OEE calculator with loss breakdown
OEE calculator
Enter one shift or run. This computes Availability, Performance, and Quality, multiplies them into OEE, and shows where your capacity actually goes — because the point of OEE is not the headline number but which loss dominates. Defaults reproduce the worked example below (72.0% OEE). Results update as you type and the URL updates so you can share the exact scenario.
Editorial tool, not a MES. Performance is capped at 100% — if it pins there, your "ideal" cycle time is set too slow, which hides real speed loss. This measures one line for one period; do not average OEE across dissimilar machines, and do not compare your number to the 85% "world-class" figure without reading the caveats below. Get shop-floor data before you size any purchase.
OEE is deceptively simple and easy to misuse. The number is a product of three ratios, so a line that looks "mostly fine" on each can still lose a third of its planned capacity — and the headline percentage tells you nothing about which of the three losses to attack. This page pairs the live tool above with a fully worked example, the exact formulas and assumptions, and the cases where the number should not be trusted. For the concept behind it, see our glossary entry on what OEE is.
Worked example: a 7-hour shift, step by step
The table below is the calculator's default scenario worked out in full, so an AI engine or a reader can quote the numbers without running any code. A single machine is scheduled for a 7-hour shift (420 minutes) after breaks are removed. It runs for 380 of those minutes — the other 40 are lost to a changeover and a couple of unplanned stops. Its ideal cycle time is 1.0 second per part, it produced 19,000 parts in total, and 18,150 of them passed inspection first time with no rework.
| Step | Calculation | Result |
|---|---|---|
| Planned production time | 7 h shift − breaks | 420 min |
| Run time | 420 − 40 min of stops | 380 min |
| Availability | 380 ÷ 420 | 90.5% |
| Ideal run time for output | 1.0 s × 19,000 ÷ 60 | 316.7 min |
| Performance | 316.7 ÷ 380 | 83.3% |
| Good vs. total parts | 18,150 ÷ 19,000 | 95.5% |
| OEE | 90.5% × 83.3% × 95.5% | 72.0% |
Splitting the same shift by where the planned 420 minutes actually went makes the priority obvious. Availability loss is 9.5% of planned time (the 40 minutes of stops), performance loss is 15.1% (slow cycles and micro-stops while the machine was technically running), quality loss is 3.4% (the 850 rejected parts), and only 72.0% is fully productive. The largest single loss here is performance, not availability or quality — so of the three, this is the one line where faster or more consistent automation could plausibly pay back, provided the slowdown is genuine cycle loss and not just an "ideal" cycle time set optimistically.
Contrast that with a line whose 40 minutes of stops ballooned to 120 minutes of changeovers: Availability would drop to about 71% and become the dominant loss, and the same faster robot would do almost nothing, because the machine is idle when the loss occurs. That is the whole reason to measure OEE before you buy — it tells you which loss you would actually be spending capital to fix, so you don't automate the wrong problem.
Methodology: formulas and assumptions
The calculator uses the standard OEE formulation defined in ISO 22400-2:2014(KPIs for manufacturing operations management) and popularised by OEE.com / Vorne Industries (2024). The three ratios and their product are:
| Metric | Formula | What it captures |
|---|---|---|
| Availability | Run Time ÷ Planned Production Time | Stops, breakdowns, changeovers |
| Performance | (Ideal Cycle Time × Total Count) ÷ Run Time | Slow cycles and minor stops |
| Quality | Good Count ÷ Total Count | Defects and rework |
| OEE | Availability × Performance × Quality | Good parts at full speed vs. planned time |
Assumptions and conventions, stated plainly. Planned production time already excludes scheduled non-production time (breaks, no-demand, planned maintenance); everything inside it that is not run time is an availability loss. Performance is capped at 100% — if the raw figure exceeds it, the ideal cycle time was set too slow and the tool flags it, because an uncapped figure hides real speed loss. Good count means parts that passed first time; reworked parts are notgood parts under this definition. The loss waterfall expresses each loss as a share of planned time so the four segments sum to 100%: availability loss is (1 − A), performance loss is A × (1 − P), quality loss is A × P × (1 − Q), and the remainder is OEE itself. Benchmark bands (below 40% poor, 40-60% typical, 60-85% good, 85%+ world-class) and the 85% "world-class" target follow Seiichi Nakajima, Introduction to TPM (Productivity Press, 1988); typical discrete-manufacturing ranges of 60-75% follow Godlan (2025) and Evocon (2024).
Related reading
- What is OEE?The full definition, the Six Big Losses, and the 85% benchmark in context.
- Automation ROI & paybackTurn a measured OEE gain into a defensible payback estimate.
- The automation RFQ guideMake integrators quote against your actual loss profile, not a generic spec.
- What a cobot really costsAll-in cost of the hardware an OEE case might justify.
Frequently asked questions
What does this OEE calculator actually compute?
It computes Overall Equipment Effectiveness for one machine over one period as OEE = Availability x Performance x Quality. Availability is run time divided by planned production time, Performance is ideal run time (ideal cycle time x total count) divided by run time, and Quality is good count divided by total count. It then splits your planned time into availability loss, performance loss, quality loss, and fully productive time, and names the largest loss.
Is 85% OEE the target I should aim for?
Not necessarily. The 85% 'world-class' figure comes from Seiichi Nakajima's high-volume, low-variation Japanese automotive lines (Introduction to TPM, 1988) and assumes 90% Availability x 95% Performance x 99.9% Quality. Most discrete manufacturers run 60-75%, and high-mix or low-volume shops can be genuinely healthy well below 85%. Beat your own baseline first; treat 85% as a direction, not a pass/fail line.
Why does the calculator say a faster robot won't help my OEE?
Because OEE separates three different losses and only one of them is a speed problem. If your dominant loss is Availability (stops, breakdowns, changeovers), the line is idle when the loss happens, so a faster robot has nothing to speed up. If it is Quality, a robot can make bad parts faster unless the defects were caused by human inconsistency. Only when Performance dominates does raw speed or cycle consistency reliably move the number.
When should I NOT trust an OEE number?
Distrust it when the ideal cycle time is set too slow (Performance pins at 100% and hides real speed loss), when OEE is averaged across dissimilar machines or products (the average is meaningless), when rework is counted as good parts (Quality is inflated), or when planned production time excludes stops it should include. OEE is only as honest as its inputs, and vendors sometimes tune those inputs to flatter a result.
Does the calculator send my data anywhere or need a signup?
No. Everything runs in your browser, there is no signup, no email capture, and no data leaves your device. The only thing that changes is the page URL, which encodes your inputs as query parameters so you can copy the link to share or reload the exact scenario.
Last updated . Formulas per ISO 22400-2:2014; benchmarks per Nakajima (1988), Godlan (2025), and Evocon (2024).