Maintenance indicators are metrics that measure the performance of assets and of the maintenance team across four dimensions: reliability, availability, cost and execution quality. Without them, management decides by perception; with them, it decides by fact. This guide brings together the 12 indicators that a mature maintenance operation tracks, each one with the formula written in plain text, the target range when a reliable reference exists, and the measurement catch: what the number hides when field reporting is poor.
Before the list, a principle that underpins everything that follows: an indicator is not born on the dashboard, it is born in the maintenance order. The work order is the mother cell of information, and the quality of any KPI is directly proportional to how seriously it is filled in out in the field. We will come back to this point after the list.
Master table: the 12 indicators and their formulas
| No. | Indicator | Formula in plain text | Unit | Better when |
|---|---|---|---|---|
| 1 | MTBF | Total operating time ÷ number of failures | Hours | Rises |
| 2 | MTTR | Total repair time ÷ number of failures | Hours | Falls |
| 3 | Physical availability | Hours in operating condition ÷ total hours in the period × 100 | % | Rises |
| 4 | OEE | Availability × performance × quality × 100 | % | Rises |
| 5 | Backlog | Labor hours of orders in the portfolio ÷ available labor hours per week | Weeks | Stable between 2 and 4 |
| 6 | Corrective index | Labor hours on corrective work ÷ total maintenance labor hours × 100 | % | Falls |
| 7 | Preventive plan compliance | Preventive tasks done on time ÷ scheduled preventive tasks × 100 | % | Rises |
| 8 | Schedule compliance | Orders executed as scheduled ÷ scheduled orders × 100 | % | Rises |
| 9 | Rework index | Labor hours on reopened orders ÷ total labor hours worked × 100 | % | Falls |
| 10 | Maintenance cost over revenue | Total maintenance cost ÷ gross revenue × 100 | % | Falls with stable availability |
| 11 | Cost over replacement value (CPMV) | Annual maintenance cost of the asset ÷ replacement value of the asset × 100 | % | Falls |
| 12 | Wrench time | Hours of effective execution ÷ total shift hours × 100 | % | Rises |
1. MTBF (mean time between failures)
MTBF (Mean Time Between Failures) is the thermometer of reliability: it measures how long, on average, a repairable piece of equipment operates between one failure and the next.
Formula: MTBF = total operating time ÷ number of failures in the period.
How to measure: use real operating time (subtracting stoppages), count only unplanned failures and stratify by equipment or family. A rising MTBF is the clearest sign that corrective interventions are decreasing.
Target range: there is no universal value; the number depends on the asset type, the operating regime and criticality. The correct reference is the equipment's own history and that of sister equipment in the plant.
The catch: failures resolved without a work order do not enter the count, and MTBF becomes artificially high. The indicator measures what was recorded, not what happened.
2. MTTR (mean time to repair)
MTTR (Mean Time To Repair) measures maintainability: how quickly the team returns the asset to operation after a failure.
Formula: MTTR = total repair time ÷ number of failures in the period.
How to measure: the clock runs from the moment the equipment stops until it returns to operation, including diagnosis and waits. That breadth is what makes MTTR useful: it exposes weak support processes (missing part, missing tool, missing procedure), not just the technician's skill.
Target range: the lower, the better, always compared with the asset's own history.
The catch: reporting done from memory at the end of the shift turns the timestamps into estimates and the indicator into fiction.
3. Physical availability
Physical availability is the final product that maintenance delivers to operations: the fraction of time the equipment is in condition to produce.
Formula: Physical availability = hours in operating condition ÷ total hours in the period × 100. In its intrinsic form, calculated from the previous indicators: MTBF ÷ (MTBF + MTTR) × 100.
How to measure: define with operations what counts as maintenance unavailability (a stoppage for lack of raw material, for example, is not maintenance's) and keep the criterion stable over time.
Target range: it is defined by contract with operations and varies by sector and by asset criticality. More important than chasing an absolute number is ensuring a consistent measurement criterion.
