How to measure OEE: start tracking without a big IT project

To measure OEE, record planned time, downtime, total count and good count for one machine on every shift.

Overall equipment effectiveness multiplies availability, performance and quality into one number. The number matters less than the losses behind it.

You can start measuring OEE this week with a sheet of paper. This guide explains the formula, what to record and when software starts to help.

How to calculate OEE

Availability is run time divided by planned production time. Performance is ideal cycle time times total count, divided by run time. Quality is good count divided by total count.

OEE is the three multiplied together. The OEE calculator does the math and walks through an example step by step.

Agree these rules before the first shift

OEE is only comparable from shift to shift if the inputs are recorded the same way. If one shift counts a changeover as planned time and another as downtime, the review turns into an argument. Agree these first:

Scope: Exactly which machine, line or cell the number covers.
Planned production time: The time the asset is meant to run.
Planned exclusions: Which breaks, meetings or no-order periods come out, and which do not.
Ideal rate: The expected speed for each product or product family.
Good count: How first-pass good parts are counted.
Reject count: How scrap, rework and start-up rejects are handled.
Stop reasons: A short list operators can pick from quickly.
Review rhythm: Who looks at the numbers, and how often.

Write the rules down on one page and keep it at the machine. Change them only on purpose, and note the date, or trends become meaningless.

The six major losses

OEE is usually analyzed through six big losses: breakdowns, setup and adjustment, minor stops, reduced speed, start-up rejects and production rejects. Seiichi Nakajima described them in his work on TPM.

Each factor of OEE holds two of the losses. Availability holds breakdowns and setup and adjustment, since both stop the machine. Performance holds minor stops and reduced speed, since the machine runs but makes less than it should. Quality holds start-up rejects, made while the process settles after a start or changeover, and production rejects, made during steady running.

Knowing which loss is largest tells you which lean tool to use next.

These losses add up across an economy. A NIST study of US manufacturing estimated that preventable maintenance issues cost manufacturers $119.1 billion in 2016, mostly in sales lost to delays and defects, on top of $57.3 billion spent on machinery maintenance.

OEE for one shift: a worked example

Here is an illustrative shift with made-up numbers. A packing machine is planned to run for 480 minutes. It stops for 60 minutes: a 35-minute breakdown and a 25-minute changeover. Its ideal cycle time is 1.0 minute per case. It makes 380 cases, and 361 pass inspection.

Availability: run time is 480 − 60 = 420 minutes, so availability is 420 / 480 = 87.5%.

Performance: at the ideal rate, 420 minutes would make 420 cases. The machine made 380, so performance is 380 / 420 = 90.5%.

Quality: 361 / 380 = 95.0%.

OEE: 87.5% × 90.5% × 95.0% = 75.2%. A quicker check gives the same result: 361 good cases × 1.0 minute / 480 minutes = 75.2%.

In minutes, the shift lost 60 minutes to stops, 40 minutes to running below the ideal rate and 19 minutes making cases that failed. That is 119 minutes of the 480, and it shows the losses in a form the team can compare.

Start with one constraint

Measure the bottleneck machine first. It sets the output of the whole line, so gains elsewhere may not raise output.

Measuring is a management habit in its own right. A Census Bureau survey of about 35,000 US manufacturing plants, analyzed by Nicholas Bloom and colleagues in the American Economic Review in 2019, found that structured management practices, such as monitoring performance and setting targets, account for more than 20% of the variation in productivity between plants.

A good pilot area usually has these traits:

A steady schedule: a clear shift pattern.
A known product family: with a realistic target rate.
People who can record: operators or team leaders able to log stops and counts.
Losses worth understanding: problems the team already feels.
Someone who will review: a supervisor or improvement lead who looks at the numbers every week.

An OEE tracking sheet on paper

Put a sheet at the machine with one row per hour. Operators record the planned output, the actual output, stops longer than a few minutes and a reason code for each.

Keep the reason codes short and agreed with the team, such as no material, changeover, breakdown and quality check. Review the sheet at the daily huddle.

Here is the illustrative shift above as an hourly sheet. Planned output is 60 cases an hour at the ideal rate.

Hour 1: 32 made. Changeover 25 min (C/O). 12 start-up rejects.

Hour 2: 55 made. Three jams, 6 min in total (minor stops).

Hour 3: 52 made. Jams, 5 min (minor stops).

Hour 4: 20 made. Hydraulic leak, 35 min (breakdown).

Hour 5: 56 made.

Hour 6: 57 made. 3 production rejects.

Hour 7: 54 made. Waiting for cartons, 4 min (no material). 2 production rejects.

Hour 8: 54 made. 2 production rejects.

The total is 380 made and 19 rejected, matching the example. The recorded minor stops add up to 15 minutes, so 25 of the 40 performance minutes went to slow running that nobody wrote down. That gap is normal on paper, and it is one reason teams later add a counter on the machine.

Sort the losses into a Pareto, largest first: breakdown 35 min, reduced speed 25 min, changeover 25 min, rejects 19 min, minor stops 15 min. The leak goes first, through TPM, and the changeover second, through SMED.

Make losses visible

Show the loss Pareto at the shift huddle, so the team can act on it. Andon signals help the team react while the loss is happening, not the next day.

Setup losses point to SMED. Breakdowns and minor stops point to TPM.

