OEE: Measuring and Improving Equipment Effectiveness

A practical OEE course for plant managers, supervisors, maintenance leads and engineers in small and mid-sized manufacturers. You will learn to calculate availability, performance and quality correctly, classify the six big losses, avoid the pitfalls that make OEE misleading and put a value on lost time. You will then set up data collection, reason codes and daily reviews, attack breakdowns, changeovers and minor stops, build a business case and run a 90-day OEE programme on your bottleneck.

Course structure

• Module 1: What OEE Measures (4 lessons). OEE and its three factors: availability, performance and quality, with one worked example carried through.
• Module 2: Losses, Benchmarks and Value (4 lessons). The six big losses, utilization and TEEP, benchmarks and pitfalls, and what lost time is worth on and off the bottleneck.
• Module 3: Measuring OEE (4 lessons). Manual data collection, reason codes and Pareto analysis, automatic data capture and daily visual reviews.
• Module 4: Improving OEE (4 lessons). Choosing targets, autonomous and preventive maintenance, SMED quick changeover, and minor stops, speed and start-up losses.
• Module 5: Business Case and Sustaining (4 lessons). Valuing an OEE gain, building a business case with payback, sustaining gains and a 90-day programme.

Interactive labs

• OEE calculator for one shift: Enter one shift of data from a machine or line to calculate availability, performance, quality and OEE.
• Which OEE factor does this loss reduce?: Classify twelve real losses from a packaging line and a machining cell by the OEE factor they reduce, then check your answers.
• What an OEE gain is worth: Enter your current and target OEE, planned hours, ideal rate, contribution margin and the one-time cost to estimate extra units, freed hours, annual value and payback.

How it works

• Work through 20 short lessons at your own pace. Each has key takeaways, an exercise for your own workplace and a quick knowledge check.
• Try the three interactive labs as you reach them in the course.
• Pass the 20-question final exam (70% or higher, unlimited retakes) to unlock your Chmltech certificate of completion.
• Free, with no sign-up. Your progress is saved in this browser.

Overall equipment effectiveness (OEE) answers one question: of the time a machine or line was scheduled to run, what share produced good product at full speed? It multiplies three factors:

OEE = Availability × Performance × Quality

There is a simpler equivalent: OEE = (good units × ideal cycle time) ÷ planned production time. It is the share of scheduled time that would have been needed to make the good output at full speed, with no stops. An OEE of 60% means 40% of scheduled time was lost somewhere.

Why it matters

Many plants believe they are capacity-constrained and plan to buy another machine or add a shift. OEE often shows a large hidden capacity inside the existing equipment: time lost to stops, slow running and scrap. Recovering part of it can deliver more output for far less money than new equipment, and it reveals exactly where the losses are.

Where to use it

OEE suits repetitive equipment with a definable ideal rate: packaging lines, CNC machines, presses, moulding, filling, bottling, extrusion and similar. It is most valuable on the bottleneck, the equipment that limits the whole plant's output, because every hour gained there is an hour gained for the business. The Insights article What OEE really measures covers the concept in more depth.

Availability compares the time the equipment actually ran with the time it was scheduled to run.

Availability = run time ÷ planned production time

Planned production time is the shift time minus planned breaks and periods when no production was scheduled, such as no orders or planned maintenance (more on what to exclude in a later lesson). Run time is planned production time minus stop time.

What counts as a stop

Availability losses are stops long enough to be recorded with a reason. Many plants use a threshold, such as five minutes, with shorter stops counted as minor stops under performance. Typical availability losses:

• Breakdowns and equipment failures.
• Changeovers and setups between products.
• Waiting for materials, operators, quality approval or maintenance.
• Upstream starving or downstream blocking, if you count them here; decide once and be consistent.

Example

An eight-hour shift has a 30-minute meal break, so planned production time is 450 minutes. The line stopped for a 25-minute changeover and a 20-minute jam. Run time is 450 − 45 = 405 minutes. Availability = 405 ÷ 450 = 90%.

