Maintenance KPIs: 12 metrics that reveal eliability, cost and team performance

Maintenance KPIs (maintenance key performance indicators) are measurable indicators that show how effectively an organisation maintains its physical assets, controls cost, delivers planned work and limits the risk of unplanned failure. A strong set usually contains 10–15 key indicators – not dozens of metrics without an owner or a target.

Throughout this article the abbreviation PM stands for preventive maintenance – the planned inspections, servicing and interventions carried out before a failure occurs. That is where indicator names such as PM compliance come from.

What are maintenance KPIs?

Maintenance KPIs are measurable indicators used to assess the effectiveness, efficiency and cost of the maintenance function. They are not simply a list of numbers: a usable KPI has a definition, a formula, a data source, an owner, a reporting frequency and a decision attached to it. Without those six attributes, a number on a dashboard is a curiosity rather than a management tool.

Maintenance KPIs help teams to:

• measure asset reliability;
• reduce reactive work;
• justify the maintenance budget;
• detect planning problems;
• control maintenance cost;
• assess the effectiveness of repairs and inspections;
• compare results across lines, plants or periods;
• prioritise improvement actions;
• track maintenance maturity.

Maintenance KPI vs. maintenance metric – what is the difference?

A maintenance metric is any measurable value describing maintenance work. A KPI is a metric tied to a target, an accountable owner and a decision.

Example: the number of work orders closed per month is a metric. It becomes a KPI once the organisation defines the expected result, names who is accountable for it and states what decision follows from a deviation.

Concept	Meaning	Example
Metric	Any measurable value describing maintenance activity	Number of work orders closed in a month
KPI	A metric with a target, an owner and a decision attached	PM compliance above 90%, reviewed weekly by the planner
Target	The expected value of the KPI	Availability above 95%
Decision	The action triggered by a deviation	Re-plan PM routes if compliance falls below 85%

How many maintenance KPIs should you track?

Most plants are best served by 10–15 KPIs at plant level, with a smaller subset – typically four to six – reported to management. Fewer indicators mean each one gets a genuine owner and a genuine review cadence.

Indicators should be selected to match the maintenance strategy, asset criticality and the plant’s operating model. A discrete manufacturing site needs a different set from a process plant, and a different set again from energy infrastructure or a utility.

A good KPI set should show both the consequences of failures and repairs and the quality of planned work. That is why lagging indicators such as MTBF or availability need to sit alongside leading indicators such as PM compliance or the share of planned work.

This article focuses on the KPIs themselves and how to read them. The wider context includes predictive maintenance and the taxonomy of maintenance strategies (types of maintenance).

Leading and lagging maintenance KPIs

Lagging KPIs describe what has already happened: how often assets failed, how long repairs took, how much maintenance cost. Leading KPIs describe the work that shapes those outcomes: whether PMs were delivered on time, whether the schedule held, how much of the workload was planned rather than reactive.

Reporting only lagging indicators means the organisation learns about its problems after they have cost money. Reporting only leading indicators means the organisation may deliver plans faithfully without ever proving that reliability improved. Both groups are needed, and they should be read together.

KPI group	What it shows	Examples
Leading	The quality and delivery of planned maintenance work	PM compliance, schedule compliance, share of planned work, backlog
Lagging	The outcome of reliability and repair performance	MTBF, MTTR, availability, maintenance cost as % of RAV

The KPI chain: PM compliance → MTBF → availability → OEE

These indicators are not independent. Delivering preventive maintenance on time reduces the number of unplanned failures, which lifts MTBF. Higher MTBF combined with a stable MTTR increases availability, and availability feeds directly into the availability component of OEE. Reading the chain in that order explains why a plant with weak PM compliance rarely fixes availability by pushing repair crews harder.

Reliability KPIs

Reliability KPIs answer one question: how dependably do the assets run? They are lagging indicators, and they are the numbers operations and management care about most.

Mean time between failures – MTBF

MTBF measures the average operating time between two failures of the same asset or asset group. It is the primary reliability indicator for repairable equipment.

Formula: MTBF = total operating time / number of failures

Indicative target: 500–5,000 hours depending on the industry and asset class. Higher is better. A rising MTBF over several periods is one of the strongest signals that a reliability programme is working.