The catch: changing the criterion for what counts as a "maintenance stoppage" in the middle of the year creates an indicator improvement that no machine ever felt.
4. OEE (overall equipment effectiveness)
OEE (Overall Equipment Effectiveness), consolidated by the TPM methodology, measures total equipment effectiveness by combining the maintenance view with the production view.
Formula: OEE = availability × performance × quality × 100. In plain text: the fraction of time the equipment operated, multiplied by the fraction of nominal speed it reached, multiplied by the fraction of good products on the first pass.
How to measure: the three factors come from different sources (stoppages, production count, scrap), which makes OEE especially sensitive to the quality of each record.
Target range: it depends on the process and the regime; the value of the indicator is in revealing where the loss lives: breakdown, microstop, reduced speed or scrap.
The catch: equipment can be mechanically available and still destroy OEE with microstops and reduced speed. For maintenance, the controllable share is availability; handling it well is the biggest contribution the planning function makes to the index.
5. Backlog
Backlog measures the accumulation of pending work: the relationship between recorded demand and the team's capacity to meet it.
Formula: Backlog = sum of the labor hours of orders in the portfolio ÷ the team's available labor hours per week. Available labor hours are total labor hours multiplied by the productivity factor, because no technician dedicates 100% of the shift to work orders.
How to measure: include planned, pending, scheduled and executed-but-not-closed orders; stratify by specialty to find bottlenecks.
Target range: market references such as Engeteles and Fracttal place the healthy range between 2 and 4 weeks. Below that, there is not enough schedulable work; above it, the queue grows and prioritization becomes reactive.
The catch: an unpurged portfolio and late reporting inflate the numerator, and the hasty conclusion ("we lack people") attacks the wrong problem.
6. Corrective maintenance index
The corrective index shows how much of the team's effort is consumed reacting to failures, rather than preventing them.
Formula: Corrective index = labor hours applied to corrective orders ÷ total maintenance labor hours × 100.
How to measure: the classification of the order type has to be disciplined. Planned corrective work (failure identified, intervention scheduled) and emergency corrective work have very different impacts and deserve separate reading.
Target range: in PM Run's experience, when more than half of the team's labor hours go to corrective work, the planning function loses the ability to schedule: the agenda comes to be dictated by breakdowns, the backlog inflates and costs soar. The healthy movement is emergency corrective work giving up space, month by month, to planned work.
The catch: plants that classify everything as corrective (or everything as preventive) to simplify reporting make this indicator, and the next one, useless.
7. Preventive plan compliance
This indicator answers whether the maintenance plan comes off paper: how much of what was scheduled as preventive work actually happened on time.
Formula: Plan compliance = preventive tasks executed on time ÷ preventive tasks scheduled in the period × 100.
How to measure: define an explicit time tolerance (for example, executed within the scheduled week) and keep the criterion fixed.
Target range: as close as possible to 100%. Preventive work systematically postponed is the first link in the chain that ends in breakdown, emergency corrective work and MTBF falling.
The catch: meeting the plan does not prove the plan is good. A plan copied from the manual, without adherence to the reality of the asset, can be 100% complied with and 0% effective. This indicator is read in pair with MTBF: a plan fulfilled with MTBF falling is the wrong plan.
8. Schedule compliance
While the previous indicator looks at the preventive plan, this one measures how well execution adheres to the weekly schedule as a whole: was the promise the planning function made for the week delivered?
Formula: Schedule compliance = orders executed as scheduled ÷ orders scheduled for the week × 100.
How to measure: freeze the schedule at close (the Friday snapshot) and compare it with what was executed. Orders that came in afterwards are a break in the schedule, even if they were executed.
Target range: the higher, the better; the value of the indicator is in the trend and in analyzing the reasons for deviation (urgencies, lack of material, denied release).