Changeovers belong in OEE

Some plants leave changeovers out of OEE because "they are planned". Do not. A changeover is time the equipment could have produced. Counting it keeps changeover reduction visible, and changeover time is often one of the largest and most fixable losses.

Performance compares what the equipment actually produced in its run time with what it could have produced at its ideal cycle time.

Performance = (ideal cycle time × total units) ÷ run time

Total units include good and bad. The ideal cycle time is the fastest the equipment can reliably make one unit of that product under good conditions, from the equipment design rating or the best sustained rate you have demonstrated. Do not use the average or the "standard" rate that already includes losses; that hides the performance gap you want to see.

What performance losses look like

• Minor stops: brief jams, misfeeds, sensor trips and blocked chutes, often under five minutes and rarely recorded. They add up quickly.
• Reduced speed: running below the ideal rate because of material problems, worn equipment, operator caution or habit.

Example

Continuing the example: the line ran 405 minutes. The ideal cycle time is 0.5 minutes per unit (120 units an hour), and it made 700 units. Performance = (0.5 × 700) ÷ 405 = 350 ÷ 405 ≈ 86.4%. About 55 minutes of run time were lost to minor stops and slow running that nobody logged.

A check on your numbers

Performance above 100% means the ideal cycle time is set too slow. Fix it rather than capping the result. If products have different ideal rates, calculate the ideal time for each product's units and add them up.

Quality compares the good units with all the units produced.

Quality = good units ÷ total units

Good means good the first time: units that passed without rework. Reworked units count as bad for OEE even if they are eventually shipped, because the equipment's time to make them the first time was wasted and the rework uses more time and labour.

Quality losses

• Start-up rejects: scrap made while warming up, after a changeover or after a stop, until the process stabilizes.
• Production rejects: scrap and rework during stable running.

Start-up losses are often treated as normal and never measured. On lines with frequent changeovers they can be a large share of all scrap.

Completing the example

Of the 700 units made, 672 were good first time. Quality = 672 ÷ 700 = 96%.

OEE = 90% × 86.4% × 96% ≈ 74.6%. Check with the simple formula: good units × ideal cycle time ÷ planned production time = 672 × 0.5 ÷ 450 = 336 ÷ 450 ≈ 74.7%. The small difference is rounding.

Notice how three factors that each look respectable multiply into an OEE well below any of them. That is why OEE is useful: it shows the combined effect of all losses in one number. Use the lab after this lesson to calculate OEE from one shift of your own data.

The total productive maintenance (TPM) tradition groups equipment losses into six big losses, two for each OEE factor. Classifying every loss into one of the six makes Pareto analysis and improvement focus straightforward.

Turn time into a loss waterfall

Express every loss in minutes of equipment time. Start with planned production time, subtract each loss in turn and finish with fully productive time: the time it would take to make the good units at ideal speed. A waterfall chart of the six losses shows where the time goes at a glance and makes the biggest loss obvious.

Quality losses can be converted to time as rejected units × ideal cycle time; performance losses as run time minus (total units × ideal cycle time).

The lab after this lesson asks you to classify common losses. Getting the categories right matters: a jam logged as a breakdown sends maintenance after the wrong problem.

OEE measures how well equipment uses the time it was scheduled to run. It does not tell you how much of the calendar is scheduled. Two more measures complete the picture.

Example

A line runs two eight-hour shifts, five days a week, with breaks excluded: about 75 planned hours out of 168 in a week, a utilization of about 45%. With OEE of 70%, TEEP is about 31%. The line has two very different sources of extra capacity: improve OEE within the current schedule, or schedule more time.

What to exclude from planned time

Decide once and write it down. Common exclusions are planned breaks, no orders, planned preventive maintenance windows and planned shutdowns. Everything else, including changeovers, waiting for material and unplanned maintenance, stays in planned time as a loss. Be wary of definitions that quietly exclude awkward losses: they raise OEE on paper while hiding the problem.

