Every factory has a number nobody wants to look at too closely. It's the total time the machines stood idle last month, adding up micro-stoppages, tool changes, missing operators, equipment failures, waiting on raw material. That number is usually far larger than management imagines, and the financial impact goes well beyond the parts that never got produced.
Siemens, in its True Cost of Downtime report, calculated that the world's 500 largest companies lose 11% of annual revenue to unplanned downtime. In a tight-margin operation, that's the difference between a profitable year and a year in the red.
And the worst part is that most of this cost is invisible: it never shows up in any consolidated report because it was never measured properly.
Why machine downtime is invisible
In most factories, downtime is logged the same way it was in 1985: an operator notes down on a spreadsheet (or on a sheet of paper taped to the machine) the time it stopped, the reason, and the time it came back. That note depends on the operator's goodwill and memory, on legible handwriting, and on someone later transcribing it into a system nobody ever opens.
The result is predictable. Micro-stoppages of 2 or 3 minutes don't get logged because "it's not worth writing down". Long stoppages get generic reasons like "maintenance" or "adjustment", without the detail needed to attack the root cause. And when the production engineer pulls the report at month's end, they're looking at a sanitized version of reality, with 30 to 50% of the stoppages missing. You can't reduce what you don't measure properly.
How to truly know why machines stop
The first change when you apply industrial IoT is simple: the machine starts logging the stoppage on its own, the second it happens. The current sensor or the PLC signal detects that the cycle stopped, records the exact time, and the micro-stoppage nobody would have noted down enters the system automatically.
But knowing that it stopped is only half the problem. The other half is understanding why, and this is where reason classification by the operator on screen comes in. Instead of writing a justification from memory at the end of the shift, they get a notification the moment the stoppage occurs and classify it in 5 seconds: tool change, waiting on raw material, mechanical failure, quality adjustment. The information comes out clean, classified, and traceable.
The Pareto that appears once you start measuring right
Every time a factory turns on automatic downtime monitoring for the first time, it discovers the same pattern. It's not the big, rare stoppages that are eating into production. It's the small, frequent ones, which individually seem irrelevant but together account for 60 to 70% of the idle time.
A tool change that could be optimized, a recurring wait for the forklift, a recipe adjustment that always takes longer than it should. This invisible Pareto is where the greatest potential for gains lies, and most of these problems have a solution the factory's own team already knows. What's missing isn't the solution, it's visibility into the frequency and the accumulated cost.
From knowing to reducing
With stoppages mapped and classified, the path to reducing them opens up on three fronts. The first is the most obvious: attack the Pareto, project by project, with method. The second is predictive maintenance: use the same data that reveals the stoppage to anticipate the next one. Motor current analysis detects wear weeks before the failure.
McKinsey notes that real-time monitoring programs combined with predictive maintenance reduce maintenance costs by up to 25% and unplanned breakdown events by up to 70%.
The third front is the one that shows up in more mature factories: using AI to suggest adjustments that prevent the stoppage before it even appears. This is the territory of Software Defined Manufacturing, where the intelligence layer above the PLC proposes setpoints that reduce equipment stress, balance load across machines, and adapt the operation to process variations. Those who reach this level stop fighting downtime and start operating in a regime where it simply happens less often.
The cost you can't see is the one costing you the most
Machine downtime doesn't fix design, it fixes decisions. And decisions are fixed with data. Every factory that still logs downtime on a spreadsheet, or that accepts reports with vague reasons like "maintenance" and "adjustment", is leaving money on the table for a simple reason: it can't see where it's losing it.
The question worth asking today is how much of your margin is being consumed by stoppages nobody is really measuring, and how long this has been happening without anyone noticing.