The catch: underscheduling to "hit the target" creates a pretty indicator and an idle team. The reading pair here is backlog: high compliance with a growing backlog indicates an undersized schedule.
9. Rework index
Rework measures execution quality: how much of the team's effort is spent redoing what was already reported as done.
Formula: Rework index = labor hours applied to reopened orders ÷ total labor hours worked × 100.
How to measure: define an objective reopening criterion (same failure, same equipment, within a time window) and record the link between the new order and the original one.
Target range: the technical ideal is zero. Every percentage point of rework consumes labor hours that were already scarce, undermines asset reliability and erodes maintenance's credibility with operations.
The catch: without the link between orders, rework disguises itself as a new failure, the index looks good and the problem (palliative intervention, missing procedure, wrong part) remains invisible.
10. Maintenance cost over revenue
This is the indicator that translates maintenance into the language of the board: how much of the company's revenue is consumed to keep the assets operating.
Formula: Maintenance cost over revenue = total maintenance cost ÷ gross revenue × 100.
How to measure: total cost includes in-house labor, materials, third-party services and the structure of the area. Rigor in allocating costs by order is what later allows the number to be broken down by asset, by area and by maintenance type.
Target range: as a public historical reference, the editions of ABRAMAN's Documento Nacional recorded the maintenance cost of Brazilian companies fluctuating between roughly 3.6% and 4.4% of revenue in the 1990s measurements, a level the association kept tracking in the following editions of the survey. Your plant's number should be compared with its own history and with companies in the same sector.
The catch: cutting maintenance cost drops this indicator in the short term and drops availability in the medium term. The correct reading is always in pair: cost falling with stable availability is efficiency; cost falling with availability falling is technical debt being taken on.
11. Maintenance cost over replacement value (CPMV)
CPMV compares the annual cost of maintaining an asset with the cost of buying it new, and it is the indicator that grounds overhaul-or-replacement decisions.
Formula: CPMV = annual maintenance cost of the asset ÷ replacement value of the asset × 100.
How to measure: it requires cost allocation by equipment (not just by cost center) and an updated replacement value, not the depreciated book value.
Target range: the lower, the better. When the annual cost of keeping the asset becomes a relevant fraction of the cost of replacing it, and especially when that fraction grows year after year, maintenance engineering should open the feasibility study between overhaul (retrofit) and decommissioning.
The catch: this indicator only works for those who allocate cost to the right order of the right equipment. Generic allocation to "umbrella" orders hides exactly the assets that are devouring the budget.
12. Wrench time
Wrench time measures the fraction of the shift in which the technician is effectively performing maintenance work, tool in hand, in contrast to travel, waits and paperwork.
Formula: Wrench time = hours of effective execution ÷ total shift hours × 100.
How to measure: by structured activity sampling or by detailed execution records. The goal is not to watch the technician, it is to see how much of the contracted capacity is being consumed by waiting for a part, waiting for release and travel.
Target range: in plants without active management of this indicator, effective execution is usually a minority of the shift, and that is why it is maintenance's biggest hidden capacity mine: every point recovered equals new capacity without hiring.
The catch: attacking the indicator by pressuring the technician, instead of attacking the waits that surround them, produces doctored reporting and not a single extra hour of real execution. The topic has its own article: wrench time and the operational friction that consumes your team.
The catch that runs through all of them: reporting quality
Reread the list and notice the pattern: all 12 indicators depend on data born in field execution. Stop and restart timestamps (MTBF, MTTR, availability), estimated and actual labor hours (backlog, wrench time), order type (corrective, preventive), the link between orders (rework), cost allocated by equipment (cost over revenue, CPMV).
When the record is made hours or days after execution, from memory, all those fields become approximations, and the entire panel starts to display fictional indicators: numbers precise in appearance, wrong at the source. Management then makes correct decisions about a plant that does not exist. The mechanism of this effect, with the fields most affected, is in the article on late reporting and the distortion of maintenance KPIs.
