Overall equipment effectiveness (OEE) tells you what share of your planned production time produced good parts at full speed. It is one number built from three factors. Get the inputs honest and it points straight at where the hours are going. To skip the arithmetic, enter your shift in the free OEE calculator; for a short definition, see the glossary entry for OEE.
The formula
OEE = Availability x Performance x Quality
The factors multiply, they do not average. A line running 90% availability, 90% performance and 90% quality is at 0.9 x 0.9 x 0.9 = 72.9% OEE, not 90%.
Availability
Availability = run time / planned production time.
- Planned production time is the shift minus time that was never meant to produce: breaks, meetings, planned maintenance.
- Run time is planned production time minus stops: breakdowns, changeovers, material waits, jams long enough to log.
Example: a 480 minute shift with a 30 minute break and a 20 minute meeting has 430 minutes of planned production time. Stops total 43 minutes, so run time is 387 minutes. Availability = 387 / 430 = 90.0%.
Performance
Performance = (ideal cycle time x total count) / run time.
Ideal cycle time is the fastest rate the machine sustains in normal running, not the nameplate and not the standard you quote to customers. Performance captures slow cycles and micro-stops that never made it into the downtime log.
Example: ideal cycle time is 0.5 minutes per part and the machine made 696 parts in 387 minutes. Performance = (0.5 x 696) / 387 = 348 / 387 = 89.9%.
Quality
Quality = good parts at first pass / total parts.
Count first pass. A part that is reworked or scrapped later is a quality loss even if it ships eventually.
Example: 696 parts, 656 good first time. Quality = 656 / 696 = 94.3%.
Putting it together
0.900 x 0.899 x 0.943 = 76.3% OEE for the shift. These numbers are an arithmetic illustration, not data from any plant.
Check it with the shortcut: OEE = good count x ideal cycle time / planned production time = 656 x 0.5 / 430 = 328 / 430 = 76.3%. The two agree because the factors cancel.
OEE with mixed parts and different cycle times
When one machine runs two or more parts in a shift, each part has its own ideal cycle time. A single cycle time for the whole shift gives a wrong performance figure. Weight each part by its ideal time instead:
- Availability = run time / planned production time, as before. A changeover between the parts is a stop, so it is an availability loss.
- Performance = sum of (ideal cycle time x total count) for each part / run time.
- Quality = sum of (ideal cycle time x good count) / sum of (ideal cycle time x total count). Weighting quality by ideal time keeps a scrapped long-cycle part from counting the same as a scrapped short one.
Example: 430 minutes planned. Part A has a 0.5 minute ideal cycle, part B 1.2 minutes. A 25 minute changeover and 20 minutes of other stops leave 385 minutes of run time. The machine made 400 of part A (384 good) and 120 of part B (114 good).
- Availability = 385 / 430 = 89.5%
- Performance = (0.5 x 400 + 1.2 x 120) / 385 = (200 + 144) / 385 = 344 / 385 = 89.4%
- Quality = (0.5 x 384 + 1.2 x 114) / 344 = (192 + 136.8) / 344 = 328.8 / 344 = 95.6%
- OEE = 0.895 x 0.894 x 0.956 = 76.5%. Shortcut check: 328.8 / 430 = 76.5%.
With one part, these formulas reduce to the single-part ones above and to what the OEE calculator computes. Two common errors: using part A's 0.5 minute cycle for all 520 parts gives performance of 0.5 x 520 / 385 = 67.5%, which blames the machine for running part B at its own rate; and quality as plain good / total (498 / 520 = 95.8%) no longer multiplies back to the shortcut.
Plant OEE across several machines
Roll up by time, not by averaging percentages. Plant OEE = sum of (good count x ideal cycle time) across machines / sum of planned production time.
Example: machine 1 is the shift above (328.8 good minutes in 430 planned). Machine 2 ran 430 planned minutes at 50% OEE, which is 215 good minutes. Machine 3 ran a short 200 minute shift at 90% OEE, which is 180 good minutes.
- Plant OEE = (328.8 + 215 + 180) / (430 + 430 + 200) = 723.8 / 1,060 = 68.3%
- Simple average of the three: (76.5 + 50 + 90) / 3 = 72.2%. Too high: machine 3's 200 minutes count as much as a full shift.
- Multiplying the three: 0.765 x 0.5 x 0.9 = 34.4%. Wrong: that is the math for machines in series where every part passes all three, and even then it is not OEE.
For the factor view of the plant, roll each factor up the same way: availability = total run time / total planned time, performance = total ideal minutes made / total run time, quality = total good ideal minutes / total ideal minutes made. On a line, the number that sets output is the OEE of the bottleneck, so report it next to the plant figure.
The six big losses
Map every logged stop to one of six categories so the data tells you which factor to work.
- Availability: equipment failure, setup and adjustment
- Performance: minor stops, reduced speed
- Quality: process defects, startup rejects
Five mistakes that make OEE look better than it is
- Planned time that is not really planned. Changeovers and waiting on material belong in availability losses.
- Ideal cycle time taken from the nameplate or an old standard. Use the best sustained rate.
- Quality counted at final inspection. The rework loop disappears from the number.
- Micro-stops buried in run time. Short stops are often the largest hidden performance loss.
- Averaging across lines or days. Track per shift and per line, because averages hide patterns.
What to do with the number
- Work the lowest factor first.
- Chase OEE on the constraint. Improving non-bottleneck equipment mostly builds inventory. See how to find the bottleneck.
- If downtime is the problem, start with machine downtime and low OEE, and put a price on it with the downtime cost calculator.
- For the wider method, read the OEE guide.
Calculate yours
Use the free OEE calculator. Enter planned time, stops, ideal cycle time and counts to see all three factors.
Need to keep scenarios and break losses out by category? OEE Calculator Pro does that, and it comes with every plan starting at Starter.
Questions
What is the OEE formula?
OEE = Availability x Performance x Quality, where availability = run time / planned production time, performance = (ideal cycle time x total count) / run time, and quality = good count / total count.
Is there a shortcut for the OEE calculation?
Yes. The three factors cancel down to OEE = (good count x ideal cycle time) / planned production time. In the shift on this page: 656 x 0.5 / 430 = 76.3%. Use it to check the three-factor result, but keep the three factors, because they tell you where the loss is.
Can OEE be over 100%?
No. If it comes out above 100%, the ideal cycle time is set slower than the machine really runs, or the counts and times come from different periods.
How do you calculate OEE for multiple products with different cycle times?
Weight everything by ideal time. Performance = sum of (ideal cycle time x total count) for each part / run time. Quality = sum of (ideal cycle time x good count) / sum of (ideal cycle time x total count). Changeovers between the parts are availability losses. The shortcut still holds: OEE = sum of (good count x ideal cycle time) / planned production time.
How do you calculate plant OEE across several machines?
Add up good count x ideal cycle time across all machines and divide by the total planned production time of those machines. Do not multiply the machine OEEs together and do not take a simple average: multiplying understates the plant, and a simple average lets a machine on a short shift count as much as one on a full shift.
Are breaks and planned maintenance part of OEE downtime?
No. They come out of the shift before you start, to give planned production time. Breakdowns, changeovers and material waits stay in, as availability losses.