Breakdowns that repeat
The same failure on the same unit three times isn’t bad luck: it’s a cause nobody has fixed, or an inspection that isn’t checking what it should.
Guide
The usual question is how much preventive maintenance to do. The useful one is different: what is your corrective maintenance telling you about the preventive plan you’re running?
Preventive maintenance is done before something fails, following a plan based on elapsed time or accumulated use. Corrective maintenance is done once it has already failed. Every maintenance operation has both, and the ratio between them isn’t a preference: it’s an outcome. It’s a consequence of what assets you have, what preventive plan you’re actually running, and how much it costs when something stops. That’s why there’s no universal "correct" ratio — and be wary of anyone who gives you one.
The opposite gets sold a lot, so it’s worth saying early: a facility with no unplanned breakdowns at all is probably over-maintaining, replacing parts that still had life left and shutting down lines that didn’t need to stop. Preventive maintenance also has a cost — hours, spare parts, negotiated downtime — and past a certain point each extra visit costs more than it prevents. The reasonable goal isn’t to eliminate corrective work, it’s for what remains to be the genuinely unpredictable, not the predictable that nobody predicted. You can only see the difference between those two with historical data.
Properly recorded, corrective maintenance is the best diagnosis of your preventive plan.
The same failure on the same unit three times isn’t bad luck: it’s a cause nobody has fixed, or an inspection that isn’t checking what it should.
If failures cluster toward the end of the interval, the interval is too long. It’s the cheapest adjustment to make, and almost nobody makes it.
It means the inspection task list is nominal: someone walks past it but doesn’t actually look, or nothing gets measured.
When the accumulated cost of repairs approaches the cost of replacement, it’s no longer a maintenance problem: it’s an investment decision.
If downtime is spent waiting rather than repairing, the problem is in the warehouse, not the workshop.
If both are climbing at once, the plan isn’t working, or there’s an external cause — usage, environment, operation — that maintenance alone can’t fix.
None of the above can be done with paper work orders, and not for lack of will: it requires every breakdown to be logged against its specific unit, with its date and its cause, in order to count anything. A folder with five hundred paper reports holds the information but doesn’t make it accessible, and reconstructing it by hand costs more than anyone is willing to invest. With visits logged against the asset, those six questions get answered by filtering. That’s the difference between having data and being able to use it, and it’s basically the whole reason a CMMS exists.
Leave your details and we will get in touch to see whether we fit. No commitment, no lock-in period.
The order that actually works is counterintuitive: sort out corrective maintenance first, then design the preventive plan — not the other way around. The reason is simple: a preventive plan designed without looking at the breakdown history is a plan based on the manufacturer’s catalog and whoever’s intuition put it together, and it usually overshoots in some places and falls short in others. With six months of well-recorded corrective work — against its unit, with its cause and its cost — the plan gets designed around what actually fails at your facility, which isn’t the same as what fails in general. What’s urgent at that point isn’t planning more, it’s recording better what’s already happening.
Say you have a pump on a quarterly inspection schedule that has failed four times in two years. If you look at the dates and find that three of the four failures happened in the last month of the quarter, the interval is too long: the unit runs fine for two months and the third is high-risk. Switching to a bimonthly schedule adds two visits a year and, if the pattern holds, prevents two breakdowns. The decisive comparison is simple: the cost of those two extra visits versus the cost of two unplanned stoppages, including what an hour of downtime actually costs. In the opposite case — three years of quarterly inspections finding nothing and no failures — the conversation flips: you’re probably over-inspecting. Neither calculation needs a sophisticated method; it needs the dates of failures logged against their unit, which is exactly what almost nobody has.
Between fixing it right now and preventing it altogether, there’s a third option that’s barely used: fixing it when it’s convenient. Plenty of failures don’t stop things from running — a burnt-out pilot light, a small contained leak, a backup fan that’s down — and dropping them into the urgent queue drives up cost for no reason. Logging them as scheduled corrective work lets you bundle them with the next visit to that site or the next shutdown on that line, which is where the real savings are: one trip instead of three. The condition for making this work is having them recorded instead of relying on the memory of whoever spotted them, which is exactly what doesn’t happen when the technician has nowhere to note it down in the moment.
There’s no universal figure, and anyone who gives you one without knowing your facility is guessing. It depends on your assets, on how much it costs when something stops, and on the plan you’re actually running. What you can measure is whether your corrective work is unpredictable or predictable.
No, and chasing that goal usually gets expensive: a facility with zero unplanned breakdowns is probably replacing parts that still had life left. The reasonable goal is for what remains to be the genuinely unpredictable.
Corrective work first. A preventive plan designed without looking at the breakdown history relies on the manufacturer’s catalog and intuition, and usually overshoots in some places and falls short in others. With a few months of well-recorded corrective work, the plan gets designed around what actually fails at your facility.
That comes later. Anticipating failures with condition data requires first having properly recorded the ones that already happen, because without history there’s no pattern to detect, sensors or not.
For execution, a spreadsheet can get by. For answering "how many times has this unit failed" or "where is spending concentrated," it can’t: you’d have to manually cross-reference hundreds of records every time the question comes up, which is exactly why nobody does it.
In the demo we’ll walk through it and see what would have needed to be recorded to explain why it keeps coming back.
Accessibility
Saved in this browser. Light or dark theme is set from the footer.