That is why digitizing the front line is not a luxury, it is a prerequisite of measurement: reporting at the moment and place of execution, with technical closeout releasing the asset in the system in real time. If your team still reports on paper or at the end of the shift, PM Run Mobility solves exactly this origin of the problem, with reporting on the phone integrated with SAP PM, working even without a connection. Discover PM Run Mobility and ensure indicators that reflect the field.
Safety and people indicators
Two groups of indicators complement the technical and financial panel:
- Accident frequency rate (FR): number of accidents with lost time multiplied by 1,000,000, divided by the total labor hours of exposure to risk. It is the thermometer of the team's exposure.
- Accident severity rate (SR): number of days lost and charged multiplied by 1,000,000, divided by the total labor hours of exposure to risk. It measures the severity of what happened.
- Training hours: percentage of total labor hours dedicated to technical and behavioral development. A trained team shows up in the other indicators: less rework, lower MTTR, more safety.
These indices converge everyone's interests: no availability gain justifies FR and SR getting worse.
How to roll it out without becoming bureaucracy
Twelve indicators described does not mean twelve targets on the wall. Three practical rules for rollout:
- Start with a few, chained together: availability, MTBF, MTTR, backlog and cost over revenue form a core that already tells the complete story (does the asset break? does it come back fast? can the team keep up? at what cost?). The rest come in as reporting matures.
- Every indicator needs an owner and a meeting: a number without a ritual of critical analysis is decoration. The routine is the PDCA cycle applied: plan the target, execute, check the indicator and act on the deviation, treating the cause and not the symptom.
- Automate collection before sophisticating analysis: a pretty dashboard over bad data is an expensive illusion. First ensure reporting at the source, then evolve the analysis layer.
The indicator panel is one of the pillars of Maintenance Planning and Control; to see how it integrates with roles, routines and the order flow, consult the complete guide to maintenance planning and control. And to track workload, productivity and scheduling in real time on top of SAP PM, PM Run Planning delivers this view ready to use. Discover PM Run Planning and track your indicators in real time.
Frequently asked questions about maintenance KPIs
What are the 5 main maintenance indicators?
An effective core brings together physical availability, MTBF, MTTR, backlog and maintenance cost over revenue. Together, they answer whether the asset breaks frequently, whether it returns to operation fast, whether the team can handle the demand and how much that costs. The other indicators refine this reading.
What are maintenance KPIs?
Maintenance KPIs (Key Performance Indicators) are the key metrics that measure the performance of the maintenance function: asset reliability and availability, workload management, execution quality, costs and safety. They turn work order records into an objective basis for management decisions.
What is the difference between MTBF and MTTR?
MTBF is the mean time between failures: it measures reliability, that is, how long the equipment operates between one breakdown and another. MTTR is the mean time to repair: it measures how quickly the team returns the equipment to operation. The first should rise; the second should fall.
What is world-class maintenance?
It is the management standard in which maintenance operates with a predominance of planned work, indicators measured with reliable field data and structured continuous improvement. In practice, it means emergency corrective work in constant decline, a preventive plan fulfilled and cost decisions based on precise allocation by asset.
How many indicators does a maintenance team need?
Fewer than it seems. A core of 4 to 6 indicators well measured and analyzed as a routine is worth more than a panel with 20 numbers that nobody audits. Start with the core (availability, MTBF, MTTR, backlog, cost) and expand as reporting quality matures.
From data to decision
The 12 indicators in this guide form a system: the reliability ones say where the asset hurts, the workload ones say whether the team keeps up, the economic ones say how much it costs and the quality ones say whether the work is being done well. None of them is worth anything without the common raw material: faithful reporting of field execution.
The PM Run platform integrates Planning and Mobility connected to SAP PM in real time, and already supports more than 12,000 users in maintenance teams across Brazil. If the next step of your management is to measure for real, discover the PM Run platform and request a demonstration.
This content was developed by the PM Run team, specialists in maintenance management integrated with SAP.
