How to calculate OEE: formula and example
OEE says how much of a machine's planned time went into good parts at full speed. The formula, a worked example for one CNC machine and one shift, and what a good OEE is.
Ward De Coster
Published 26 October 2026

What is OEE?
OEE, overall equipment effectiveness, is the share of a machine's planned time that went into good parts at full speed. At 100 % it made only good parts, as fast as it can, all the time it was meant to run. OEE is availability, performance and quality multiplied, so it also shows which of the three costs you most.
Seiichi Nakajima made it known with Total Productive Maintenance (TPM), in his book Introduction to TPM (1984, in English 1988).
How is OEE calculated?
OEE = availability × performance × quality
- Availability is the time the machine ran, divided by its planned time. Breakdowns, setups and waiting for material lower it.
- Performance is the parts it made times their ideal cycle, divided by the time it ran. Slower cycles and short stops lower it.
- Quality is the good parts divided by all the parts it made. Scrap lowers it.
Multiplied out, the three come to one line, OEE = good parts × ideal cycle ÷ planned time.
The ideal cycle is the time one part takes at full speed. Take the machine's own quickest cycles for that program, not the time in the quote. A generous quote makes performance look better than it is.
A worked example
Take one CNC machine on an early shift from 06:00 to 14:00, with a 30-minute break. The program it runs has an ideal cycle of 4 minutes. During the shift it stands still for 120 minutes, with a 45-minute setup, 35 minutes waiting for material, 25 minutes of tool changes and a 15-minute alarm. It makes 72 parts, 3 of them scrap.
| Factor | Worked out | Result |
|---|---|---|
| Planned time | 480 min less the 30 min break | 450 min |
| Availability | 330 min run ÷ 450 min planned | 73,3 % |
| Performance | 72 parts × 4 min = 288 min ÷ 330 min run | 87,3 % |
| Quality | 69 good ÷ 72 made | 95,8 % |
| OEE | 73,3 % × 87,3 % × 95,8 % | 61,3 % |
In one line, 69 good parts × 4 minutes is 276 minutes of good parts at full speed, of 450 planned, so 61,3 %. Availability is the biggest loss here, and the setup the longest stop in it.
What counts as planned time?
Planned time is the time a machine is meant to make parts, its shifts less their breaks. A weekend or a night nobody works isn't planned, so a machine standing then costs no OEE. Some shops also take out planned maintenance or meetings. Whatever you choose, keep it the same, or one week's OEE won't compare with the next.
Do setups lower OEE?
Yes. A setup is planned time in which the machine makes no parts, so it lowers availability, as the 45-minute setup did above. Vorne, who publish oee.com, count setup and adjustments among the six big losses, as an availability loss, and changeovers as its largest source.
A job shop that changes over several times a day loses more to setups than a plant that runs one part all week, so its OEE sits lower for the same good work. Leaving setups out of planned time would raise the figure, and hide the hours a quicker setup wins back.
What is a good OEE?
85 % is the figure most often called world class. Vorne trace it to Seiichi Nakajima, who noted that the plants winning Japan's Distinguished Plant Prize all had an OEE above 85 %. They also write that most manufacturers are closer to 60 %.
With short runs and many setups, a job shop will sit lower than a plant that makes one part all week. Compare each machine with itself, week on week. An OEE that rises says more than a benchmark.
Sources
- Seiichi Nakajima, Introduction to TPM: Total Productive Maintenance, Productivity Press, 1988 (in Japanese 1984).
- Vorne Industries, World-Class OEE, oee.com, read on 26 October 2026.
- Vorne Industries, OEE Six Big Losses, oee.com, read on 26 October 2026.
How Dextrolab works it out
Dextrolab works out OEE for every machine and every shift, straight from the machine's control and the scrap your operators log. No forms to fill in, and you see at once which of the three costs you most. See OEE in Dextrolab, or book a demo.