How to calculate MTBF and MTTR (with a step-by-step example)

MTBF and MTTR are two simple divisions. The hard part is choosing what goes into the numerator and the denominator, and that is where most maintenance metrics stop meaning anything.

This guide covers the formulas, an example worked out by hand, the availability you get from the two numbers and the mistakes that make the metric tell a different story from what actually happens on the shop floor.

What MTBF and MTTR measure

MTBF (Mean Time Between Failures) answers one question: on average, how long does the equipment run before it breaks down? It is a reliability metric. Higher is better.

MTTR (Mean Time To Repair) answers another: when it breaks, how long does it take to get it running again? It is a measure of maintainability and of how quickly the team responds. Lower is better.

Both look at the same event, the failure, from opposite sides. An asset can fail rarely and take days to come back, or fail every week and come back in minutes. You only see the full picture when you look at the two together.

The formulas

Three definitions hold these calculations together:

• Operating time is the time the equipment was scheduled to run and did run. It is not calendar time.
• A failure is a loss of function that requires corrective work. A scheduled stop for preventive maintenance is not a failure.
• Repair time is the time the asset was down because of the failure, from the moment it stopped until it was released back to operation. It includes waiting, not just the hours with tools in hand.

Before you calculate anything, write these three definitions down and agree on them with the team. If each shift understands "failure" differently, the metric measures the interpretation, not the equipment.

Step-by-step example: Pump B-07

Take a fictional booster pump, Pump B-07, that runs two shifts: 16 hours a day, 26 days a month. In the month under review it had one scheduled 4-hour preventive and three stops due to failure.

Item	Value	Note
Scheduled time	416 h	16 h × 26 days
Scheduled preventive	4 h	Comes out of scheduled time, not counted as a failure
Failure 1: seal leak	2.5 h down	Corrective
Failure 2: seized bearing	4.0 h down	Corrective, included waiting for the part
Failure 3: tripped breaker	1.5 h down	Corrective
Total repair time	8 h	2.5 + 4.0 + 1.5
Operating time	404 h	416 − 4 − 8
MTBF	134.7 h	404 ÷ 3
MTTR	2.7 h	8 ÷ 3 (about 2 h 40 min)
Availability	98.1%	134.7 ÷ (134.7 + 2.7)

In practical terms: B-07 runs, on average, a little over eight production days between failures and is back up in under three hours. The 98.1% availability is the share of expected operating time during which it was actually available, counting failures only.

Notice that failure 2 alone accounts for half of the repair time. It is the first candidate for a cause analysis and for a review of bearing stock. To see where those hours went, it helps to break the repair down into the stages described in the 5 components of MTTR.

The most common mistakes

1. Using calendar time

If you divide the month's 720 hours by three failures, B-07's MTBF becomes 240 hours. The number looks good, but it includes nights and Sundays when the pump was switched off. It doesn't compare to anything and doesn't help you plan. Use the time the asset was supposed to be running.

The exception is continuous-duty assets, such as a cold room or a hospital elevator, which really do run 24 hours a day. For them, calendar time and operating time are almost the same.

2. Counting preventive work as a failure

If the 4-hour preventive stop goes in as a fourth "failure," MTBF drops to 101 hours (404 ÷ 4) and MTTR goes up. Worse, the more preventive work you do, the worse the metric gets, which punishes exactly the behavior you want to encourage. Separate work orders by type and calculate MTBF and MTTR from corrective work only.

3. An MTTR that ignores waiting time

Many teams record only the hours the technician logged. In B-07's failure 2, the technician worked 1.5 hours; the other 2.5 were spent waiting for a bearing. An MTTR of 1.5 hours hides the very problem you need to fix. The MTTR that matters to operations is equipment downtime.

4. An average that hides the extremes

Picture four repairs of 1 h, 1 h, 1.5 h and 12 h. MTTR is 3.9 hours, yet no repair took anywhere near that: three were quick and one was a disaster. Always look at the list of work orders behind the average, and consider tracking the median as well (here, 1.25 h) along with the longest repair in the period.

5. Too few failures to conclude anything

With one or two failures in the period, MTBF is close to a coin toss. An asset that failed once in January and not at all in February did not "improve infinitely." For assets that rarely fail, use longer windows (quarter, half year, year) or group similar equipment into families.

Using both metrics to make decisions

The numbers are there to help you choose where to act. A simple way to read the combination:

• Low MTBF, low MTTR: it fails often but comes back quickly. Investigate the cause of the repeat failures; the team is fighting fires efficiently. The 5 Whys technique is a good place to start.
• High MTBF, high MTTR: it rarely fails, but when it does, recovery is slow. Review critical spare parts, repair procedures and access to specialists.
• Low MTBF, high MTTR: this is the worst quadrant and where your maintenance plan should start.
• High MTBF, low MTTR: keep an eye on it, without spending extra effort for now.

Compare the same asset over time before comparing different assets. A booster pump and a hydraulic press don't need the same MTBF to be well maintained.

Where the data comes from

In practice, MTBF and MTTR are only as good as your work order records. If a failure was solved over the phone and nobody opened a work order, it doesn't exist for the metric. If the work order was closed at the end of the day with an approximate time, MTTR is a memory. If that sounds familiar, take a look at the signs of reactive maintenance.

A CMMS like taskfield records the time of each work order step in the technician's app, even with no signal, separates corrective from preventive work by order type and calculates MTBF and MTTR from closed work orders, with no end-of-month spreadsheet. In manufacturing, this is often the first step toward comparing assets with numbers the team trusts.