What OEE actually measures
Overall Equipment Effectiveness (OEE) is a single percentage that answers one blunt question: of all the time a machine or line was supposed to be producing good parts at its rated speed, how much of it actually did? It rolls three separate losses — stopped time, slow running and bad parts — into one number you can track shift over shift and compare across machines.
The power of OEE is that it refuses to let any one loss hide behind another. A press can run every minute of the shift and still be a poor performer if it ran at half speed or produced a tray of rejects. A machine can hit its part count and still be losing money if a fifth of those parts get reworked. OEE catches all of that because it multiplies three factors together rather than averaging them.
The OEE formula, factor by factor
OEE has one master equation and three component equations underneath it. Learn the three components and the rest follows.
| Factor | What it captures | How it is calculated |
|---|---|---|
| Availability | Stopped-time loss — breakdowns, setups, changeovers, material starvation | Run time ÷ Planned production time |
| Performance | Speed loss — running slower than the ideal cycle, minor stops, idling | (Ideal cycle time × Total count) ÷ Run time |
| Quality | Defect loss — rejects and parts needing rework | Good count ÷ Total count |
| OEE | Availability × Performance × Quality | |
Availability starts from planned production time — the shift time you actually intended to run, after removing planned non-production such as scheduled breaks or a shift you never staffed. From that you subtract unplanned stops: breakdowns, waiting for material, over-long setups. What remains is run time. The work order and its operations define what should have run; the gap between that and reality is the availability loss.
Performance compares how fast the machine actually ran against its ideal cycle time — the standard time per part from the route sheet. If the standard says one minute a part and you made 350 parts in 420 run-time minutes, you ran slower than standard, and performance captures exactly that slippage.
Quality is the simplest to state and often the most painful: of everything produced, how much was good the first time. Rework counts against you here — a part that needed a second pass was not good first time, even if it eventually shipped. This is why capturing reject and rework at source matters so much to an honest OEE.
A full worked example
Take a single machine on an eight-hour shift. Work through it in order and the arithmetic is straightforward.
- Planned production time: 480 minutes (a full 8-hour shift you intended to run).
- Downtime: 60 minutes lost to a breakdown and a long setup, leaving 420 minutes of run time.
- Ideal cycle time: 1.0 minute per part (the route-sheet standard).
- Total count: 350 parts produced in those 420 minutes.
- Good count: 330 parts passed; 20 were rejected or sent to rework.
Now apply the three formulas:
- Availability = 420 ÷ 480 = 87.5%
- Performance = (1.0 × 350) ÷ 420 = 83.3%
- Quality = 330 ÷ 350 = 94.3%
- OEE = 0.875 × 0.833 × 0.943 = 68.7%
Notice what the single number does. No individual factor looked alarming — the worst was performance at 83.3% — yet the combined OEE is only 68.7%. That is roughly a third of the shift's potential quietly lost, and the multiplication is exactly why. If you improved only quality to 100% and left the rest alone, OEE would rise to 72.9%; the bigger prize here is the speed loss hiding in performance.
What is a good OEE score?
The famous benchmark is 85% — "world-class" — built from roughly 90% availability, 95% performance and 99% quality. It is a genuine target for a mature, single-product, high-volume line. It is also the wrong number to fixate on when you are starting out.
A discrete manufacturer measuring OEE honestly for the first time typically lands between 40% and 60%. That is not a failure; it is a baseline. The mistake is to see 85% quoted everywhere, compare your 52%, and conclude the metric is depressing or the data is wrong. The correct response is the opposite: a low first number means there is a lot of recoverable capacity, and the job is to move it up steadily and visibly.
Where the data for each factor comes from
OEE fails in most plants not because the maths is hard but because the data collection becomes a project of its own. The trick is to source each factor from execution data the floor already captures rather than a parallel logging exercise.
Availability comes from the gap between planned run time and actual — which the route and process-status booking already frames, because every operation carries a standard time and a completion record. Performance comes from comparing that standard cycle time against the actual quantity completed in the run window. Quality comes straight from the good-versus-reject split booked at each operation — the same good and reject WIP a disciplined shop already records. When those three streams are captured as a by-product of normal work-order execution, OEE stops being a separate initiative and becomes a report.
Want OEE from data you already book?
We can show you how operation standard times, process-status completion and good/reject WIP feed availability, performance and quality — on your own parts, in 30 minutes.
Common mistakes when measuring OEE
1. Starting the clock from the wrong place
If you calculate availability from calendar time rather than planned production time, an unstaffed night shift drags your number into meaninglessness. Define planned production time as the time you intended to run, and losses become actionable.
2. Guessing the ideal cycle time
Performance is only as trustworthy as the standard cycle time behind it. If standards are stale or invented, performance is fiction. Maintain operation standard times on the route sheet — the same standards that drive costing and scheduling.
3. Forgetting that rework fails the quality factor
A part that shipped only after a second operation was not good first time. Counting reworked parts as "good" flatters quality and hides the exact loss OEE exists to expose. Capture reject and rework where they happen.
4. Rolling one plant-wide OEE number
A single factory OEE averages away the constraint. OEE is most useful per machine or per work center, so you can see which station is actually holding back the line rather than a blurred plant average.
OEE data in Fast Production Software
Fast Production Software does not ask you to run a separate OEE data-capture project, because the three factors are already produced by ordinary shop-floor execution on the platform. Each OEE component maps to a place the data already lives:
The result is that OEE is grounded in the same linked chain the plant already runs — a released work order, its route, its material issue, its good and reject WIP — rather than a spreadsheet a supervisor updates by hand at the end of each shift. To go deeper on the execution layer that feeds these numbers, start with the pillar guide, what is production management software.
Frequently asked questions
What is OEE?
OEE — Overall Equipment Effectiveness — is a single percentage that measures how much of a machine or line's planned production time produced good parts at rated speed. It is the product of three factors: Availability (was the machine running when it should have been), Performance (did it run at its rated speed) and Quality (were the parts good the first time). OEE = Availability × Performance × Quality. An OEE of 100% means every planned minute made a good part at full speed, which never happens in practice; most plants that begin measuring honestly land between 40% and 60%.
What is the OEE formula?
OEE = Availability × Performance × Quality. Availability = run time ÷ planned production time. Performance = (ideal cycle time × total count) ÷ run time. Quality = good count ÷ total count. Because the three factors are multiplied, a weak result in any one of them pulls the whole score down — 90% availability, 90% performance and 90% quality multiply to only 72.9% OEE, not 90%.
What is a good OEE score?
The widely quoted world-class benchmark is 85% (roughly 90% availability, 95% performance, 99% quality). But 85% is a destination, not a starting point. A typical discrete manufacturer that has never measured OEE begins around 40% to 60%. The right target is not 85% on day one — it is a measured baseline plus steady, visible improvement against it.
Where does OEE data come from on the shop floor?
OEE needs three streams of data: planned versus actual run time (Availability), standard cycle time versus actual output (Performance), and good versus reject quantity (Quality). In a production system such as Fast Production Software, all three already exist as by-products of normal execution — operation standard times from the route sheet, actual completion from process-status booking, and good and reject WIP captured at each operation. That means OEE can be measured from data the floor already books.
What is the difference between OEE and utilisation?
Utilisation usually means only how much of the available time a machine ran — it is close to the Availability factor alone. OEE is stricter: a machine can be running all shift (high utilisation) yet still score a low OEE because it ran slowly or produced rejects. OEE therefore exposes hidden losses that a simple utilisation figure conceals, which is exactly why it is the more honest shop-floor metric.