How to read it: a falling MTBF on a critical asset points to a root cause that inspections are not catching – often lubrication, alignment, contamination or an operating condition outside the design envelope. Compare MTBF by failure mode rather than by asset alone; an average across ten failure modes hides the one that matters.

Asset availability

Availability shows the share of required time an asset is actually able to perform its function. It is the single clearest expression of what maintenance contributes to production time.

Formula: Availability = MTBF / (MTBF + MTTR)

Indicative target: above 90%; mature organisations run above 95%. Availability responds to two levers at once – fewer failures and faster restoration – so it should always be read next to MTBF and MTTR rather than on its own.

How to read it: if availability is flat while MTBF improves, repair and logistics time is absorbing the gain. If availability drops without a change in MTTR, failure frequency is rising and the PM programme needs review.

Maintenance efficiency KPIs

Efficiency KPIs describe how well the maintenance organisation converts its own capacity into completed, effective work. They are where planning quality, spare-parts availability and technical competence become visible.

Mean time to repair – MTTR

MTTR measures the average time needed to restore an asset to service after a failure, including diagnosis, repair and functional testing. It is the primary measure of maintainability.

Formula: MTTR = total repair time / number of repairs

Indicative target: under 2 hours is good, under 1 hour is very good, though the realistic figure depends heavily on asset complexity. Lower is better.

How to read it: a high MTTR is rarely a technician-speed problem. The usual causes are waiting for parts, waiting for permits or isolation, missing documentation and travel time. Break MTTR down into waiting time and hands-on time before drawing conclusions.

Wrench time

Wrench time is the share of a technician’s paid time spent with tools in hand on the asset, as opposed to travelling, waiting, searching for parts or looking for information.

Formula: Wrench time = actual hands-on repair time / total maintenance labour time × 100%

Indicative target: 30–40% is typical; above 55% indicates a very mature planning and kitting process. A low figure is an organisational loss, not a motivation problem.

How to read it: wrench time is the fastest way to quantify the cost of poor planning. If it sits in the low twenties, the improvement lever is job preparation – kitted parts, pre-written procedures, permits raised in advance – not additional headcount.

First-time fix rate

The first-time fix rate shows how many jobs are completed on the first visit, without a return trip for parts, tools, information or a second attempt at diagnosis.

Formula: First-time fix rate = repairs completed on the first visit / all repairs × 100%

Indicative target: above 80%.

How to read it: a low rate points to one of three gaps – diagnosis, spare-parts availability or the quality of information reaching the technician. Splitting repeat visits by reason code tells you which of the three to fix first.

Maintenance cost KPIs

Cost KPIs connect maintenance activity to money. Used alone they invite the wrong conclusion – the cheapest maintenance function is usually the one quietly building a reliability debt – so they belong next to the reliability group.

Maintenance cost as a percentage of replacement asset value (RAV)

This indicator expresses annual maintenance spend as a share of what it would cost to replace the asset base. It is the most widely used benchmark for judging whether maintenance spend is proportionate.

Formula: Maintenance cost % RAV = annual maintenance cost / replacement asset value × 100%

Indicative target: 2–5% of RAV.

How to read it: too low is as informative as too high. A figure well under 2% usually means deferred work and an accumulating reliability debt; a figure well above 5% suggests inefficiency, a reactive workload or an ageing asset base. Confirm that the RAV figure itself is current before comparing against any benchmark.

Maintenance cost per unit of production

This KPI ties maintenance spend to output, which makes it the most persuasive cost indicator in front of operations and finance.

Formula: Maintenance cost per unit = total maintenance cost / number of units produced

Indicative target: industry-specific; the trend should decline. Judge it as a trend, never as an absolute.

How to read it: the figure moves with production volume, so a drop during a high-output month is not automatically an efficiency gain. Normalise for volume before presenting it to management.

Spare parts inventory turnover

Turnover shows how many times the spare-parts inventory is consumed and replenished over a year. It is the balance point between availability of parts and capital tied up on shelves.

Formula: Spare parts turnover = annual value of parts consumed / average inventory value

Indicative target: 1.0–3.0 turns per year.

How to read it: too high risks stock-outs on critical items; too low means frozen capital and obsolescence. Segment the calculation between critical insurance spares – which are meant to sit still – and consumables, where slow turnover is genuine waste.

Compliance and maturity KPIs

This group is made up of leading indicators. They describe the discipline of the maintenance process – whether planned work is genuinely delivered – and they are the earliest place to see a reliability problem forming.