Which measure for which decision

Use OEE to drive daily improvement on the shop floor. Use TEEP and utilization when deciding whether to add shifts, buy equipment or outsource: those decisions are about total capacity, not just how well scheduled time is used.

An OEE of 85 percent is often quoted as "world class", usually broken down as 90 percent availability, 95 percent performance and 99.9 percent quality. Treat it as a rough reference, not a target for every machine. Many plants that measure OEE honestly for the first time find it well below that, and that is normal.

Compare equipment with itself

OEE depends on product mix, batch sizes, how many changeovers the schedule requires and how ideal cycle times and planned time are defined. A filling line making one product for a week and a CNC cell making forty different parts will never have comparable OEE. The useful comparison is each piece of equipment against its own baseline over time.

Pitfalls that make OEE misleading

• Generous ideal cycle times: using a slow "standard" rate makes performance look high and hides speed losses.
• Excluding losses from planned time: calling changeovers or material shortages "planned" inflates availability.
• Counting rework as good: hides quality losses.
• Averaging OEE across machines: a simple average gives a slow machine the same weight as the bottleneck. Weight by time, or focus on the bottleneck.
• Using OEE to judge operators: people will protect themselves by bending definitions. OEE should point at losses and causes, not at people.
• Chasing OEE on non-bottlenecks: raising OEE on equipment with spare capacity may just build inventory.

Honesty first

A lower but honest OEE that shows where time is lost is far more valuable than a high number nobody trusts. When you introduce OEE, explain that the first figure will probably look low and that the point is to find and remove losses.

To decide which losses to attack and how much to spend, convert lost time into money. The answer depends on one question: is this equipment a bottleneck?

On a bottleneck with sellable demand

If the equipment limits output and customers would buy more, an hour recovered becomes an hour of extra sales. Its value is the contribution margin of the extra units: selling price minus the variable costs of making them (materials, energy, consumables, variable labour and freight). Do not use full revenue, and do not use the full-cost margin that includes fixed overhead already paid for. The Insights article What OEE losses cost works through this in detail.

Example: a bottleneck makes 120 units an hour at ideal rate, and each extra unit sold brings $2.50 of contribution margin. One hour recovered is worth up to 120 × $2.50 = $300. Recovering 400 hours a year is worth up to $120,000 a year, but only if the extra units are sold.

On a non-bottleneck, or without extra demand

Extra output that cannot be sold is just inventory. Here the value of better OEE comes from cost avoidance:

• Fewer overtime hours or a shift no longer needed.
• A machine purchase that can be deferred or avoided.
• Less scrap: material, energy and disposal costs.
• Less expediting and fewer late deliveries.

Scrap is valued differently

A scrapped unit loses its materials and processing cost even on a non-bottleneck. On a bottleneck it also loses the margin of the unit that could have been made in its place. Count both where they apply, but never count the same benefit twice.

You do not need sensors or software to start measuring OEE. A well-designed paper or tablet form, filled in during the shift, is enough for the first months and teaches the team what matters.

What to record each shift

Design the form with operators

Operators know which stops happen and what to call them. Involve them in designing the form and the reason codes, keep it to one page, and make it quick: tick boxes and times, not paragraphs. A form that takes more than a few minutes per shift will be filled in at the end of the shift from memory.

Start on one machine

Begin with the bottleneck. Run the form for two to four weeks before drawing conclusions, so the data covers different products, shifts and conditions. Expect the first week's data to be patchy; review it daily with the operators, fix confusing codes and thank people for recording problems honestly.

Hourly counts reveal hidden losses

Adding a simple hourly output count, written on a board next to the target for that hour, is one of the most effective additions. It shows when performance drops and prompts the question "what happened in that hour?" while the answer is still fresh.

OEE tells you how much time is lost. Reason codes tell you why. Good codes turn a pile of stop records into a Pareto chart that points improvement at the biggest causes.

Designing reason codes

• Two levels: a category (breakdown, changeover, material, quality, minor stop, other) and a specific reason within it (for example breakdown → conveyor motor, sealing jaw, vision system).
• Short list: 10 to 25 codes per machine is usually enough. Too many and people pick the first one; too few and the data says nothing.
• Clear definitions with examples, so two operators would code the same stop the same way.
• Limit "other": if "other" exceeds about 10 percent of stop time, review the comments and add codes.

