Overall equipment efficiency: OEE formula and example

Overall equipment efficiency usually refers to overall equipment effectiveness, or OEE: availability multiplied by performance multiplied by quality. It measures how much planned production time becomes good output at the ideal production speed. The factors separate time lost to stops, running below speed and producing rejected units.

The OEE formula

OEE = Availability × Performance × Quality

Availability = Run minutes / Planned production minutes

Performance = Ideal cycle minutes × Total units / Run minutes

Quality = Good units / Total units

Multiply the factors as decimals, then format the result as a percentage. The OEE calculation guide also gives the equivalent direct calculation: good count multiplied by ideal cycle time, divided by planned production time.

That identity matters because a production report should produce the same answer both ways. If it does not, the factors probably use different shifts, different unit definitions or rounded inputs.

The denominator is the time production was intended to run. The guide to OEE factors distinguishes time outside the production schedule from stops during planned production, including changeovers. Write down your plant's calendar policy before comparing machines. A maintenance event cannot move between those categories just because a supervisor wants a cleaner score.

Start with the production and downtime exports

Use one summarized row per machine, shift and product standard. A US manufacturer of metal components supplies the following records for October 22, 2026. Each machine made one product throughout its recorded period; their scheduled windows differ.

Planned minutes come from the approved shift schedule after excluded breaks. Stop minutes come from the downtime log, with overlapping events removed. Run minutes equal planned minutes minus stops. Counts come from completed production and inspection records, using the same unit of measure and reporting cutoff.

Keep machine ID, product code, shift date, standard revision and rejection reason alongside these columns. They let the production lead explain the result without treating the summarized table as the original evidence. Manufacturing KPIs place this equipment measure beside delivery, output and cost measures.

An ideal cycle time is a defensible fastest cycle for that machine and product, not a sales target or a deliberately reduced budget rate. Product changes require their own standards. A rate stated in units per hour must be converted before it is multiplied by minutes.

Work each machine through the factors

All percentages below are calculated from the unrounded inputs and displayed to two decimal places.

Machine B has the same speed performance but weaker availability and quality. Buying a faster machine would not directly address those two measured differences. The maintenance lead should investigate the stop reasons; the quality lead should examine the rejected production by lot and cause.

Quality here counts units accepted on their first pass through the measured process. Keep reworked units out of the good count for that original pass. Use first-pass yield for the separate quality investigation rather than expanding OEE into a complete defect report.

Prove the result with a minutes reconciliation

Each loss can be expressed in ideal production minutes. That makes the result auditable without adding percentages that use different denominators.

The proof is 110 + 70 + 45 + 585 = 810 planned minutes. Combined OEE is therefore 585 / 810 = 72.22%. The factors calculated from combined inputs also agree: availability is 700 / 810 = 86.42%, performance is 630 / 700 = 90.00%, and quality is 1,170 / 1,260 = 92.86%.

The simple average of the machine percentages is 71.75%, which is wrong for the combined period because their planned durations differ. With different product cycle times, sum each row's ideal minutes before dividing; do not multiply a mixed unit total by one convenient cycle time.

These are independent machines. For stations connected in series, overall line efficiency uses the line's final good output and bottleneck standard, avoiding repeated counts of the same product.

Where OEE goes wrong

• Changing the denominator. Removing a stop from planned time improves the percentage without producing another good unit. Keep excluded time visible and approved.
• Using an easy speed standard. Performance above the physical limit suggests an incorrect cycle time, count or time unit. Investigate the inputs before accepting the result.
• Counting rework twice. Recounting a rejected part as new production inflates total count and may hide the original quality loss.
• Subtracting small stops twice. If small stops remain inside run time and reduce performance, do not also subtract them as availability downtime.
• Mistaking OEE for available capacity. OEE describes scheduled production. Capacity utilization addresses another denominator and requires a separate statement of available capacity.

Review lost minutes with their owners before setting improvement targets. Availability belongs with maintenance and production scheduling; speed losses need process review; quality losses need inspection evidence. A rising total is useful only when the underlying records show which loss actually fell.

Compare your machine records in Covirage

Bring a production export, downtime records and the approved cycle-time standards, with machine and shift identifiers. Covirage maps the columns, and its tools compute the factors and check the minutes reconciliation; the external AI model explains the differences with every figure traced to its rows. Create a free Covirage account, which takes about a minute, and bring a sample of your own data after setting up your business.

See how the same file-based approach supports capacity planning.