PM compliance

PM compliance shows what share of scheduled preventive maintenance work orders were actually completed within the required window.

Formula: PM compliance = completed PM work orders / scheduled PM work orders × 100%

Indicative target: above 90%; mature organisations run above 95%.

How to read it: define the compliance window before you report the number – a PM completed three weeks late on a monthly route is not compliant in any meaningful sense. Persistently low compliance is the most reliable predictor of rising unplanned failures.

Schedule compliance

Schedule compliance widens the lens from PM alone to the whole planned workload: how much of the scheduled work was executed on time.

Formula: Schedule compliance = scheduled work orders completed on time / all scheduled work orders × 100%

Indicative target: above 90%.

How to read it: a gap between high PM compliance and low schedule compliance means PMs are being protected while corrective planned work is repeatedly bumped – which is how a backlog of known defects builds up unnoticed.

Planned versus reactive work

This ratio shows what share of the total workload was planned in advance rather than executed as a reaction to a breakdown. It is the clearest single measure of maintenance maturity.

Formula: Share of planned work = planned work orders / all work orders × 100%

Indicative target: above 80%; mature organisations exceed 90%.

How to read it: a low ratio describes a reactive culture, and it explains poor wrench time, poor first-time fix and volatile cost at the same time. Improving this one indicator usually moves several others.

Planning KPIs

Planning KPIs describe the load in front of the team and the organisation’s ability to work through it at a controlled pace.

Maintenance backlog

The backlog is the volume of approved work that has been identified but not yet executed, expressed in labour hours rather than in a count of work orders.

Formula: Backlog = labour hours of approved, outstanding work orders

Indicative target: 2–4 weeks of crew capacity.

How to read it: a backlog near zero is not a success – it usually means inspections are not generating findings, or that work is being executed the moment it is raised, with no planning window. A backlog beyond six weeks means either a capacity shortfall or a prioritisation failure, and defects will start failing before they are reached. Report it in weeks of capacity so the number is comparable between crews.

Maintenance KPI summary table

The twelve indicators below are grouped by category, with formulas, indicative target ranges and the reading each one supports.

KPI	Category	Formula	Indicative target range	Reading
MTBF – mean time between failures	Reliability	Total operating time / number of failures	500–5,000 h depending on industry	Higher = greater reliability
Availability	Reliability	MTBF / (MTBF + MTTR)	Above 90%, very mature above 95%	Direct maintenance impact on production time
MTTR – mean time to repair	Efficiency	Total repair time / number of repairs	Under 2 h good, under 1 h very good	Lower = faster restoration to service
Wrench time	Efficiency	Actual hands-on repair time / total maintenance labour time × 100%	30–40% typical, above 55% very mature	Low = organisational losses
First-time fix rate	Efficiency	Repairs completed on first visit / all repairs × 100%	Above 80%	Low = diagnosis, parts or information gap
Maintenance cost % RAV	Cost	Annual maintenance cost / RAV × 100%	2–5% of RAV	Too low = under-maintenance risk; too high = possible inefficiency
Maintenance cost per unit	Cost	Total maintenance cost / units produced	Industry-specific; trend should decline	Links maintenance to production value
Spare parts turnover	Cost	Annual value of parts consumed / average inventory value	1.0–3.0 turns per year	Too high = stock-out risk; too low = frozen capital
PM compliance	Compliance	Completed PM work orders / scheduled PM work orders × 100%	Above 90%, very mature above 95%	Low = higher risk of unplanned failures
Schedule compliance	Compliance	Scheduled work orders completed on time / all scheduled work orders × 100%	Above 90%	Shows delivery of the whole planned workload
Share of planned work	Compliance	Planned work orders / all work orders × 100%	Above 80%, very mature above 90%	Low = reactive culture
Maintenance backlog	Planning	Labour hours of approved, outstanding work orders	2–4 weeks of capacity	Shows planning capability and team load

How to design a maintenance KPI dashboard

A dashboard is not a list of every number the CMMS can produce. It should answer three questions on one screen: are the assets reliable, is the team working effectively, and is the cost proportionate. Everything else belongs in a drill-down.

