You calculate OEE by multiplying three percentages: availability × performance × quality. Below we work this out with an example of one shift on one machine.
The starting point
- Planned production time: 480 minutes (an eight-hour shift, breaks already deducted)
- Downtime during that shift: 60 minutes (breakdowns, changeovers and waiting)
- Ideal cycle time: 1 minute per product
- Produced: 380 products
- Of which good: 370 products
Step 1: availability
The machine ran for 480 − 60 = 420 minutes. Availability is then 420 / 480 = 87.5%.
Step 2: performance
In 420 minutes, at the ideal cycle time, the machine could have made 420 products. It made 380. Performance is 380 / 420 = 90.5%.
Step 3: quality
Of the 380 products, 370 were good. Quality is 370 / 380 = 97.4%.
Step 4: OEE
0.875 × 0.905 × 0.974 = 0.771. The OEE of this shift is therefore about 77%.
The quick check
There is a shorter route to the same result: good products × ideal cycle time, divided by planned time. Here: 370 × 1 minute / 480 minutes = 77.1%. If the two calculations don’t match, there is an error somewhere in your data.
Common mistakes
- Using a cycle time that is too generous. Taking the average instead of the ideal cycle time hides speed losses and gives a performance figure that looks too good.
- Counting planned maintenance as downtime, or not. Either is fine, as long as you do it the same way everywhere. Write the agreement down.
- Forgetting short stops. Recording downtime by hand mainly misses stops of a few minutes. They then end up unnoticed in performance.
With automatically collected machine data, these calculations are made continuously, per machine, line and shift. That is how the FactoryNebula OEE dashboard works too.



