Last reviewed: June 17, 2026
OEE is calculated by multiplying three ratios: Availability times Performance times Quality. Scrap rate equals scrap units divided by total units produced. First-pass yield equals units that pass all quality checks on the first attempt divided by total units started. All three formulas are simple. The errors in how most plants apply them are not.
Key takeaways:
OEE = Availability x Performance x Quality. A plant with Availability of 0.90, Performance of 0.85, and Quality of 0.98 has an OEE of 75%.
First-pass yield is not the same as scrap rate. A line with 2% scrap and 13% rework has 85% FPY. The rework cost is real even when the part eventually ships.
The most common OEE calculation error is using average cycle time instead of best-demonstrated cycle time in the Performance component. It makes Performance look better than it is and hides speed losses.
Use the Sharpen OEE Calculator to run your numbers without building a spreadsheet.
THE OEE FORMULA, EXPLAINED
OEE = Availability x Performance x Quality.
Each component is a decimal between 0 and 1. The three multiply together, so losses in any one component compound against the others.
Availability is the share of planned production time the equipment was actually running. Downtime from breakdowns, unplanned changeovers, and tooling failures reduces Availability. Planned events such as scheduled maintenance and standard break times are excluded from the denominator before you start.
Availability = (Planned production time - Downtime) / Planned production time
Performance is how fast the equipment ran compared to its best-demonstrated cycle time. A machine that can run 60 parts per hour when everything is right, but averaged 51 parts per hour during a shift, has a Performance of 0.85. The denominator is best-demonstrated rate, not the average and not the nameplate spec.
Performance = (Total parts produced x Ideal cycle time) / Run time
Quality is the share of output that passed inspection on the first attempt. Parts that were reworked and later shipped do not count as good for OEE purposes. Only parts that came off the line clean go in the numerator.
Quality = Good parts / Total parts produced
| COMPONENT | FORMULA | WORKED EXAMPLE |
|---|---|---|
| Availability | (Planned time - Downtime) / Planned time | (480 - 48) / 480 = 0.90 |
| Performance | (Parts x Ideal cycle time) / Run time | (367 x 1 min) / 432 min = 0.85 |
| Quality | Good parts / Total parts produced | 360 / 367 = 0.98 |
| OEE | Availability x Performance x Quality | 0.90 x 0.85 x 0.98 = 0.75 |
A WORKED OEE EXAMPLE WITH REAL NUMBERS
A stamping press runs an 8.5-hour paid shift. The 30-minute unpaid lunch break is removed, leaving 480 minutes of planned production time.
During the shift, an unplanned die change overrun and a jammed feeder strip consume 48 minutes of production time. The machine is running for the remaining 432 minutes; this is the run time used in the Performance calculation.
The best-demonstrated rate for this press is 1 part per minute. During the 432 minutes of run time, the press produced 367 parts. Seven parts failed final inspection and were scrapped.
Availability = (480 - 48) / 480 = 0.90
Performance = (367 parts x 1 min/part) / 432 min run time = 0.85
Quality = (367 - 7) / 367 = 360 / 367 = 0.98
OEE = 0.90 x 0.85 x 0.98 = 0.75, or 75%
A result of 75% sits at the top of the typical range for plants with an active OEE program. World-class benchmarks by process type start above 80% for most high-volume operations. In this example, the gap shows up almost entirely in Performance (0.85), which points to speed losses and micro-stops as the priority before addressing anything else.
THE FIRST-PASS YIELD FORMULA
First-pass yield (FPY) = units passing all quality checks on the first attempt / total units started.
This is not the same as scrap rate. Scrap rate only counts the parts you discard permanently. FPY counts every part that needed any rework before it could ship, whether scrapped or not. A plant with a 2% scrap rate and a 13% rework rate has an FPY of 85%. Both numbers are accurate. They measure different costs.
FPY = (Total units started - Scrapped units - Reworked units) / Total units started
Worked example: A machining cell starts 1,000 parts. Twenty are scrapped. One hundred and thirty require rework before they can ship.