A workable structure uses three layers. The management layer carries four to six KPIs with targets and trend arrows. The plant layer carries the full set of 10–15 indicators, broken down by line or area. The analysis layer carries the underlying work-order and failure data for root-cause work.Each KPI on the dashboard needs a target line, a trend over at least twelve periods, and a named owner. A KPI without a target cannot be breached, and a KPI without an owner will not be acted on.

Data sources for maintenance KPIs

KPI quality is limited by data quality. In practice the numbers come from a small set of systems, and the weakest link is almost always work-order discipline in the CMMS.

Data source	What it provides	KPIs it feeds
CMMS / EAM	Work orders, PM schedules, labour hours, failure records	MTTR, MTBF, PM compliance, schedule compliance, backlog, share of planned work
Production and control systems (SCADA, MES, historian)	Runtime, downtime, output volumes	Availability, OEE availability component, cost per unit, production impact of failures
Cost and ERP systems	Labour cost, contractor cost, materials, asset register	Maintenance cost % RAV, cost per unit
Inventory / storeroom	Parts consumption, stock value, stock-outs	Spare parts turnover, first-time fix rate
Condition monitoring and IIoT	Vibration, temperature, current, oil analysis	Early failure indicators, PM optimisation, MTBF

How to interpret maintenance KPIs

A single value tells you almost nothing. Interpretation rests on four habits.

Read the trend, not the point. Twelve periods of PM compliance say more than this month’s figure. A single strong month often reflects a light schedule rather than better execution.

Read indicators in pairs. Cost against reliability, MTBF against MTTR, PM compliance against unplanned failures. Every maintenance KPI can be improved in isolation by making something else worse – cost falls when work is deferred, MTTR falls when jobs are closed prematurely.

Check the definition before the number. Most disputes about a KPI turn out to be disputes about the denominator: whether standby hours count as required time, whether contractor hours are in the labour base, when the compliance window closes.

Segment by criticality. An average across the whole asset base hides the behaviour of the few assets that can stop production. Report critical assets separately.

How to improve maintenance KPIs

Improvement follows a predictable order, because the leading indicators drive the lagging ones.

Start with work-order discipline. If failures, labour hours and downtime are not captured consistently, every KPI downstream is guesswork. This is unglamorous and it is where most programmes actually stall.

Then fix planning and kitting. Job packs, kitted parts, permits raised in advance and realistic time estimates move wrench time, first-time fix and schedule compliance together.

Then optimise the PM programme. Review PM tasks against actual failure modes: remove routines that catch nothing, add condition-based tasks where they replace calendar-based intrusion. Rising PM compliance on the wrong tasks improves no outcome.

Then attack the dominant failure modes. Use the failure history to select two or three recurring modes on critical assets and run root-cause elimination rather than repeat repairs.

Only then chase the cost numbers. Cost per unit and cost as a percentage of RAV improve as a consequence of reliability and planning work, not as a target set in isolation.

Common mistakes when working with maintenance KPIs

The failure patterns below account for most abandoned KPI programmes.

Mistake	Why it fails	What to do instead
Tracking too many indicators	Thirty metrics mean none has a real owner or review cadence	Keep 10–15 at plant level, 4–6 for management
No owner or decision attached	The KPI is reported, discussed and never acted on	Name an owner and the action triggered by a deviation
Unreliable CMMS data	Missing labour hours and failure codes make every downstream KPI unusable	Fix work-order discipline before adding dashboards
Reporting cost KPIs alone	Cost falls when work is deferred, which looks like success for two quarters	Always pair cost with reliability and compliance
Benchmarking without checking definitions	Two plants report availability on different denominators	Document the formula and denominator for every KPI
Targets set to be hit, not to inform	Teams optimise the number rather than the outcome – jobs closed early, PMs signed off unperformed	Audit a sample of closed work orders periodically
No segmentation by criticality	Averages hide the assets that can stop production	Report critical assets as a separate group

How to start a maintenance KPI implementation

A first implementation should take weeks, not quarters. The sequence below works in most plants.

Step 1 – Define the business question. Reliability, cost or delivery of planned work: pick the one that matters this year.

Step 2 – Select 5–8 KPIs, no more. One or two per category, weighted towards the question in step 1.

Step 3 – Write the definition sheet. Formula, denominator, data source, owner, reporting frequency and the decision triggered by a deviation – one page per KPI.

Step 4 – Audit the data. Check that failures, labour hours and downtime are actually being captured in the CMMS. Fix the gaps before building anything.

