What Is OEE (Overall Equipment Effectiveness) and How Do You Calculate It?

What is OEE?

OEE (Overall Equipment Effectiveness) is a manufacturing metric that measures how much of your planned production time is actually productive. It combines three factors into one score: OEE = Availability × Performance × Quality. A perfect OEE score of 100% means a machine ran with zero stoppages, at full speed, producing only good parts — which in practice almost never happens.

OEE was developed by Seiichi Nakajima in the early 1970s as part of Total Productive Maintenance (TPM) at the Japan Institute of Plant Maintenance. It's now the standard productivity metric across manufacturing, used by maintenance teams, plant managers and operations consultants to answer one question: of all the time this machine was scheduled to run, how much of it actually created value?

Most plants that start measuring OEE for the first time are surprised by the number. An OEE of 100% is theoretical. An OEE of 60% is closer to the honest average across manufacturing globally, and it's common for a first measurement to land lower still — not because the equipment is unusually bad, but because most factories have never measured the gap between "the machine was on" and "the machine was actually producing good parts at full speed."

How do you calculate OEE?

OEE is the product of three separate scores, each between 0% and 100%:

OEE = Availability × Performance × Quality

Each component isolates a different kind of loss:

Component	What it captures	Formula
Availability	Time lost to unplanned stops (breakdowns, changeovers) and planned stops	Run Time ÷ Planned Production Time
Performance	Speed lost to slow cycles and small stops when the machine is running	(Ideal Cycle Time × Total Count) ÷ Run Time
Quality	Output lost to defects and rework	Good Count ÷ Total Count

These are the standard definitions used across TPM literature and OEE tracking tools, and they matter because each one points to a different fix. A low Availability score is a maintenance and changeover problem. A low Performance score is a speed and micro-stoppage problem. A low Quality score is a process-control problem. Blending them into one number without knowing the breakdown tells you almost nothing about what to fix first.

A worked OEE calculation, start to finish

Here's a full calculation using one 8-hour shift on a single machine.

1. Establish Planned Production Time. Shift length is 480 minutes. Subtract 30 minutes of scheduled breaks. Planned Production Time = 480 − 30 = 450 minutes.

2. Establish Run Time and calculate Availability. During the shift, the machine was down for 45 minutes (a jam and a changeover). Run Time = 450 − 45 = 405 minutes. Availability = 405 ÷ 450 = 90.0%.

3. Count output and calculate Performance. The machine's ideal cycle time is 1 unit per minute. In 405 minutes of Run Time it produced 344 total units (running below its rated speed for part of the shift due to a partially blocked feed). Performance = (1 × 344) ÷ 405 = 84.9%.

4. Count good parts and calculate Quality. Of the 344 units produced, 14 were rejected at inspection. Good Count = 344 − 14 = 330. Quality = 330 ÷ 344 = 95.9%.

5. Multiply the three scores. OEE = 0.900 × 0.849 × 0.959 = 73.4%.

That single number — 73.4% — says that just under three-quarters of this shift's planned time went into good, full-speed output. The other 26.6% is broken down and traceable: roughly 10% lost to stoppages, another chunk to running slow, and a smaller amount to rework. That breakdown, not the headline number, is what you act on.

What is a good OEE score?

A commonly cited rule of thumb is that 85% OEE is "world-class." This figure comes from Nakajima's original TPM benchmarks, built from Japanese manufacturers winning the Distinguished Plant Prize for TPM implementation in the 1970s and 80s — companies scoring 85%+ OEE, typically made up of roughly 90% Availability, 95% Performance and 99% Quality.

It's worth being precise about what that figure actually is: an industry-cited rule of thumb from one country, one era and one industry (Japanese automotive manufacturing), not a universal engineering standard. OEE tracking sources also report that most manufacturers average closer to 60% OEE, and that more plants score below 45% than above 85%. If your first measurement comes back at 55–65%, you're not an outlier — you're typical, with real room to close the gap.

The number to compare against isn't really "85%." It's your own trendline. A factory that moves from 58% to 68% OEE over a year has made a bigger real improvement than a factory sitting flat at 80%. Use 85% as a directional target, not a pass/fail line.

How does a small manufacturer start tracking OEE without expensive software?

You don't need a full MES or IIoT sensor package to start. Most small manufacturers can get a usable first OEE reading with a clipboard, a stopwatch and a spreadsheet:

1. Pick one bottleneck machine — the one that most limits your output — rather than trying to instrument the whole floor at once.
2. Log every stop, its cause and its duration, on a simple downtime sheet at the machine. This is the same discipline that underpins reliable downtime cost tracking — you can't calculate Availability without it.
3. Record output and rejects each shift: total units and good units. This gives you Performance and Quality with no extra equipment.
4. Total it up weekly in a spreadsheet using the three formulas above, and watch the trend rather than any single shift's number.

This is also exactly the data a CMMS captures automatically once you have one — work orders logged against a machine give you Availability for free, and most CMMS platforms will chart OEE-style trends without a separate system. If you're already running a preventive maintenance schedule, you're most of the way to an OEE-ready downtime log already; OEE just turns that maintenance data into a productivity number the whole plant can act on.

The point of starting manually isn't to stay manual forever. It's to prove the number is worth tracking before you spend money proving it with software.

FAQ

Is OEE the same as machine utilisation? No. Utilisation typically only measures whether a machine was running (similar to the Availability component alone). OEE also accounts for speed losses and quality losses, which is why an OEE score is almost always lower than a utilisation figure for the same machine.

What's a bad OEE score? There's no formal cutoff, but scores below roughly 40% typically point to major, addressable losses — frequent breakdowns, chronic small stops, or a quality problem generating significant rework — rather than a single obvious cause.

Can OEE be measured for a whole factory, not just one machine? Yes, but it's most useful at the individual machine or line level first, ideally starting with your bottleneck. A single factory-wide OEE figure blends good and bad performers together and hides which specific machine to fix.

Do you need sensors or IoT to calculate OEE accurately? No. Manual logging of stops, output and rejects produces a mathematically identical OEE score to an automated system — it's just slower to collect and easier to under-record. Automation improves accuracy and reduces admin effort; it doesn't change the formula.

How often should OEE be recalculated? Daily logging with a weekly review is a practical starting cadence for most small manufacturers — frequent enough to catch a slipping trend, infrequent enough not to become an admin burden on its own.

If you want a straight answer on where your bottleneck machine is actually losing time — and whether the fix is maintenance, changeovers, speed or quality — book a lean and process audit with Ardmor. We work with Victorian manufacturers on exactly this: turning downtime, maintenance and production data into a small number of decisions worth making.

This article is general information for manufacturers, not professional advice. Last updated: 20 September 2026.