FPY = (1,000 - 20 - 130) / 1,000 = 850 / 1,000 = 85%
Scrap rate = 20 / 1,000 = 2%
The plant tracking only scrap rate sees a 2% problem. The plant tracking FPY sees a 15% loss, 13 points of which it is paying labor to rework. That rework cost is real even though the parts shipped.
| METRIC | WHAT IT COUNTS | FORMULA | EXAMPLE (1,000 UNITS) |
|---|---|---|---|
| Scrap rate | Parts permanently discarded | Scrapped / Total produced | 20 / 1,000 = 2% |
| First-pass yield | Parts right on the first attempt | (Total - Scrapped - Reworked) / Total | (1,000 - 20 - 130) / 1,000 = 85% |
THE SCRAP RATE FORMULA
Scrap rate = scrap units / total units produced.
In unit-based manufacturing (assembly, stamping, machining), count is the right denominator. In weight-based operations (casting, forging, injection molding), cost or weight is often more meaningful because a large defect costs more to produce than a small one.
| DENOMINATOR | WHEN TO USE IT | FORMULA |
|---|---|---|
| Unit count | Assembly, stamping, machining | Scrap units / Total units produced |
| Weight | Casting, forging, injection molding | Scrap weight / Total material input |
| Cost | Financial reporting | Scrap cost / Total material cost |
Scrap rate should be tracked by reason code, not just in aggregate. An overall rate of 3% tells you almost nothing actionable. A Pareto that shows 60% of scrap comes from tooling failures on one press gives you a place to go. In plants we have worked with, the aggregate scrap rate is tracked weekly; the reason-code breakdown is what drives corrective action.
THE FOUR MISTAKES THAT MAKE YOUR NUMBERS LIE
Mistake 1: Using average cycle time as the ideal rate. Average cycle time includes speed losses by definition. If you set the ideal rate equal to what you average, Performance starts at 100% and you have no speed-loss visibility at all. The ideal rate must be the best-demonstrated actual rate from a sustained run under normal conditions, not the nameplate speed and not the average.
Mistake 2: Inconsistent planned downtime treatment. Planned downtime is excluded from planned production time before calculating Availability. Whether changeovers, scheduled maintenance, and operator breaks are classified as planned or unplanned must be defined once and applied the same way across all shifts and all equipment. In plants we have worked with, inconsistent definitions between shifts create 5 to 10 point Availability differences that are entirely definitional, not operational.
Mistake 3: Counting rework as good output in Quality. OEE Quality measures parts that passed clean on the first attempt. If a reworked part goes into the good count, the Quality component is overstated and the OEE number is higher than actual line performance. In high-rework operations, this error inflates OEE by 5 to 10 points, which means the plant is chasing an improvement target that does not reflect reality.
Mistake 4: Treating OEE as a capacity utilization metric. OEE only measures performance during scheduled production time. If a machine sits idle because there are no orders, that idle time does not reduce OEE at all. A plant with 80% OEE and 30% unscheduled idle time is not performing at 80% of its potential output. Total Effective Equipment Performance (TEEP) captures this by including idle time in the denominator, but most plants do not need TEEP unless they are evaluating capital decisions or justifying new equipment purchases.
FREQUENTLY ASKED QUESTIONS
WHAT IS A GOOD OEE NUMBER FOR A SMALL MANUFACTURER?
For most discrete manufacturers, 65% to 75% is the typical range for plants with an active OEE program, and 80% is a reasonable near-term target. World-class benchmarks in high-volume operations run above 85%, but the right number depends on your process type. See the full breakdown in OEE, Scrap Rate, and FPY Benchmarks by Process Type.
DOES REWORK COUNT AGAINST FIRST-PASS YIELD?
Yes. Any unit that required rework before it could ship counts as a first-pass failure, even if the rework was minor and the part eventually met spec. FPY measures what came off the line right the first time.
WHAT CYCLE TIME SHOULD I USE IN THE OEE FORMULA?
Use the best-demonstrated actual cycle time from a sustained production run under normal conditions. Do not use the nameplate speed from the equipment manual and do not use the average cycle time. Using the average sets your baseline at your current performance level, which makes Performance 100% by definition and hides every speed loss.
CAN I CALCULATE OEE ON A SPREADSHEET?
Yes, the formulas are simple arithmetic. The harder part is collecting consistent shift-level data for downtime minutes, parts produced, and parts rejected. Many plants start with a paper log per shift and a daily spreadsheet rollup before investing in automated data collection. The Sharpen OEE Calculator runs the calculation interactively if you would rather not build the spreadsheet yourself.