Step 5 – Establish the baseline. Three to six months of history, so the first target is grounded rather than aspirational.

Step 6 – Set targets with the owners. A target agreed with the accountable planner or supervisor survives contact with the shop floor.

Step 7 – Build the dashboard in layers. Management view first, plant view second, analysis drill-down third.

Step 8 – Run a weekly review. Fifteen minutes on deviations and the actions arising, not a monthly presentation.

Step 9 – Validate by sampling. Audit closed work orders quarterly to confirm the numbers reflect work actually performed.

Step 10 – Review the KPI set annually. Retire indicators that no longer drive a decision and add the ones the next improvement phase needs.

FAQ

What are maintenance KPIs?

Maintenance KPIs are measurable indicators that show how effectively an organisation maintains its assets, controls maintenance cost and delivers planned work. Each one has a formula, a target, a data source and an owner.

Which 5 maintenance KPIs matter most?

The five most widely used are PM compliance, MTBF, MTTR, availability and the share of planned work. Together they cover reliability, repair performance and process discipline.

What are maintenance metrics?

A maintenance metric is any measurable value describing maintenance activity. It becomes a KPI once a target, an owner and a decision are attached to it.

Which KPIs should a maintenance manager track?

At plant level, 10–15 indicators across reliability, efficiency, cost, compliance and planning; four to six of those reported upwards to management.

How do you calculate MTBF?

MTBF = total operating time / number of failures, calculated per asset or asset group over a defined period.

How do you calculate MTTR?

MTTR = total repair time / number of repairs, including diagnosis, repair and functional testing.

How do you calculate PM compliance?

PM compliance = completed PM work orders / scheduled PM work orders × 100%, measured within a defined compliance window.

What is wrench time?

Wrench time is the share of paid maintenance labour time spent working directly on the asset, as opposed to travelling, waiting or searching for parts and information.

How do you calculate maintenance cost as a percentage of RAV?

Maintenance cost % RAV = annual maintenance cost / replacement asset value × 100%.

What is a good level of PM compliance?

Above 90% is the accepted benchmark; mature organisations sustain above 95%.

What is a good maintenance cost as a percentage of RAV?

Typically 2–5% of replacement asset value. Below 2% often signals deferred work rather than efficiency.

How much planned work should a maintenance team have?

Above 80% of the total workload, with mature organisations exceeding 90%.

What is a good wrench time?

30–40% is typical. Above 55% indicates a very mature planning and kitting process.

How do you set up a maintenance KPI dashboard?

Build it in three layers: four to six KPIs with targets for management, the full 10–15 for the plant, and a drill-down to work-order and failure data for analysis.

What data do maintenance KPIs require?

Work orders, PM schedules, labour hours and failure records from the CMMS; runtime, downtime and output from production systems; cost data from ERP; stock data from the storeroom; condition data from monitoring systems.

What is the difference between leading and lagging KPIs?

Leading KPIs describe the planned work that shapes future outcomes, such as PM compliance. Lagging KPIs describe results that have already occurred, such as MTBF or availability.

How do you improve maintenance KPIs?

In order: work-order discipline, planning and kitting, PM programme optimisation, elimination of dominant failure modes, and only then cost reduction.

Which 5 skills does maintenance work require?

Technical diagnosis, planning and scheduling, reliability analysis, data literacy in the CMMS, and communication with production.

Which 5 KPIs are worth tracking first?

PM compliance, share of planned work, MTTR, availability and maintenance cost as a percentage of RAV – one leading indicator, one maturity indicator, one efficiency indicator, one reliability indicator and one cost indicator.

What are the 7 elements of maintenance management?

Asset register and criticality, maintenance strategy, planning and scheduling, work execution, spare-parts management, data and KPIs, and continuous improvement.

Sources

1. SMRP (Society for Maintenance & Reliability Professionals) – Best Practices Metrics: definitions and benchmarks for maintenance indicators.
2. EN 15341:2019+A1:2022 – Maintenance – Maintenance Key Performance Indicators.
3. ISO 55000:2024 – Asset management – Vocabulary, overview and principles (2nd edition, superseding ISO 55000:2014).
4. Noria Corporation / ReliablePlant – educational material on reliability, maintenance and KPIs.
5. Smart RDM internal material on industrial data, asset monitoring and maintenance indicators.