What are MTTF and MTTR?
Mean Time to Failure (MTTF) is the average time a piece of equipment operates before it fails, calculated as total operational hours divided by the number of failures. It applies to non-repairable assets. Mean Time to Repair (MTTR) is the average time required to diagnose, repair and restore equipment after a failure, calculated as total repair time divided by the number of repairs. MTTF measures reliability; MTTR measures maintenance efficiency.
What does the R in MTTR stand for?
This is the single most common source of confusion with the metric, and it is worth settling before you benchmark anything. Depending on the source, the R means:
- Repair — hands-on wrench time only, from starting work to finishing it.
- Recovery — everything until the asset is back in service, including detection, waiting, travel and testing.
- Respond — the time from alert to someone beginning work.
- Resolve — the full lifecycle of the incident, often used in IT and service-desk contexts.
The same organization can report wildly different “MTTR” figures depending on which clock it starts. Mean time to recovery is almost always the larger number, because it includes the waiting and travel that mean time to repair excludes — and it is usually the one customers actually experience.
How to calculate MTTF and MTTR
MTTF = Total operational hours ÷ Number of failures
If a population of non-repairable components accumulates 50,000 operating hours and 20 of them fail, MTTF = 50,000 ÷ 20 = 2,500 hours.
MTTR = Total repair time ÷ Number of repairs
If a maintenance team spends 60 hours on repairs across 24 separate repair events, MTTR = 60 ÷ 24 = 2.5 hours.
How MTTF, MTBF and MTTR fit together
MTTF applies to assets that are replaced rather than fixed. For repairable assets, the equivalent reliability measure is MTBF (Mean Time Between Failures). Reliability and repair speed combine into availability:
Availability = MTBF ÷ (MTBF + MTTR)
Raising MTBF (more reliable equipment) and lowering MTTR (faster restoration) both increase uptime — but they are different projects with different owners. MTBF is largely an engineering and maintenance-strategy problem. MTTR is an execution problem.
What is a good MTTR?
There is no single correct target, because MTTR depends on asset criticality, on staffing, and above all on which clock you start. What is available is a directional benchmark for industrial recovery time.
| Measure | Figure | Source (year) |
|---|---|---|
| Average time to get production running again after unplanned downtime | 81 minutes (was 49 minutes five years earlier) | Siemens / Senseye, The True Cost of Downtime 2024 |
| Downtime incidents per facility per month | 25 (down from 42 in 2019) | Siemens / Senseye (2024) |
| Downtime hours lost per plant per month | 27 (down from 39 in 2019) | Siemens / Senseye (2024) |
The Siemens data is based on 181 completed online interviews with maintenance, engineering and IT professionals at large industrial organizations across automotive, FMCG, heavy industry and oil & gas, covering April 2019 to March 2023.
The most practical target is not an absolute number but a downward trend against your own baseline, measured with a definition you have not changed mid-year.
Why recovery times are getting longer
The Siemens finding is counter-intuitive and worth sitting with: manufacturers have successfully cut the number of downtime incidents — 25 a month per facility, down from 42 in 2019 — yet each incident now takes longer to recover from. Siemens attributes the increase partly to businesses losing skilled maintenance labour during the post-COVID “great resignation,” creating a skills and knowledge gap, and partly to supply-chain difficulty sourcing emergency replacements.
That distinction matters when choosing where to invest. Fewer failures is a reliability win. Slower recovery is an access-to-expertise problem — the right person is not in the room when the machine stops.
How to reduce MTTR
- Fix the definition first. Decide whether you are measuring repair, recovery or resolution, and keep it stable.
- Cut the time before work starts. In distributed operations, detection, dispatch and travel routinely dominate the hands-on repair time.
- Shorten diagnosis. Time spent establishing what is actually wrong is often the largest controllable block.
- Standardize the procedure. Consistent work instructions reduce variance between an experienced technician and a new one.
- Separate the parts clock. If parts lead time dominates, no amount of process improvement in the workshop will move the number — that is an inventory problem, and worth reporting separately.
How VSight helps
Two of the blocks above — travel time for an expert and diagnosis time — are what AR remote assistance directly compresses. With VSight Remote, a senior expert joins the on-site technician’s live camera view within minutes and annotates the equipment itself, instead of the asset waiting for a scheduling and travel window. That addresses precisely the gap Siemens identifies: the knowledge left with the people who retired or resigned, which now has to reach the machine some other way. VSight Workflow delivers the repair procedure as standardized digital work instructions so execution is consistent. VSight is a connected worker platform, and is GDPR, HIPAA and ISO 27001 certified.
Being precise about the limits: VSight does not measure or store MTTR — it is an input to whatever your maintenance system reports. It does not shorten parts lead time, and it does not predict failures, so it moves MTTR rather than MTTF or MTBF.
Request a demo to see how removing expert travel from the repair clock changes your recovery time.
Related terms: MTBF, uptime and downtime, first time fix rate, preventive maintenance, predictive maintenance, CMMS.
Frequently asked questions
What is the difference between MTTF and MTTR? MTTF (Mean Time to Failure) is the average time equipment operates before it fails, and it applies to non-repairable assets. MTTR (Mean Time to Repair) is the average time needed to diagnose, repair and restore equipment after a failure, and it measures maintenance efficiency rather than equipment reliability.
How is MTTF calculated? MTTF is total operational hours divided by the number of failures. For example, 50,000 operating hours across a population of components with 20 failures gives an MTTF of 2,500 hours. MTTR, by contrast, is total repair time divided by the number of repairs — 60 hours of repair across 24 repairs gives an MTTR of 2.5 hours.
What does the R in MTTR stand for? It depends on the source, which is why MTTR comparisons so often break down. R can mean repair, recovery, respond or resolve. Mean time to repair usually counts only hands-on repair time; mean time to recovery counts everything until the asset is back in service, including waiting and travel. Agree which one you mean before benchmarking against anyone.
What is the difference between MTTR and MTBF? MTBF (Mean Time Between Failures) applies to repairable assets and measures average uptime between failures, while MTTR measures how long restoring the asset takes. They combine into availability, commonly expressed as Availability = MTBF divided by (MTBF + MTTR). Raising MTBF and lowering MTTR both increase uptime.
What is a good MTTR? There is no single good number, because MTTR depends on asset criticality and on which clock you start. As a directional benchmark, Siemens reports that manufacturers now take an average of 81 minutes to get production running again after unplanned downtime, up from 49 minutes five years earlier. A useful target is simply a downward trend against your own baseline, measured consistently.
How does VSight help reduce MTTR? VSight compresses two specific components of MTTR — expert travel time and diagnosis time. A remote expert joins the technician’s live camera view immediately and annotates what they see, instead of the asset waiting for a scheduling and travel window. It does not shorten parts lead time, and VSight does not measure or store MTTR itself; it is an input to the number your maintenance system reports.
Sources
- Siemens / Senseye Predictive Maintenance — The True Cost of Downtime 2024 (181 completed interviews, April 2019 – March 2023)