Overall Equipment Effectiveness isn't one number — it's three independent loss factors multiplied together.
A machine can run 100% of its scheduled time — perfect Availability — and still post a terrible OEE if it runs slow (poor Performance) or turns out defective parts (poor Quality). That's not a contradiction; it's exactly what OEE is designed to reveal. OEE = Availability × Performance × Quality, and because the three factors are multiplied rather than averaged, a genuine weakness in any single one of them drags the whole number down no matter how excellent the other two are. Collapsing OEE straight to one blended percentage — without also reporting the three factors that produced it — throws away the exact information a maintenance or production team needs to know what to actually go fix.
Availability — was the machine actually running? Of the time it was scheduled to produce, how much time was it actually running versus stopped for breakdowns, changeovers, or other unplanned/planned downtime? Availability = Run Time ÷ Planned Production Time.
Performance — while it ran, was it running at full speed? Of the time it was running, how much output did it actually produce compared to what it could have produced at its ideal, rated cycle time? This captures slow cycles and small stops that never show up as recorded downtime. Performance = (Ideal Cycle Time × Total Count) ÷ Run Time.
Quality — of what it made, how much was actually good? Of the total units produced, how many were good, first-pass units versus scrap or units needing rework? Quality = Good Count ÷ Total Count.
If OEE were an average of the three factors, a machine could be terrible at Quality and still post a respectable score as long as Availability and Performance were excellent. Multiplication doesn't allow that: 87.5% × 80% × 90% is not close to the average of those three numbers (85.8%) — it's meaningfully lower, at 63%, precisely because every factor has to individually survive the multiplication. A weakness anywhere shows up everywhere in the final number. That's the entire reason OEE is structured this way instead of as one blended efficiency percentage: it forces every loss category to earn its place in the final score, so a team can't hide a bad Quality number behind a good Availability number. The tradeoff is that the single OEE percentage, taken alone, tells you almost nothing about which factor is the problem — for that, you need to look at all three numbers, not just their product.
Not quite. A single OEE percentage describes the combined effect of three losses without telling you which one is actually driving the number. As the two-machine comparison above shows, a mediocre 60% OEE can come from a pure downtime problem, a pure quality problem, or any blend in between — and those call for completely different fixes. Throwing resources at faster changeovers won't help Machine B at all; its equipment runs fine, it just makes bad parts. Tightening process control won't help Machine A; its parts are fine, it just doesn't run enough. This is exactly why OEE is reported as three separate factors, not collapsed into one undifferentiated score from the start— the single number flags that there's a problem; the three factors tell you where to look.
Overall Equipment Effectiveness (OEE) measures shop-floor productivity as the product of three independent loss factors: Availability (the share of scheduled production time the equipment was actually running, versus downtime from breakdowns and changeovers), Performance (the share of its ideal, rated production speed the equipment actually achieved while running, capturing slow cycles and minor stops), and Quality (the share of total units produced that were good, first-pass units, versus scrap or rework). OEE = Availability × Performance × Quality. Because the factors multiply rather than average, a genuine weakness in any single factor pulls the whole OEE score down even when the other two factors are excellent — which is exactly why OEE is reported as three separate numbers, not a single blended score, from the start.
If OEE combined its three factors by averaging them, a machine with excellent Availability and Performance could mask a genuinely poor Quality rate and still post a deceptively respectable overall score. Multiplying the three fractions together removes that loophole: every factor has to individually survive the calculation, so a weak factor anywhere in the chain drags the final percentage down everywhere. An 8-hour, 480-minute shift with 87.5% Availability, 80% Performance, and 90% Quality doesn't average to something close to 85.8% — it multiplies down to 63%, because losses compound rather than cancel out.
Two machines can land on nearly identical OEE percentages for entirely different reasons. A machine with 63% Availability, 95% Performance, and 100% Quality has a pure downtime problem — it runs fast and clean whenever it's actually running, but it isn't running often enough, so the fix is reducing breakdowns and changeover time. A machine with 98% Availability, 96% Performance, and 64% Quality has a pure quality problem — it's almost always running at full speed, but a third of its output is scrap or rework, so the fix is process control, tooling, or incoming material, not uptime. Both machines can land at roughly the same 60% OEE despite needing completely different corrective actions.
A single OEE number is genuinely useful as a top-level health indicator and for tracking trend over time, but on its own it cannot tell a team which of the three losses to attack first — that information only exists in the three individual factors that produced it. Reporting OEE alongside its Availability, Performance, and Quality components (rather than the blended percentage alone) is what turns OEE from a scoreboard number into an actual diagnostic tool that points a maintenance or production improvement effort at the right root cause.
OEE = Availability × Performance × Quality. Availability = Run Time ÷ Planned Production Time. Performance = (Ideal Cycle Time × Total Count) ÷ Run Time. Quality = Good Count ÷ Total Count. All three are expressed as percentages (or decimal fractions) and multiplied together, not averaged.
Multiplying the three factors ensures that a genuine weakness in any single one of them meaningfully lowers the overall score, even if the other two factors are excellent. Averaging would let a strong Availability and Performance mask a poor Quality rate (or any other single weak factor), producing a misleadingly favorable overall number.
Yes. A machine can run for its entire scheduled time — 100% Availability — and still post a poor overall OEE if it runs below its ideal rated speed (poor Performance) or produces a high proportion of scrap or rework (poor Quality). Availability alone says nothing about how fast the machine ran or how good the output was.
Because the single OEE number is the product of three factors, and many different combinations of Availability, Performance, and Quality can multiply to roughly the same overall percentage. A machine weak mainly in Availability has a downtime problem that calls for reducing breakdowns and changeover time; a machine weak mainly in Quality has a scrap/rework problem that calls for tightening process control instead. The overall OEE score alone doesn't distinguish between these cases — the three individual factors do.
The Six Big Losses framework groups OEE's losses under its three factors: Availability losses are equipment failures/breakdowns and setup/changeover time; Performance losses are idling/minor stops and reduced/slow-running speed; Quality losses are process defects (scrap/rework) and reduced-yield losses at startup. Each of the six maps to exactly one of the three OEE factors.
Try our Industrial & Systems Engineering Studio
More calculators, simulators, and guides for this discipline.