The two formulas
MTBF = operating time ÷ number of failures MTTR = total repair time ÷ number of repairs
MTBF (mean time between failures) measures reliability: how long the equipment keeps running. MTTR (mean time to repair) measures responsiveness: how long it takes you to get it back into service.
They are independent. One asset can rarely fail and take a week to repair. Another can fail often and be fixed in twenty minutes.
A complete example
Example
Compressor FRG-PLT1-CMP-002, last 12 months.
| Data | Value |
|---|---|
| Hours in the period | 8,760 h |
| Downtime hours from failures, from failure to restart | 62 h |
| Failures | 4 |
Operating time = 8,760 − 62 = 8,698 h MTBF = 8,698 ÷ 4 = 2,174.5 h (about 91 days) MTTR = 62 ÷ 4 = 15.5 h Availability = MTBF ÷ (MTBF + MTTR) = 99.3%
99.3% sounds great. But it means 62 hours with the compressor down. If that facility keeps food cold, the useful question is not the percentage: it is what those 62 hours cost.
Mistake 1: calculating it on an incomplete history
It is the most common mistake and the most serious one. Say only 2 of those 4 failures were recorded, because two were fixed on the spot without opening a work order. The MTBF comes out twice as good as it really is.
The KPI is only as good as the records behind it. Before you show an MTBF to anyone, ask whether every job is in the system. If the answer is “almost all of them,” the number is not good enough to base a decision on.
Mistake 2: not deciding when the MTTR clock starts
When the equipment fails? When someone reports it? When the technician arrives? An MTTR of 2 hours and another of 15 can describe the same service.
The useful approach is to measure the whole time, from failure to restart, and break it down:
| Stage | Hours | Share of total |
|---|---|---|
| Detection and reporting | 8 | 13% |
| Travel | 6 | 10% |
| Waiting for parts | 36 | 58% |
| Actual repair | 12 | 19% |
That breakdown changes the whole conversation. This asset’s problem is not the technicians: it is the parts room. Without the breakdown, the meeting would have been about productivity.
Mistake 3: using it to compare different things
MTBF only makes sense within the same family of equipment. Comparing a pump’s MTBF with an elevator’s tells you nothing.
Where it does help is comparing an asset with itself over time, or twin assets with each other. If two identical compressors have an MTBF of 2,175 and 900 hours, something about the second one deserves a visit: the installation, how it is used or its environment.
What to do with the numbers
- Rank your assets by MTBF within each family. The bottom ten are your to-do list for the quarter.
- Cross MTBF with criticality. An unreliable asset that is not critical is not a problem. An unreliable, critical one is urgent.
- Break down MTTR and go after the biggest stage, which is rarely the repair itself.
- Watch the trend, not the single value. A family whose MTBF drops three quarters in a row tells you more than any isolated figure.
An MTBF calculated on incomplete records is worse than none at all, because it gives false confidence. Before you build the KPI, make sure every job goes into the system. That is the real work; the formula is the easy part.
Last updated: