Every ISO audit and every serious customer visit gets to the same question: what is your MTTR and MTBF? Most plants produce a spreadsheet computed the night before. The auditor knows. You know. Everyone moves on — and the numbers change nothing.
The definitions, briefly
- MTTR (Mean Time To Repair) = total repair time ÷ number of repairs. How fast you recover.
- MTBF (Mean Time Between Failures) = total running time ÷ number of failures. How reliable the machine is.
- Availability = MTBF ÷ (MTBF + MTTR). The one number that combines both.
Why hand-maintained versions are fiction
The numbers are only as honest as the timestamps, and recalled timestamps are systematically wrong in one direction: they flatter. The breakdown "started" when someone finally wrote it down. The repair "took two hours" because the gap waiting for approval and waiting for the spare quietly disappeared into it.
The fix is not a better spreadsheet. It is capturing time at the moment of the event: the operator reports the fault when it happens, the approver's click stamps the response, the technician's start and finish stamp the repair. Nobody fills a register — the register fills itself.
Reading the numbers like a professional
- Falling MTBF on one machine = it is asking for root-cause review, not another patch-up. Three bearing failures in 60 days is a lubrication system problem, not bad luck.
- High MTTR with fast repairs = your delay is logistics: approvals, technician availability, spares. Fix the process before blaming the fitter.
- Watch trends, not absolutes. An MTBF of 180 hours means little; an MTBF that fell from 300 to 180 in a quarter means everything.
In AssetAI these numbers exist as a by-product: the QR report, the approval tap and the technician's start/complete generate the timestamps, and MTTR/MTBF per machine are simply displayed — honest because no human ever typed them.