MTTR (Mean Time To Repair) and MTBF (Mean Time Between Failures) are the two core maintenance metrics. MTBF measures how reliably a machine runs — the average time between breakdowns. MTTR measures how effectively your team responds — the average time from breakdown to machine back in production. Together they give a complete picture of maintenance performance.
Most resources covering MTTR and MTBF are written for IT and DevOps teams — servers going down, incidents getting resolved, software deployments rolling back. This guide is for manufacturing: hydraulic presses, bearing failures, spare-parts shortages, and overnight repair waits. The numbers work the same way; the context is entirely different.
What MTTR measures
MTTR is the average time from the moment a machine stops to the moment it is back in production. It is a measure of your team's recovery speed — not just the physical repair, but every step between breakdown and machine-back-up.
Formula: MTTR = Total repair time ÷ Number of breakdowns (in a defined period, for a specific asset)
What "total repair time" actually captures in a factory setting:
- Detection time: the lag between the failure occurring and someone noticing (sometimes zero on a running line; sometimes significant on unmanned auxiliary equipment)
- Response time: how long it takes a technician to arrive at the machine
- Diagnosis time: identifying the root cause — is it the seal, the pump, the hydraulic line?
- Parts wait time: the time spent waiting for the correct spare if it is not in stock
- Repair time: the actual hands-on fix
- Testing and restart: confirming the machine is running correctly before the production line restarts
Each of these sub-components is a separate lever. A plant with a high MTTR is not necessarily slow at the physical repair — it might have a parts availability problem, or technicians covering too large a floor area, or a shift-handover gap where breakdowns notified at end-of-shift don't get actioned until the next shift starts.
For the full treatment — including how to break MTTR down by component and how to benchmark it against your own trend — see the MTTR in maintenance guide.
What MTBF measures
MTBF is the average time a machine runs between one breakdown and the next. It is a measure of equipment reliability — how often the machine fails, not how long it takes to fix it.
Formula: MTBF = Total operating time ÷ Number of failures (same period, same asset)
"Total operating time" means the time the machine was actually running, excluding downtime. It is not planned production time, and it is not calendar time — it is the hours the machine was turning parts.
What MTBF captures:
- Equipment reliability: a machine with an MTBF of 500 hours is more reliable than one with an MTBF of 80 hours
- Preventive program effectiveness: rising MTBF over time means the PM program is working — the machine is failing less often. Falling MTBF is a signal that something is deteriorating: a PM being missed, a component reaching end of life, a change in operating conditions
- Asset health trend: a single MTBF number is informative; the trend over six to twelve months is diagnostic
What MTBF does not capture: how fast you recover when failures do occur. A machine can have excellent MTBF — it rarely fails — and still cause significant production loss if the repair takes twelve hours every time it does fail. That is where MTTR does the work.
For a full walkthrough of the MTBF formula, including how to handle planned downtime in the calculation, see the MTBF for machines guide.
The key difference: MTBF is about reliability; MTTR is about responsiveness
These two metrics measure fundamentally different things, and they respond to different interventions.
MTBF measures your preventive program. If your PM tasks are being executed on time, the machine is being lubricated correctly, worn components are being replaced before they fail, and operating conditions are within spec — MTBF should rise over time. A falling MTBF means something in your prevention program is not working: the PM is slipping, a component has reached end of life, or a new failure mode has appeared that your PM schedule does not address.
MTTR measures your reactive capability. Even a well-run PM program does not eliminate breakdowns — it reduces them. When failures happen, MTTR tells you how fast your team can get the machine back. MTTR falls when technicians have better spares availability, better fault history to draw on, faster response protocols, and clearer handover procedures.
This distinction matters for prioritization:
-
A plant with high MTBF but poor MTTR has a functioning prevention program but a slow repair process. Failures are infrequent, but when they happen, the recovery is painful. Worth investigating: parts availability, technician skill on specific failure modes, overnight and weekend coverage.
-
A plant with poor MTBF but excellent MTTR is firefighting efficiently. The team is fast at putting out fires, but the fires are happening too often. The PM program is where to invest — not in making the firefighting faster.
