What Is OEE? Meaning, Formula and Calculation Example
Learn what OEE means, calculate it with a worked example, and track production losses using shift records, machine monitoring or Bulk.
OEE stands for overall equipment effectiveness. In manufacturing, it measures how much of the time scheduled for production becomes good output made at the ideal speed. The formula is OEE = Availability × Performance × Quality. A machine can be running for most of a shift and still have low OEE because it runs slowly or makes parts that need rework. Vorne’s introduction to OEE explains this distinction between running and fully productive time.
The useful part is the loss behind the percentage. This guide shows how to calculate OEE for one hypothetical shift, keep changeovers in the calculation and turn the result into a specific improvement decision.
What is the OEE formula?
The three factors answer different questions. Keep their inputs within the same machine or line, product and measurement period.
| Factor | Calculation |
|---|---|
| Availability | Run time ÷ planned production time |
| Performance | (Ideal cycle time × total count) ÷ run time |
| Quality | First-pass good count ÷ total count |
Availability shows the time lost to stops. Performance compares output with the ideal production rate. Quality measures the share that passed without rework.
Multiply the ratios as decimals, then multiply by 100 to express OEE as a percentage. For a single product, the same calculation simplifies to:
OEE = (Good count × ideal cycle time) ÷ planned production time.
Use matching time units. If the cycle time is in seconds, convert production time to seconds too. The shorter formula checks the result; the three-factor version explains the losses. Both are set out in Vorne’s OEE calculation guide.
Define planned production time before the shift
Planned production time is the period when you intend to produce. A lunch break with no production scheduled can sit outside it. A changeover during that production period stays inside it and reduces availability, even when the changeover was planned. A planned stop is different from time when no production was scheduled. Vorne’s OEE factors guide separates those two cases.
Write down that boundary before collecting data. Otherwise, two supervisors can record the same shift and report different scores simply by excluding different stops.
Set the ideal cycle time for the product
Ideal cycle time is the fastest achievable time per unit under optimal conditions. It is not the average cycle time or a relaxed planning target. A machine’s headline rate may need translating into time per part for the actual product and tooling. If performance exceeds 100%, check the cycle-time setting, counts and time units. Vorne explains the ideal cycle time input.
For a mixed-product shift, calculate each run using its own ideal cycle time. To combine them, sum each run’s good count multiplied by its ideal cycle time, then divide by the combined planned production time. Do not take a simple average of percentages from runs of different lengths.
How to calculate OEE: a worked example
Imagine a press making one type of bracket during an eight-hour shift. These figures are hypothetical, not Bulk customer results.
| Shift record | Value |
|---|---|
| Shift length | 480 minutes |
| Lunch with no production scheduled | 30 minutes |
| Planned production time | 450 minutes |
| Changeover within production time | 30 minutes |
| Breakdown | 45 minutes |
| Run time | 375 minutes |
| Ideal cycle time | 1.5 seconds per bracket |
| Total output | 13,500 brackets |
| First-pass good output | 13,230 brackets |
| Scrap or parts requiring rework | 270 brackets |
Availability: (450 − 30 − 45) ÷ 450 = 83.33%.
Performance: (1.5 × 13,500) ÷ (375 × 60) = 90%.
Quality: 13,230 ÷ 13,500 = 98%. The 270 rejected parts do not count as first-pass good output even if some are later repaired.
OEE: (375 ÷ 450) × 0.90 × 0.98 = 73.5%. Use unrounded ratios in the calculation and round the displayed result.
Check it with the shorter formula: 13,230 × 1.5 seconds is 19,845 seconds, or 330.75 minutes of fully productive time. Dividing that by 450 planned minutes gives the same 73.5%.
The largest loss here is 75 minutes of stops, followed by 37.5 equivalent minutes of speed loss and 6.75 equivalent minutes of quality loss. Start by inspecting the breakdown and changeover records. The final percentage alone cannot tell you which to tackle.
The six big losses behind OEE
Availability, performance and quality organise the loss. A reason code explains what happened. The six big losses framework gives you a useful starting point:
| Loss | Example |
|---|---|
| Equipment failure (availability) | A failed drive stops the press |
| Setup and adjustments (availability) | Tooling is changed for the next bracket |
| Idling and minor stops (performance) | Repeated short misfeeds interrupt the cycle |
| Reduced speed (performance) | The press runs below its ideal rate |
| Process defects (quality) | Brackets fail inspection during the run |
| Startup rejects (quality) | Initial pieces fail while settings stabilise |
Agree when a short interruption becomes a separately recorded stop. A small stop left within run time reduces performance through the missing output. If you instead subtract it from run time, it reduces availability. Use one consistent rule so the same lost time is not deducted twice.
What is a good OEE score?
85% is a widely cited reference for discrete manufacturing, not a universal pass mark. Vorne’s discussion of world-class OEE explicitly cautions against applying one target to every process. Product mix, changeover frequency and measurement rules affect the comparison.
Use a stable baseline for the same process and inspect the factors alongside it. In our example, investigate the 45-minute breakdown before declaring 73.5% acceptable or unacceptable. A useful next target could be tied to preventing that specific failure, with quality and safe operation checked at the same time.
