How is OEE calculated in manufacturing?
OEE multiplies availability, performance and quality, so every loss shows up in the result.
- Availability = run time ÷ planned production time. It is lost to breakdowns, changeovers and other stops.
- Performance = (ideal cycle time × total count) ÷ run time. It is lost to minor stops and slow cycles.
- Quality = good count ÷ total count. It is lost to defects and rework, including start-up scrap.
- OEE = availability × performance × quality.
Worked OEE example
A packaging line is planned to run for 480 minutes in a shift, with an ideal cycle time of 0.1 minutes per unit.
| Factor | Shift data | Result |
|---|---|---|
| Availability | Ran 400 of 480 planned minutes | 83% |
| Performance | Made 3,600 units; 4,000 were possible in 400 minutes | 90% |
| Quality | 3,420 of 3,600 units were good | 95% |
| OEE | 0.83 × 0.90 × 0.95 | 71% |
A quick check: 3,420 good units at 0.1 minutes each is 342 fully productive minutes, and 342 ÷ 480 is also 71%.
The six big losses
OEE comes from total productive maintenance, where it makes six losses visible: breakdowns, set-up and adjustment, minor stops, reduced speed, defects and rework, and start-up losses. Each one lowers one of the three factors, so a falling OEE points to where the problem is.
Using OEE well
- Compare a machine with itself over time, not one machine with another. Different equipment has different ideal cycle times and planned schedules.
- Look at the three factors, not only the headline. A 70% OEE from low availability needs maintenance and changeover work; the same figure from low quality needs process control.
- Record the losses behind the numbers. OEE tells you how much time was lost; the reason codes tell you why.
- Review it daily. Yesterday's OEE and its biggest loss belong on the Tier 1 board, with chronic losses escalated into problem solving.
Common mistakes
- Chasing a benchmark figure such as "world class 85%" without understanding what is in the planned time.
- Excluding changeovers or planned maintenance to make the number look better, which hides the losses that could be reduced.
- Treating OEE as a report for management rather than a signal for the team running the machine.
See also why OEE is important and what to look for in OEE software.