MTBF and MTTR Maintenance KPIs: How to Read and Use Them

What do MTBF and MTTR mean?

The International Electrotechnical Vocabulary (IEC 60050-192) defines mean operating time between failures (MTBF) as the expectation of the duration of operating time between failures, and limits it to repairable items; for non-repairable items, mean operating time to failure (MTTF) is used. ^[2] Put simply, the higher the MTBF, the longer the asset runs between failures.

In the same vocabulary, MTTR stands for mean time to restoration; “mean time to repair” and “mean time to recovery” are listed as deprecated terms for it. ^[3] The Society for Maintenance & Reliability Professionals (SMRP) calls it “mean time to repair or replace” and places it, with MTBF, under the equipment reliability pillar of its metrics. ^[6]

The two indicators answer different questions: how often does the asset fail, and how long does it take to get it back into service?

How do you calculate them?

The basic formulas are:

• MTBF = total operating time in the period ÷ number of failures in the same period.
• MTTR = total time to restoration for those failures ÷ number of failures.

Hypothetical example: a pump at a station is required to run around the clock. Over 30 days (720 hours) it failed 4 times, with restoration times of 6, 10, 3 and 5 hours, giving 24 hours of downtime in total, and there were no planned stops.

Note that 174 ÷ (174 + 6) gives the same result: when both indicators are calculated over the same period, this simple calculation links them to the share of time the asset was running. Had there been planned stops, they would be deducted from operating time without being counted as failures.

Choose a measurement period that contains enough failures. An asset that fails only once a month is better read over a quarter or a rolling year than month by month. Keep the same period in every report so that comparisons stay fair.

What goes into MTTR: repair time or restoration time?

This is the most important question before comparing any two figures. In IEC terminology, time to restoration starts at the instant of failure and ends at restoration; if the instant of failure is not known, it starts when the failure is detected. ^[4] Repair time, by contrast, covers fault localisation, diagnosis, correction and function checkout, and excludes technical, administrative and logistic delays. ^[5]

The practical consequence: a team that measures MTTR from when the technician starts work until they finish will report a much smaller figure than a team that measures from when the asset stopped until it was handed back to operations, even if performance is identical. So write the agreed definition into your maintenance procedure, and if you use maintenance contractors, state in the contract exactly when the clock starts and stops.

Where do the indicators sit in EN 15341 and SMRP?

The European standard EN 15341:2019, as amended by A1 in 2022, lists key performance indicators for the maintenance function and gives guidance on selecting them to appraise and improve the effectiveness, efficiency and sustainability of maintenance, covering industrial assets, infrastructure, facilities, civil buildings and transport systems. ^[1] The standards it refers to include the maintenance terminology standard EN 13306 and the IEC 60050-192 vocabulary, ^[1] so it makes sense to align your internal definitions with both.

SMRP organises its metrics into five pillars: business and management, manufacturing process reliability, equipment reliability, organisation and leadership, and work management. MTBF, MTTR and mean time between maintenance (MTBM) sit under equipment reliability as metrics 3.5.1, 3.5.2 and 3.5.3. ^[6]

You do not have to pick a single framework, but your reports should state which definition you follow, especially when comparing with other sites or organisations.

How should you read the results?

Do not judge a single month’s figure. An MTBF calculated from only two failures is inherently volatile; the right reading is the trend across several periods. SMRP material notes that MTBF can be measured by individual asset or by equipment class and subclass, ^[6] which is better than a single overall figure, because an average across different assets hides the problem asset.

Remember too that MTBF is an average: it does not mean the asset will run that long before its next failure. It is an indicator for comparing and tracking reliability, not a prediction of when the next failure will occur.

When comparing two sites or contractors, check three things before drawing conclusions: the assets are similar in type and operating conditions, the definition of failure is the same, and restoration time starts and ends at the same points. If any of these is missing, the difference in the figures reflects how they were measured, not how well the assets performed.

Which common mistakes distort the indicators?

• Using calendar time instead of operating time: an asset that runs one shift a day will look more reliable than it is if measured over 24 hours.
• Counting planned stops as failures: this lowers MTBF and penalises the team for doing preventive maintenance.
• Closing work orders late: if a work order is closed days after the actual handover, MTTR is inflated for no real reason.
• No record of the actual failure time: logging only when the report was raised drops detection and reporting time from the calculation.
• Combining dissimilar assets: an average of a small pump and a large compressor supports no decision.

How do you turn the indicators into real improvement?

If MTBF is low: start with root cause analysis of repeat failures, review the preventive maintenance plan and operating conditions, and make sure repeat failures are classified by a clear failure mode in the maintenance management system.

If MTTR is high: break it down into its components using the timeline above: how much time was lost before the technician arrived, waiting for parts, or in diagnosis? SMRP material notes that trending MTTR can indicate whether unnecessary delays are actually being eliminated. ^[6] The remedy depends on where the delay sits: critical spares held in stock, written repair procedures, diagnostic training, or a better reporting route.

Data rules before dashboards: define the mandatory work order fields (failure time, start time, repair completion time, handover time and failure mode), and review recording quality monthly for critical assets before presenting any figure to management.

A monthly review template for critical assets

Make the two indicators part of a short monthly meeting between maintenance and operations, shown in a single table for each critical asset:

The key column is “largest component of restoration time”, because it turns the figure into a specific question: why was time lost here? The last column makes sure every meeting ends with an action that has an owner and a date, reviewed at the next meeting.

Conclusion

MTBF tells you how often an asset fails; MTTR tells you how much time each failure costs. Fix the definitions first, calculate on actual operating time, read the trend for each asset, then break restoration time down to see where improvement should go.