Compare like with like. A long run of one product and a shift with several necessary setups have different demands. Removing setups from the denominator makes the numbers look closer while hiding the work involved.
OEE versus utilisation and TEEP
OEE measures effectiveness during planned production time. It does not tell you how much of the entire week was scheduled. TEEP, or total effective equipment performance, extends the calculation to all calendar time: TEEP = OEE × utilisation, where utilisation is planned production time divided by all time. Vorne’s TEEP guide describes that boundary.
For example, a line with 80% OEE that is scheduled for half the calendar period has 40% TEEP. That does not mean the other half should automatically be filled with production. Demand, staffing and maintenance still determine whether more scheduled time would be useful.
How to track OEE on the shop floor
Start with one machine and one shift. You need a production window, a reliable time record and counts that refer to the same run. Choose where parts will be counted and who will confirm first-pass acceptance before collecting the first score.
Start with a shift sheet or spreadsheet
Keep a run summary and a separate stop log. The summary holds the calculation inputs; the log explains the downtime behind them. A practical starting sheet needs these fields:
| Record | What to capture |
|---|---|
| Run identity | Machine, date, shift, job and product |
| Planned production | Start and end times, with any unscheduled periods identified |
| Cycle standard | Ideal cycle time for this product, its units and parts per cycle |
| Stops | Start, end and reason for each stop counted as downtime |
| Output | Total parts and first-pass good parts from the same counting point |
| Quality losses | Scrap and parts needing rework, kept distinct |
Record stops close to the event. At a product change, start a new run record so the correct ideal cycle time follows the work. If a counter runs continuously, use the difference between its opening and closing readings and account for any reset.
At handover, reconcile the sheet before calculating the score. In our example, the stop log must explain 75 minutes, and the quality record must explain the 270 brackets that did not pass first time. If inspection is still outstanding, mark the result as provisional rather than assuming every part is good. Our scrap and rework guide explains the distinction behind those counts.
Check your own run with the OEE calculator, or use the printable OEE tracking sheet to collect the inputs before calculating a score.
Automate signals that are difficult to record by hand
For frequent short stops or fast cycles, machine signals can provide a more detailed event record than an end-of-shift estimate. Keep operator input for context the signal cannot establish, such as why the machine stopped. A stopped signal alone does not distinguish a tooling problem from missing material.
During a trial, compare the captured events with an observed run. Check the stop threshold, parts per cycle, counter resets and rejected-part counts. Make sure a short stop is not deducted from runtime and also counted as a speed loss. Keep an identifiable gap when data is missing so it can be investigated.
For the reporting tradeoffs, see OEE dashboards versus spreadsheets.
Examples of OEE tracking systems
These examples show different ways to collect and use production records. The useful test is whether a system can explain your shift using definitions your team understands.
Vorne XL: sensors and production monitoring
Vorne XL uses sensors to collect production signals such as cycle times, downtime and rejects, then presents OEE factors and underlying losses in reports and floor displays. Its XL manual describes automatic run detection using cycle inputs and configurable part settings, including the threshold that separates a short interruption from a down event.
For a trial on the example press, check that the input counts brackets correctly and that the stop settings capture the changeover and breakdown consistently. Confirm how the reason for each event reaches the report.
MachineMetrics: machine data with production context
MachineMetrics describes OEE monitoring that combines machine data with schedules, part cycle standards and quality inputs. Its published workflow uses shift, break and holiday schedules for availability; compares cycle times with part standards for performance; and accepts scrap data from operators and external inspection sources for quality.
Check which cycle standard the system is using on your parts. An ERP planning standard or a baseline derived from observed production needs checking against the ideal cycle time used in this article. Ask to trace a score back to the schedule, part record and quality entries that produced it.
Bulk: capture downtime while running production
In Bulk, downtime capture is part of the production workflow. When OEE downtime tracking is enabled for an asset type, pausing its production timer requires a reason. The pause opens a downtime event; resuming the timer closes that event and records its duration. OEE & Analytics provides recorded-time views and downtime reasons so the team can review where recorded time went.
On the example press, the operator would pause for the changeover, select the configured setup reason and resume when production restarts. The breakdown would have its own pause and reason. Those records give the handover a specific starting point: the two events account for 75 minutes to investigate.
Bulk’s recorded-time score measures running time as a share of running time plus logged downtime. With the same 375 running minutes and 75 downtime minutes, that score would be 83.3%. The conventional OEE calculation above is 73.5%, because it also includes speed and first-pass quality losses. To calculate that full measure, you still need the ideal cycle time, total count and first-pass good count for the same run.
Turn OEE tracking into an improvement routine
At handover, review the underlying losses with the people who can explain them. If a large share of downtime is labelled “Other”, inspect those events and make the reason choices more useful. A correct duration with an unclear cause gives the team little to act on.
Choose one loss, assign an owner and set a date to check the result. For the worked example, begin with the 45-minute breakdown. On the next comparable run, check the cause, downtime and accepted output together. Keep the same time boundary and counting rules so a better score reflects changed production rather than changed arithmetic.
When evaluating OEE software, ask the supplier to reproduce one complete shift from its underlying records. Have them show how setup time, product changes, missing data and rework affect the result. The team should be able to explain both the percentage and the next action it supports.
