parts piece.
Insurance spares are different. A specific high-voltage transformer for the main electrical supply. A custom-machined drive shaft for the largest mill on site. A spare PLC for the line's master controller. These parts may sit in the storeroom for five years without being touched.
They exist not because the team expects to use them, but because the cost of being without one, if and when it fails, is too high to accept on lead time.
Most plants store both populations the same way and apply the same min/max logic to both. The result: insurance spares either get over-counted (the storeroom is full of low-failure parts that are not really insurance and not really operating) or under-covered (the genuine insurance items are missing because their consumption history shows zero).
Two thresholds, both qualitative until the numbers go in.
If a single failure of this part would cost the plant more than the asset's annual operating cost divided by 12, the part is in the insurance category. A bearing that costs 4 hours of downtime to replace is operating. A transformer whose failure means a multi-week outage and a major production loss is insurance.
The calculation uses the same cost-per-production-minute number from the article on the preventive maintenance schedule .
If the part can be sourced in under two weeks even from the OEM, it is operating, not insurance. Two weeks is short enough that most plants can run on workarounds, partial production, alternative routing, hand-built temporary substitutes. Above two weeks, the workarounds run out and the plant is exposed.
Both thresholds need to be true. A high-failure-cost part with a 3-day lead time is operating (just source it when it fails). A low-failure-cost part with a 12-week lead time is also operating (accept the gap; it does not warrant insurance treatment).
The cost of holding an insurance spare is the carrying cost: storage, capital tied up, occasional inventory counting. For most parts, this is 20-30% of the part's value per year.
The cost of not holding one is the expected loss: probability of failure × cost of the resulting outage. For a transformer with a 1% annual failure probability and a large outage cost, the expected annual loss is a small fraction of that outage. If the carrying cost of keeping it on the shelf is several times that expected loss, the straight math says do not hold it.
But the straight math is misleading on tail-risk items. The expected-loss calculation assumes the plant is risk-neutral; in practice, a single catastrophic outage is much worse than several years of small expected losses, because the plant has to absorb it all at once. Insurance spares earn their keep on the variance, not the expected value. The article on manufacturing KPIs covers the trend-vs-variance tradeoff in adjacent metrics.
The practical rule: hold the spare if the carrying cost is less than 4-6x the expected annual loss. Above that ratio, the math is too unfavourable even with risk aversion baked in.
The single biggest operational mistake with insurance spares is putting them into the regular consumption-based reorder system. Operating-spare reorder is triggered when consumption drops the stock to the min level. Insurance-spare consumption is approximately zero. The trigger never fires.
So when an insurance spare finally gets used, usually at 2am in a crisis, there is no replacement order in the queue. The plant has solved the immediate crisis by consuming the insurance and is now exposed to the next failure for the entire lead time of the replacement order. Multiple months, sometimes.
The fix is structural: insurance spares are flagged as a separate inventory class with an event-based reorder rule. Any consumption of an insurance-class part triggers an immediate replacement order, regardless of remaining stock. The reorder is automatic, not dependent on someone noticing the gap. The work order management system needs to mark insurance-class parts when they are consumed against a work order so the trigger fires.
An annual audit of the insurance-spare population is one of the easiest high-leverage exercises in the storeroom. The audit asks four questions per part:
Most plants find 10-15% of their insurance-spare population needs reclassification each year. Without the audit, the population grows stale and the storeroom value grows without the protection growing. The piece on root cause analysis often surfaces these reclassifications, when a failure mode is identified, the relevant insurance spare may need to be added or upgraded.
For a typical mid-market plant with 50 critical assets, the insurance-spare list is 12-30 SKUs. Each one has:
That is the whole list. It is short on purpose. Plants whose insurance-spare list is over 60 SKUs have usually misclassified a chunk as insurance when they belong in operating.
The classification works in any inventory system that allows two distinct part classes and an event-based reorder rule.
Where a unified OEE + CMMS platform helps is in three places: the failure-cost calculation comes from the OEE event stream, the asset-link audit is automatic when assets are decommissioned, and the consumption trigger fires reliably because the work order, the parts and the asset are all in one database.
Fabrico is built so the insurance-spare class is a first-class concept rather than a workaround in the consumption-based system. To see what the classification would look like on your storeroom, book a demo .
Some OEMs offer "guaranteed availability" agreements for critical parts at a fee. The math is the same one above: compare the agreement cost to the expected loss × risk-aversion multiplier, plus the carrying cost saved. For parts where the OEM holds local stock, the agreement is often cheaper than carrying the part ourselves. For parts that would ship from overseas regardless, holding ourselves is usually better.
This is the trickiest category. The OEM may not support the part anymore; substitutes may not exist. The right answer is usually to hold one spare and to start the asset retirement calculation in parallel. The piece on preventive maintenance schedules covers how PM tuning interacts with retirement timing.
Almost never. If the failure rate is high enough to justify two on the shelf, the part is operating, not insurance. The exception is parts with a destructive failure mode that could damage a second one in close succession; even then, two is the maximum.
The reliability engineer where the role exists; the maintenance manager otherwise. Single owner. Annual audit. Changes go through the same review process as adding or removing an operating-spare SKU.
Letting the storeroom team decide which parts are insurance. The classification is an asset-criticality and failure-cost decision, not a storeroom decision. The maintenance manager or reliability engineer owns the classification; the storeroom executes the storage and reorder rules.
Login
Blog
