A single OEE percentage tells you almost nothing on its own. It is a product of three ratios, and three very different floors can arrive at the same number. Before the figure is useful you need to know which component is dragging it down — that is where the actual decision lives.
Why one number hides the answer
OEE multiplies three things together:
- Availability — was the machine running when it was supposed to be?
- Performance — when it ran, did it run at its rated rate?
- Quality — of what it made, how much was good?
Because they multiply, a 60% OEE could be a machine that is available 95% of the time but running slowly, or one running at full rate that spends a third of the shift stopped. Those two floors need opposite interventions. The composite number is a reporting metric; the components are the diagnostic.
Read availability first
Availability is usually where the recoverable hours are, and it is the component least likely to be measured accurately by hand. The reason is micro-stops: a 30-second stop for a chip clear or a tool change doesn't get written on a logbook sheet, but two hundred of them across a shift is well over an hour.
That is also why a floor's own downtime log and its measured availability tend to disagree. The log captures the stops long enough to be worth writing down. The short ones vanish, and they are frequently the larger total.
A worked example
Illustrative only — these are not customer figures, just arithmetic to show how the components combine.
Take one machine, one 8-hour shift, with a 30-minute planned break:
| Input | Value |
|---|---|
| Planned production time | 450 min |
| Unplanned stops (all causes) | 68 min |
| Run time | 382 min |
| Rated cycle time | 4.0 min/part |
| Total parts produced | 82 |
| Good parts | 79 |
That gives:
- Availability = 382 ÷ 450 = 84.9%
- Performance = (4.0 × 82) ÷ 382 = 85.9%
- Quality = 79 ÷ 82 = 96.3%
- OEE = 0.849 × 0.859 × 0.963 = 70.2%
The headline is 70%. The useful reading is that availability and performance are each costing about 15%, and quality is nearly fine — so chasing scrap here would be optimising the one component that is already working.
The number only means something if the definitions hold still
Most OEE arguments are definition arguments wearing a disguise. Three settle most of them:
- What counts as planned production time. Breaks, planned maintenance and unstaffed shifts either belong in the denominator or they don't. Either convention works; changing it mid-quarter makes the trend meaningless.
- What the rated cycle time is. Leave it blank and performance loss becomes a dumping ground for everything unexplained.
- When a stop becomes downtime. A threshold of a few seconds is a decision about whether micro-stops are visible at all.
Pick each once, write them down, and keep them fixed. A slightly wrong definition held constant still produces a usable trend. A correct definition that changes every month does not.
What to do with it
OEE is a starting point, not a target to manage directly. The workflow that actually moves it is narrow: find the largest single loss, ask why it happens, remove one cause, then check whether the component you targeted moved. If the number goes up and you cannot say which loss shrank, you have not learned anything you can repeat.
Want the arithmetic on your own shift numbers? The OEE calculator shows every step, and the full OEE guide works through the six big losses on a machining floor.