MTTR and MTBF Explained with Worked Examples

Published on 10/18/2026 • Updated on 10/23/2026

Ask a maintenance team how a machine is doing and you often hear it is fine, or it keeps giving trouble. Two simple numbers turn that feeling into facts: how often it fails and how long it takes to fix. Those are MTBF and MTTR.

This guide defines both in plain words, gives the formulas, works through examples with real arithmetic, links them to availability and shows how to use the numbers without drowning in data. The examples use assumed figures to show the method.

MTBF: mean time between failures

MTBF is the average running time between one failure and the next. It measures reliability. A higher MTBF means the machine fails less often.

Use operating time, not calendar time. Hours when the machine was switched off or waiting for orders do not count.

MTTR: mean time to repair

MTTR is the average time taken to restore a machine after a failure. It measures how quickly the team recovers. Definitions vary. Some plants count only the repair work, others count from the moment of the stop. Pick one definition and use it every time.

The M is sometimes read as meantime. The name matters less than a stable definition. Our guide on the breakdown register shows how to log the times.

Availability from both numbers

Availability is the share of time the machine is ready to run. With MTBF and MTTR it is simple.

Now see what each lever does. If MTTR falls from 2 hours to 1 hour, availability becomes 150 / 151 = 99.3 percent. If MTBF rises from 150 to 300 hours with MTTR still at 2, availability becomes 300 / 302 = 99.3 percent. Both routes reach the same place. Choose the cheaper.

Reading the pair together

The table is a quick triage. It shows where to look first.

Where the data comes from

You need three things for every machine: the operating hours, the number of failures and the time to restore each. Operating hours come from a running hour meter, a production log or a shift record. Failures and times come from the breakdown register. Without a register, the numbers are guesses.

• Define what counts as a failure, for example any unplanned stop over five minutes.
• Define when the clock starts and stops for repair time.
• Use the same definitions for every machine and every month.

A second worked example

C-3 fails more often but recovers fast, so its availability is higher. C-4 fails rarely but each failure is long. The fix for C-3 is finding the repeat cause. The fix for C-4 is spares, skills or a standby.

Using the numbers well

Do not set a target without a baseline. First measure for three months, then ask what is realistic.

Cautions

• Small samples mislead. A machine with two failures has a noisy MTBF.
• Averages hide extremes. One 20 hour outage can sit inside a good MTTR. Look at the longest repair too.
• Planned stops are not failures. Keep them out of MTBF.
• Compare like with like. A press and a pump have different normal values.

These numbers guide questions. They do not answer them on their own.

Related measures you will meet

You do not need all of them. MTBF and MTTR are the best pair to start with because the data is easy to collect and the meaning is clear.

Setting targets sensibly

After a baseline, set a modest target for the machines that matter. A reasonable aim is a small improvement per quarter in MTBF for the worst machine, or a cut in MTTR through better spares. Agree the target with the technicians who will do the work, and review progress with them. A target set in an office with no link to a plan is ignored.

Reward the behaviour you want. If the team is praised for a fast repair but not for preventing the failure, MTTR improves and MTBF stays low. Recognise both, in the same meeting.

The link to spares and skills

A high MTTR often has little to do with the repair itself. It is the waiting: for a part from the store, for a technician with the right skill, for a permit or for a vendor. Split repair time into waiting and working in the register for a month. The split often shows that half the time is spent waiting. That is the part you can cut fastest, by holding critical spares at the machine and training a second person on each critical task.

Common mistakes

Another mistake is chasing the average and ignoring the single costly failure. Review the top three longest outages each month.

A third is calculating by hand each month from scattered notes. Put the data in one place so the maths takes minutes.

A 90-day baseline

Build a trustworthy set of numbers.

Set targets only after the baseline, and share the first report with the maintenance team before it goes to management. People who see their own numbers first tend to explain them honestly and then act on them, which is the whole purpose of measuring.

Five-minute self-check

Check your numbers against these points.

Two or more no answers mean this is worth fixing before anything else on your list.

Reader Q&A

Bottom line

MTBF tells you how often a machine fails and MTTR tells you how long you take to recover. Together they give availability, and they show whether to fix reliability or recovery.

Log every stop, fix your definitions and review the numbers monthly with the people who repair the machines.

Author of the Blog Post

Aadeshwar Modi

Technical Content Writer at VisitorFlow

Aadeshwar Modi is a marketing technical content writer who loves writing on technology and AI topics.