Industry 4.0 did not change the OEE formula.
It changed who does the arithmetic, how often, and whether anyone argues with the answer. What that means in practice — including why your OEE will probably fall.
OEE predates Industry 4.0 by about thirty years. It was defined in the 1980s as part of Total Productive Maintenance, and for most of its life it was calculated by hand, from a logbook, days after the shift it described.
What Industry 4.0 changed is not the formula. It is who does the arithmetic, how often, and — most consequentially — whether anyone argues about the answer. That distinction matters, because a great deal of Industry 4.0 marketing implies the metric itself is new, and plant managers who have calculated OEE on paper for fifteen years reasonably conclude they are being sold something they already have.
Four shifts, none of them to the formula.
| Traditional OEE | OEE under Industry 4.0 | |
|---|---|---|
| Who collects the data | An operator or supervisor, by hand, at shift end | The machine, continuously, as a by-product of running |
| When it is available | Days or weeks later, after compilation | While the shift it describes is still running |
| Resolution | Shift totals, with short stops lost to memory | Second-level, so a six-minute stop that happened forty times is visible |
| Status of the number | Contested — everyone has a version | Shared, because everyone is reading the same record |
| What it enables | Retrospective reporting | Intervention during the shift, while it can still change the outcome |
The fourth row is the one that changes behaviour. An OEE figure nobody disputes is a different management tool from one that opens every review meeting with an argument about whose number is right.
Manual OEE collection has a floor below which it cannot see, and that floor is roughly the length of a stop somebody bothers to write down. In practice that is around fifteen minutes. Everything shorter — the wait for a crane, the chip clearing, the tool change, the late start after tea — is invisible, not because anyone is concealing it but because recording it would cost more time than it takes.
On unmonitored floors those sub-fifteen-minute intervals routinely add up to more than the breakdowns that dominate the morning meeting. A plant that moves from manual to automatic collection therefore usually sees its OEE fall, sometimes substantially, and this is the single most misunderstood moment in a monitoring project. Nothing got worse. The measurement got honest.
Management that is not warned about this in advance tends to conclude the system is faulty. Management that is warned tends to treat the drop as the first useful finding, because the gap between the old number and the new one is a reasonable estimate of what was never being counted.
OEE sits at layer three, which is worth noting because most Industry 4.0 disappointment comes from stopping there. A plant with excellent dashboards and no layer four has bought reporting, not improvement. The formula is not the product; the decision it prompts is.
It says nothing about whether you made the right parts
A plant can post excellent OEE while running the wrong jobs and missing dispatch. OEE is a machine metric; schedule adherence is a production metric, and they are not interchangeable.
It compresses three different problems into one number
An OEE of 60% could be an availability problem, a speed problem or a quality problem, each with a different owner. The headline figure hides which, which is why the split matters more than the score.
It is not comparable between plants without shared definitions
Two plants computing OEE differently will produce numbers that cannot be compared, and frequently do while believing otherwise. The definitions are below.
| Factor | ISO 22400-2 — the standard | MachineWise variant — internal only |
|---|---|---|
| Availability | Actual production time ÷ planned production time | Machine on-time ÷ reporting time |
| Performance / Effectiveness | (Produced quantity × ideal cycle time) ÷ actual production time | Productive time ÷ machine on-time — a time-in-cut proxy |
| Quality | Good quantity ÷ produced quantity | Identical to ISO |
| Needs a maintained standard? | Yes — an ideal cycle time per part | No |
| Comparable to published benchmarks? | Yes | No — treat as your own baseline |
The variant exists because a floor running fifty part numbers a month often has no trustworthy ideal cycle time, and an OEE that cannot be computed until standards are maintained is an OEE that never gets computed. It is useful internally and it is not the ISO definition. Quote the ISO figure externally, and say which one you used.
If you already calculate OEE manually and you are considering automating it, the realistic expectation is not a better number. It is a lower, more detailed, less disputed number, available while you can still act on it, with the losses split into categories that have owners.
The commercial case follows from the detail rather than the metric. A plant that discovers two hours per machine per day of previously unlogged stops, concentrated on three machines and two repeating causes, has something it can act on this month. At a typical machine-hour rate across a thirty-machine floor, recovering even part of that runs into tens of lakhs annually — but the number that gets a corrective action approved is usually the Pareto, not the OEE.
If you want to see how the platform computes this continuously, that is covered on the OEE monitoring page. If you want to work your own figures first, the OEE calculator runs both definitions side by side in your browser.