A maintenance KPI is a measurable indicator that shows how well your maintenance performs: how available your installations are, how reliably your assets run, how much work you have under control and what it costs. You use KPIs to steer, not to collect. The value is not in the number itself, but in the decision that follows from it. This article lists the key maintenance KPIs, groups them by what they measure, and shows how to pick a small, meaningful set.
Most KPIs fall into four groups: availability and reliability, effectiveness, work control and cost. The figure below summarises them.
Availability and reliability
This group answers the most direct question: does the installation run when you need it, and how often does it fail?
Availability (%) is the share of planned production time that an asset was actually usable. You calculate it as uptime divided by uptime plus downtime. An availability of 96% means the asset was unintentionally down 4% of the time. It is an outcome KPI: it captures the combined effect of failures, repairs and waiting time in a single figure.
MTBF (Mean Time Between Failures) measures reliability: the average running time between two failures. If MTBF rises, an asset fails less often. MTTR (Mean Time To Repair) measures recovery: the average time to resolve a failure, from report to working installation. A long MTTR often points to missing parts, unclear procedures or too little capacity, rather than to the fault itself.
Availability, MTBF and MTTR are connected: higher MTBF and lower MTTR together lead to higher availability. Measuring them side by side shows you not only that an asset fails, but whether the problem lies in the frequency or in the repair time. The formulas and pitfalls are covered in more detail in the article on MTBF and MTTR.
Effectiveness: OEE
Availability tells you whether a machine ran, but not whether it ran well. For that you use OEE (Overall Equipment Effectiveness). OEE combines three factors into one percentage: availability (did the machine run), performance (did it run at speed) and quality (was the product good). You multiply the three: a machine scoring 90% on all three lands at roughly 73% OEE.
The strength of OEE lies in that breakdown. A low performance score points to minor stops and speed loss; a low quality score to rejects and rework. That tells you where the loss sits. OEE is most useful on production lines with a clear cycle time; on standalone assets without a fixed production norm it says less. What OEE measures exactly and how to calculate the factors is explained in what OEE is.
Work control
This group is not about the machine, but about your own process: do you have the work under control, or are you running behind the facts?
The preventive/corrective ratio (PM ratio) shows what share of your maintenance hours goes to planned work and what share to breakdowns. An organisation that mostly works reactively recognises itself in a high share of corrective work. Many maintenance teams steer towards a majority of preventive and condition-based work, but the right ratio depends on your installations. The difference between the two is covered in the article on preventive vs corrective maintenance.
Backlog is the volume of approved but not yet executed work, usually expressed in labour hours. A backlog is not a problem; a growing or invisible one is. The figure shows whether your capacity and work intake are in balance.
Schedule compliance measures what share of planned work was actually carried out as planned. Low schedule compliance means urgent work keeps overriding your plan, a sign that you are still working reactively. Together these KPIs tell you whether your maintenance is controllable or whether it is happening to you.
Cost
Ultimately maintenance has to stay affordable without giving up reliability. Two KPIs keep an eye on that.
Maintenance cost as a percentage of replacement asset value (RAV) sets your annual maintenance cost against the replacement value of your assets. This ratio makes cost comparable between sites or years, regardless of the size of your installations. A rising RAV without better reliability is a signal to look at where the money is going.
Cost per asset brings spending down to the level of the individual machine or installation. That shows which assets cost disproportionately much and which are candidates for replacement or a different maintenance strategy. Cost figures only become useful alongside reliability KPIs: cheap maintenance that drives up failures is not a saving.
Pick few, but meaningful KPIs
Measuring more is not steering better. A dashboard full of numbers nobody uses is a collection of vanity metrics: it looks complete, but leads to no decision. A KPI without action is dead weight.
So choose a small number of KPIs that match a concrete goal. If you want to reduce unplanned downtime, availability, MTBF, MTTR and the PM ratio are enough. If you want a grip on cost, add RAV and cost per asset. Make sure every KPI has an owner, a target value and a clear source, and that someone acts on it periodically. A KPI that stays red for three months with no consequence no longer measures anything useful.
For those wanting to benchmark against a standard: the European norm EN 15341 describes a set of standardised maintenance KPIs (economic, technical and organisational) and is available through NEN. Such a norm is not an obligation, but a handy framework for deciding which indicators are relevant for your organisation.
Reliable KPIs start with reliable data. When failures, work orders and hours are logged consistently in one system, most of these indicators emerge on their own in reports and dashboards. Data logged loosely across scattered lists leads to figures nobody trusts, and therefore to KPIs nobody steers by.
Frequently asked questions
How many KPIs should you measure in maintenance?
Fewer than you think. Start with four to six KPIs tied to one concrete goal, each with an owner and a target. Only expand once you actively use those to make decisions. A large dashboard nobody consults steers nothing.
What is the difference between a KPI and a vanity metric?
A KPI leads to a decision or action; a vanity metric looks complete but changes nothing about what you do. The test is simple: if a number changes, does anything change in your approach? If not, it is dead weight.
Which maintenance KPI is the most important?
There is no universally most important KPI; it depends on your goal. For availability problems MTBF and MTTR lead, for production lines OEE, for cost control RAV. Choose the KPI that measures the problem you want to solve.