From codes to action

After two to four weeks, sum stop minutes by code and sort them. Typically a handful of reasons account for most lost time. Then drill down: for the top reason, look at when it happens (product, shift, time since changeover) and ask why using 5 Whys or a fishbone.

In this example the first two reasons account for 58 percent of lost time, so they become the first improvement projects. Count frequency as well as minutes: many short stops of one type may matter more to operators than one long one, and have different causes.

Once manual tracking has shown where the losses are and the team trusts the definitions, automatic data capture can make OEE more accurate and less work.

Options from simple to advanced

Automation catches minor stops

The biggest surprise when automatic capture starts is usually minor stops. Stops of 20 seconds to two minutes are almost never logged by hand but can add up to hours a shift. Seeing them by time and product is often the first big improvement opportunity.

People still provide the why

Sensors know that the machine stopped, not why. Keep operators adding reason codes for stops above the threshold, with a quick touchscreen list. Plan for network access, cybersecurity on connected machines and who maintains the system.

Before buying

Prove the value with manual data first, define what you will do with the information, and pilot on one line. A dashboard nobody acts on is just another screen.

Data only helps if people look at it and act. Simple routines around a visual board turn OEE into daily improvement.

The production board

Place a board at the line showing, for each hour: target output, actual output, the gap and the reason for any gap. Add the shift's OEE, the top three losses and open actions with owners. Anyone walking past can see in seconds whether the line is on track and what is being done.

A tiered daily review

Rules that make it work

• Stand at the board; keep to time.
• Every significant loss gets an owner and a date, or is escalated to the next tier.
• Review actions from the last meeting first.
• Celebrate improvements and thank people for raising problems.

The aim is a fast loop: a loss happens, it is visible within hours, a cause is found and a fix is tried within days. Monthly OEE reports alone are too slow; by the time they are read, the details are forgotten.

With a few weeks of data, pick improvement targets deliberately. Three questions guide the choice.

1. Which equipment? Start with the bottleneck, where recovered time becomes extra output. If there is no clear bottleneck, choose the equipment with the most overtime, late orders or scrap cost.
2. Which loss? Use the six-loss waterfall and reason-code Pareto. Pick the biggest loss that the team can realistically influence.
3. Which method? Match the method to the loss.

Small teams, short cycles

Form a small team of operators, maintenance and an engineer or supervisor for each focused improvement. Give it one loss, a measurable target and four to eight weeks. Results on one loss build the credibility to tackle the next.

Waiting for materials or operators is an OEE loss whose cause usually sits outside the machine, in scheduling, supply or staffing. Raise it through the daily review rather than asking maintenance to fix it.

Breakdowns are rarely random. Most have early signs: loose bolts, leaks, wear, dirt, heat, noise and vibration. Total productive maintenance (TPM) aims to catch these signs early and remove their causes, with operators and maintenance working together.

Autonomous maintenance

Operators take on basic care of their equipment: cleaning, inspecting, lubricating and tightening. Cleaning is inspection: wiping down a machine reveals cracks, leaks and loose parts before they cause failures. Typical steps:

1. Initial clean and inspection, tagging every abnormality.
2. Eliminate sources of contamination and hard-to-clean areas.
3. Set clean, inspect and lubricate standards with simple checklists and visual aids.
4. Train operators to understand the equipment and spot problems.

Maintenance technicians then focus on skilled work instead of routine checks.

Planned and preventive maintenance

Planned maintenance schedules work before failure: time-based (every 500 hours), usage-based (every 100,000 cycles) or condition-based (when vibration, temperature or wear crosses a limit). Use your breakdown history to set priorities: components that fail repeatedly need either a better maintenance task or a design change.

