How to calculate OEE
OEE answers one question: of all the time the machine was supposed to be producing, how much became good parts at the ideal speed? The math splits the answer into three losses, and each has a different cause.
OEE = Availability × Performance × Quality
| Factor | Formula | What it measures |
|---|---|---|
| Availability | run time ÷ planned production time | losses from stops (breakdowns, changeovers, no material) |
| Performance | (ideal cycle time × total count) ÷ run time | speed losses (minor stops, reduced speed) |
| Quality | good count ÷ total count | scrap and rework losses |
Planned production time is the shift minus planned stops such as meals, breaks and scheduled maintenance. Run time is planned production time minus unplanned stops.
Worked example
The values already filled in the calculator describe an 8-hour shift with a 30-minute meal break:
- Planned production time: 480 − 30 = 450 min
- Unplanned stops: 50 min, so run time = 400 min
- Ideal cycle time: 30 s per part
- Total count: 680 parts, of which 20 were rejected
The three factors:
- Availability = 400 ÷ 450 = 88.9%
- Performance = (30 s × 680) ÷ (400 × 60 s) = 20,400 ÷ 24,000 = 85.0%
- Quality = 660 ÷ 680 = 97.1%
OEE = 0.889 × 0.850 × 0.971 = 73.3%.
In minutes, the 450 planned minutes produced 330 minutes’ worth of good parts at ideal speed. The other 120 were lost like this: 50 to stops, 60 to running below ideal speed and 10 making the 20 rejected parts. Here the biggest loss is performance, not the stops that usually get the blame.
How to read the result
- 85% or more: world class for discrete manufacturing. Few lines sustain it.
- 60% to 85%: where most plants that measure and work on losses sit.
- Below 60%: there are big losses, usually easy to find. 40% is common when measurement starts.
The number alone says little. What tells you what to do is which factor is lowest:
- Low availability: log the stops and build a Pareto by cause. Breakdowns call for preventive maintenance and root cause analysis; long changeovers call for SMED.
- Low performance: look for minor stops nobody records (jams, adjustments, waiting for an operator) and machines running below rated speed. Also check that the ideal cycle time is right.
- Low quality: go after the main scrap cause with a fishbone diagram and 5 Whys, and count rework as a loss even when the part is saved.
Common OEE mistakes
- Using the average cycle instead of the ideal cycle. The average already includes speed losses and makes performance look close to 100%.
- Taking changeovers out of planned time. The loss disappears from the metric and stops being worked on.
- Counting rework as good. Quality in OEE means good the first time.
- Comparing OEE across different machines. OEE tracks how one machine or line improves over time; it is not a league table.
- Measuring by hand once a month. Short, forgotten stops are exactly what drags performance down. The closer to real time the logging, the more honest the OEE.