Key takeaways
In most plants the spare-parts policy is a layer cake of decisions made years ago. Some parts are stocked because the OEM said to. Some because a maintenance lead two roles ago had a bad weekend. Some because the supplier offered a kit discount. Almost none because the actual failure pattern justifies it. The result is a storeroom that is full of low-failure-rate items and out of stock on the high-failure-rate ones.
The cost shows up in two places. The first is working capital tied up in dead stock — often 20–30% of a storeroom's value sits as parts that have not been pulled in over a year, and in a neglected storeroom it can run higher. The second is unplanned downtime extended by hours or days because the part that actually failed has to be sourced from scratch. Both costs are invisible until you map them, and most plants never do.
How long from "we need this part" to "the part is on site." For an in-region distributor that is usually one to three days. For an OEM-direct, often two to six weeks. For a custom-fabricated part — longer, sometimes months. Lead time is the first number because it bounds everything else.
What does the asset do when this part fails? Three buckets:
Criticality should come from the asset's role in the line, not from the operator's emotional response when it last failed. A single shared asset hierarchy in the work order management system makes this assessment honest because the asset's actual contribution to OEE is already visible.
How many times has this part failed in the last 12 to 24 months? This is the number that most spare-parts policies skip, because nobody has been tracking part-level consumption. A CMMS that records the parts consumed by each work order makes the number available; a CMMS that does not, does not. See the article on manufacturing KPIs for how this connects to MTBF.
Cross the three numbers and the stocking decision falls out:
The single largest source of dead stock is spares for assets that no longer exist on the floor. An asset gets replaced, decommissioned, or relocated; the spares stay. Sometimes for years. The fix is quarterly:
Most plants find 8–15% of storeroom value falls into this bucket the first time they do it. The follow-up rule is what matters: any time an asset is decommissioned in the CMMS, the spare-parts list is reviewed within 30 days. The article on preventive maintenance schedule covers the related cleanup of PMs against decommissioned assets.
The min/max rule for stocked parts is usually overthought. A working approach:
The dangerous version is treating min/max as a one-time set-up exercise. Failure rates drift, lead times change, suppliers go out of business. Without a quarterly review the min/max becomes fiction within 18 months.
Every part consumption gets logged against a work order. Every work order is linked to an asset. The asset is in a hierarchy that maps to a line that maps to an OEE event stream. This is the chain that makes the three numbers — lead time, criticality, failure rate — calculable from real data rather than guessed.
Plants that do not have this chain end up running spare-parts policies on someone's memory, which decays. Plants that do have it can recalibrate min/max levels every quarter from actual consumption, and can identify obsolete spares automatically rather than at the next physical inventory. The article on root cause analysis covers how repeated consumption of the same spare against the same asset feeds the deeper "is the part the right design" question.
The decision grid works in any CMMS that tracks parts against work orders, but the loop closes cleanly when the OEE event, the work order, the parts consumed and the asset hierarchy live in one database. Fabrico is built that way — the failure rate per spare is a live calculation rather than a quarterly project, and the obsolete-spare check is automatic when an asset is decommissioned. To see how the inventory would look against your live consumption data, book a demo.
By count, hundreds. By line items, usually 1,200–3,000 SKUs depending on asset count. By value as a percentage of replacement asset value (RAV), top-performing plants keep spare-parts inventory at or below roughly 1.5%. Persistently above that is usually a sign of over-stocking worth investigating rather than a target to grow into.
Treat it as one input, not the answer. The OEM does not know your failure rates, your lead times to your region, or your asset criticality. Some recommendations are essential; many are reflexive. Evaluate each against the three numbers.
For most mid-market plants, parts tracking inside the CMMS is enough. The separate inventory system makes sense at the point when you have multiple storerooms with shared SKUs, or where the supply chain side needs visibility into part demand. Below that threshold, two systems just create reconciliation work.
Identify them in the storeroom as multi-asset, and calculate failure rate against the sum of failures on all linked assets, not against any single one. The decision grid does not change; the inputs just aggregate.
Setting it once and not reviewing it. Lead times, failure rates and asset registers all drift. A policy set in 2022 will be wrong in 2026, sometimes by a factor of two. Quarterly review is the difference between a working storeroom and an expensive one.
Вход
Блог