Which loss to attack first

Start with the factor that loses the most time, then break it into reasons the team can influence:

Main loss	Look first at	Typical action
Availability	Breakdowns, changeovers, waiting for material or people	Maintenance response, setup preparation, scheduling, material readiness
Performance	Slow cycles, minor stops, unstable feed, worn parts	Settings, tooling, work methods and repeated small stops
Quality	Scrap, rework, start-up defects, inspection holds	Process control, materials, training and quality standards

Availability

Breakdowns, changeovers, waiting for material or people

Typical action: Maintenance response, setup preparation, scheduling, material readiness

Performance

Slow cycles, minor stops, unstable feed, worn parts

Typical action: Settings, tooling, work methods and repeated small stops

Quality

Scrap, rework, start-up defects, inspection holds

Typical action: Process control, materials, training and quality standards

Give each action one owner and a review date, and keep it small enough to finish. Look for patterns too. If Monday mornings always show more start-up loss, look at weekend shutdowns, warm-up routines or material preparation.

What an OEE report needs

A report earns its place if it makes the next action obvious. Six fields are enough to start:

Context

Date, shift, machine and product.

Time base

Planned production time and run time.

The three factors and OEE

So the calculation can be checked.

Top three losses

Where attention goes first.

Shift notes

What the numbers can't show.

Owner and next action

What happens because of the report.

Shift results drive the fast response. Weekly trends show repeat losses, and monthly views support planning and maintenance priorities.

Common OEE mistakes

The formula is simple. Most problems come from the inputs and from how the number is used:

1
Setting the ideal cycle time too slow. If you use the rate the machine usually runs at instead of its best sustainable rate, performance looks close to 100% and the speed loss disappears. Use the design rate or the best rate seen over a long run.
2
Leaving planned stops out. Some sites take changeovers and planned maintenance out of planned time, which raises availability. They are real losses, and the six-loss model counts setup and adjustment for that reason. Only take out time the machine was never meant to run, such as breaks or no orders.
3
Changing the time base. Mixing planned time, staffed time and machine-available time between shifts or months makes the trend unreliable.
4
Too many reason codes. A long list slows entry and makes operators pick whatever is nearest. Start with six or seven.
5
Measuring every machine at once. A plant-wide rollout produces reports before anyone knows how to use them.
6
Comparing different machines. A press, a packing line and a CNC machine have different products, cycle times and loss patterns. Compare a machine with itself over time.
7
Chasing a benchmark. A single OEE figure from another company says little about yours. The trend and the loss Pareto on your own machine are what guide the next improvement.
8
Chasing the number alone. A higher OEE is not better if it costs safety, quality or maintenance.

A good habit: check the report against the line. If it says breakdowns were the main loss, ask whether operators and maintenance agree.

When to automate OEE tracking

Manual sheets catch long stops well but miss minor stops and slow running. Move to OEE software when those losses matter, or when you track more than a few machines.

The lean manufacturing guide shows how OEE fits with the other shop-floor tools.

Before buying, ask five questions:

Easier for operators? Will it make recording stops and reasons quicker?
Your time rules? Can it handle the way you define planned and unplanned time?
Product rates? Does it support a different ideal rate per product?
Fast enough? Can supervisors see losses while they can still act on them?
For the floor? Will the reports help frontline teams, not just managers?

Don't automate confusion. If the rules are still disputed, software only calculates disputed numbers faster.

Expanding beyond the first machine

Once the pilot is stable, add similar machines first, then more complex areas. Carry forward the definitions, report format, review routine and training from the pilot.

A packing line, a machining center and a batch process may need different stop reasons and rate rules. The formula stays the same. Capture it all in a one-page OEE standard: the calculation rules, what to record, the reason codes, the review rhythm and when a repeat loss is escalated.

Frequently asked questions

How is OEE calculated?

OEE is availability multiplied by performance multiplied by quality. The OEE calculator works it out from run time, output and good parts.

What are the six big losses in OEE?

They are breakdowns, setups and adjustments, minor stops, reduced speed, start-up rejects and production rejects. The first two lower availability, the middle two lower performance and the last two lower quality.

Do you need software to start measuring OEE?

No. A paper sheet with one row per hour at one constraint machine is enough to start. Paper usually misses slow running and short stops, so automate once those losses matter or you track more than a few machines.

Which machine should you measure first?

Start with the constraint, the machine that limits the output of the whole line. Gains there raise output for the whole process.

Choosing OEE Software?

Read the buying guide: what to look for, mistakes to avoid and questions to ask vendors.

Open the guide

Related guides

Sources

Books and papers show their publication year. Web pages show the month we last checked them.

1.Seiichi Nakajima, Introduction to TPM: Total Productive Maintenance (Productivity Press, 1988)
2.Shigeo Shingo, A Revolution in Manufacturing: The SMED System (Productivity Press, 1985)
3.Douglas S. Thomas and Brian A. Weiss, NIST, Economics of Manufacturing Machinery Maintenance: A Survey and Analysis of U.S. Costs and Benefits (NIST AMS 100-34) (nist.gov, checked October 2026)
4.Nicholas Bloom, Erik Brynjolfsson, Lucia Foster, Ron Jarmin, Megha Patnaik, Itay Saporta-Eksten and John Van Reenen, What Drives Differences in Management Practices? (American Economic Review 109(5), May 2019)