-
The goal is both: a preventive program that reduces failure frequency (MTBF up) and a repair process that minimizes downtime when failures happen anyway (MTTR down).
Neither metric alone tells the full story. A single MTBF number tells you how reliable the machine is — not whether that reliability is improving, and not whether the recovery process is costing you when it does fail. MTTR fills exactly that gap.
Worked example — same machine, same month, both metrics
One hydraulic press, March data. Three breakdowns during the month.
| Breakdown | Machine stopped | Machine back up | Repair time |
|---|---|---|---|
| 1 | 3 Mar 07:15 | 3 Mar 09:00 | 1 hr 45 min |
| 2 | 11 Mar 14:30 | 11 Mar 17:00 | 2 hr 30 min |
| 3 | 24 Mar 22:00 | 25 Mar 08:30 | 10 hr 30 min (overnight — parts wait) |
| Total repair time | 14 hr 45 min (14.75 hrs) |
Planned operating hours in March: 480 hours. Less downtime: 480 − 14.75 = 465.25 operating hours.
Calculating MTTR:
MTTR = 14.75 ÷ 3 = 4 hours 55 minutes
Calculating MTBF:
MTBF = 465.25 ÷ 3 = 155.1 hours
At 24-hour operation, 155 hours is roughly 6.5 days between failures. At 8-hour shifts, it is approximately 19 shifts between failures — about once every three to four working weeks.
What this tells us:
Three breakdowns in a single month on one machine is a high failure rate. An MTBF of 155 hours means this press is failing roughly every three weeks. That is primarily a PM problem — the prevention program is either not in place, not being executed consistently, or not addressing the right failure modes for this machine.
MTTR of 4 hours 55 minutes is tolerable for in-hours failures. Breakdown 1 (1h 45min) and breakdown 2 (2h 30min) are reasonable numbers — diagnosis was likely straightforward and parts were available. Breakdown 3 (10h 30min) is the outlier, and it is almost entirely driven by parts wait time, not repair complexity. That is a spares management problem on a specific component — not a technician skill problem.
The priority diagnosis from these two numbers:
The MTBF is the more urgent problem. Three breakdowns a month on one asset, regardless of how fast each is repaired, is not a sustainable position. The first action is to look at the PM schedule for this press: is the service interval appropriate? Are the right components being checked? Are the PMs being completed on time?
The MTTR for in-hours breakdowns is acceptable. The overnight parts wait on breakdown 3 is worth a separate investigation — which component failed, whether it should be a stocked spare, and whether the overnight coverage process needs adjustment.
This is exactly what the two-metric combination surfaces: MTBF points to the prevention gap; MTTR points to the spares gap on a specific failure mode. Either metric alone would give an incomplete diagnosis.
MTTR and MTBF as a diagnostic pair
Tracking both metrics together per asset produces a simple 2×2 that tells you what kind of problem each machine has — and what kind of intervention it needs.
| Low MTTR (fast recovery) | High MTTR (slow recovery) | |
|---|---|---|
| High MTBF (rare failures) | Well-managed maintenance. Failures are infrequent and resolved quickly. This is the goal. | Reliable equipment, slow recovery when it does fail. May be tolerable depending on failure frequency and production impact; worth investigating the recovery bottleneck. |
| Low MTBF (frequent failures) | Firefighting efficiently. Fast at fixing, but still failing too often. PM program is the priority. | Worst case — frequent failures and slow recovery. Highest production impact. Both the prevention program and the recovery process need attention simultaneously. |
The 2×2 is most useful as an asset-ranking tool. For a plant with 50 machines, most assets will sit in the top-left quadrant (acceptable) or close to it. The handful that fall into the bottom-right quadrant — frequent failures, slow recovery — are where the loss is concentrated. Start there.
One practical use: review this grid monthly for your top 10 assets by downtime hours. The machines that migrate from bottom-right toward top-left over six months tell you the PM program is working. The machines that stay in the bottom half despite intervention tell you the intervention is wrong.
MTTD and MTTF: two related terms worth knowing
Two other acronyms appear in MTTR and MTBF discussions. Neither is as commonly tracked on factory floors, but knowing the distinction prevents confusion.
MTTD — Mean Time To Detect
MTTD is the average time between a fault occurring and someone noticing it. On a manned production line, detection is usually immediate — the machine stops and the operator sees it. On auxiliary equipment or unmanned areas, detection can add significant time: a compressor that starts running hot at 2am might not be noticed until the first shift starts at 6am, adding four hours to what looks like "repair time" in the MTTR calculation.
MTTD is not commonly tracked as a separate metric in most plants. It is, however, embedded in your MTTR as "detection time." If your MTTR is high and your repair time itself is fast, MTTD is worth pulling out separately — it often points to a monitoring gap on specific equipment or shifts.
MTTF — Mean Time To Failure
MTTF (Mean Time To Failure) is used for non-repairable components, not for the machine as a whole. When a bearing is replaced, the old bearing is not repaired — it is discarded. The MTTF for that bearing type is its expected lifespan before replacement is needed again.
MTBF, by contrast, applies to the whole repairable machine: how long it runs before the next breakdown (whatever that breakdown turns out to be). A single machine will have one MTBF; the same machine will have multiple MTTFs — one for each category of component that gets replaced rather than repaired.
The practical implication: MTTF data on critical components (bearings, seals, belts, filters) feeds your PM interval planning. If the MTTF for a specific bearing on a specific press is 2,200 hours, your PM inspection interval for that bearing should be shorter — not longer — than 2,200 hours.
How to start tracking both
You do not need a separate calculation process. You need three data points captured consistently on every breakdown work order:
- Breakdown start time: the moment the machine stopped (or the work order was raised)
- Machine back in production: the moment the repair was complete and production restarted
- Planned production time for the period: how many hours the machine was scheduled to run
From those three data points, both MTTR and MTBF compute directly — total repair time, total operating time, and number of breakdowns are all derivable.
Where to start:
Pick 3–5 machines that have caused the most downtime in the past three months. If you don't have that data, start with your highest-impact assets — the ones where a breakdown stops the most output. Track those machines for 90 days. At the end of 90 days you have a baseline MTTR and MTBF for each. More importantly, you have 90 days of trend — whether each number is moving in the right direction.
The 90-day mark is when the metrics become actionable. A single month's MTBF is a data point; three months is a trend you can build a PM decision on.
The biggest mistake: collecting the data but not reviewing it. MTTR and MTBF are not reporting metrics — they are diagnostic inputs. Schedule a monthly 30-minute review of both numbers per machine with the maintenance supervisor. If MTBF is falling on a specific asset, ask what changed — did a PM slip? Did a component reach end of life? Did operating conditions change? The review is where the value is; the calculation is just the starting point.
For how both metrics sit inside a wider maintenance KPI framework, see the maintenance KPIs guide.
The conventional advice is to fix MTBF first and treat MTTR as a later refinement. I don't fully agree. MTTR is where I put my attention: even when MTBF is poor, a low MTTR means production resumes quickly — and the two improvements land on different people. Pushing MTBF up mostly adds load to the maintenance team, while pulling MTTR down depends on the whole plant: stores, shift handover, whoever can authorize an escalation at 2am. I don't think there is a clean threshold where the priority flips. It comes down to the equipment — if a machine is cheap enough to run to failure, or fails often enough that aggressive PMs are not economic, put the effort into recovering fast.
A cone making machine I saw is the clearest example I have of why one number is not enough. Its MTTR was excellent — a broken spare could be swapped in about five minutes — so on repair speed alone there was nothing to fix. The MTBF was what exposed the real problem: it was breaking that often. The fix was a design change to the machine, and it deliberately made MTTR worse, because the replacement became harder to get at. MTBF improved by a wide margin, and that was the right trade.
Where MachDatum fits: both MTTR and MTBF per asset update automatically with every closed work order — no spreadsheet, no end-of-month calculation. The breakdown start time is captured when the work order is raised; machine back-in-production is captured at close. The analytics view shows both metrics per machine, so the 2×2 picture above becomes visible without anyone calculating anything. We're onboarding our first group of manufacturing teams right now — see how it works at machdatum.com/cmms.