Learn from every significant failure

For each breakdown over a set duration, record the failed component and ask why it failed, not only how it was repaired. Track mean time between failures (MTBF) and mean time to repair (MTTR) for the bottleneck. Rising MTBF shows prevention is working; falling MTTR shows better spares, tools and procedures.

Maintenance and operating work on equipment must follow your lockout and safety procedures; autonomous maintenance tasks should be limited to work operators are trained and authorized to do safely.

Changeovers are often the largest availability loss on high-mix equipment. SMED (single-minute exchange of die), developed by Shigeo Shingo, is a structured way to cut them. The goal of "single minute" means under ten minutes; even halving a long changeover is a large gain.

The SMED steps

1. Observe and record: film a complete changeover and list every task with its time, from the last good part of the old product to the first good part at full speed of the new one.
2. Separate internal and external tasks: internal tasks can only be done with the machine stopped; external tasks can be done while it runs.
3. Move external tasks out: fetch tools, dies, materials and paperwork, pre-heat, and pre-set before the machine stops, using a checklist and a changeover cart.
4. Convert internal to external: pre-assemble parts, use standard heights so tools need no shimming, pre-set settings off-line.
5. Streamline what remains: quick clamps instead of bolts, one-turn fasteners, eliminate adjustments with fixed stops, work in parallel with two people.
6. Standardize: write the new method as standard work and train every shift.

Example

A packaging line's 70-minute changeover was filmed. Twenty-five minutes went to fetching parts and tools, finding the right settings and waiting for first-article approval. A changeover cart, a settings sheet per product and a quicker approval routine cut it to 40 minutes, with no capital spending. Quick clamps on the forming section later brought it to 28 minutes.

Use the time well

Shorter changeovers can be taken as extra running time or as more frequent changeovers with smaller batches, which cuts inventory and improves responsiveness. Decide which matters more for this equipment.

Performance and quality losses are often the least visible and the most persistent. They respond to careful observation and small, specific fixes.

Minor stops

Minor stops are short and frequent: a carton jams, a sensor misreads, product tips over on a conveyor. Each takes seconds to clear, so nobody logs it, but dozens a shift can cost an hour. To attack them:

• Stand at the machine for an hour and tally every stop by location and type.
• Pick the most frequent and ask why five times: worn guide rail, dirty sensor, inconsistent material, wrong setting.
• Fix the cause and set a cleaning or adjustment standard so it stays fixed.

Reduced speed

Ask why the equipment is not run at its ideal rate. Common answers: it jams at higher speed, a particular material runs badly, the operator was told years ago to slow it down, or a worn component limits it. Test speed increases in small steps, with maintenance and quality present, and fix what breaks at each step.

Start-up and changeover scrap

Track scrap for the first period after each start-up or changeover separately. Standard start-up settings per product, pre-heating, checked materials and a clear first-piece approval routine often cut it sharply.

Production rejects

Scrap and rework during stable running are process problems. Use the DMAIC method from the Lean Six Sigma Foundations course: measure, find the verified causes, fix and control.

Before investing in an OEE programme or a specific fix, estimate what it is worth. The method from the earlier lesson on lost time applies to the gain.

Extra output on a bottleneck

Extra good units per year = (target OEE − current OEE) × planned production hours × ideal rate per hour.

Example: a bottleneck runs 4,000 planned hours a year with an ideal rate of 120 units an hour. Raising OEE from 55% to 65% gives 0.10 × 4,000 × 120 = 48,000 extra good units a year. At $2.50 contribution margin per extra unit sold, that is up to $120,000 a year, if the market will take the extra units.

Capacity freed for the same output

If demand is fixed, higher OEE means the same output in fewer hours. Hours needed at the new OEE = current hours × current OEE ÷ target OEE. Here, 4,000 × 55 ÷ 65 ≈ 3,385 hours, so about 615 hours a year are freed. Their value depends on what you do with them: cut overtime, drop a weekend shift, delay a new machine or take on new work.

Other benefits

• Lower scrap: material, energy and disposal savings.
• Less expediting, better on-time delivery and fewer penalties.
• Lower maintenance cost per unit as breakdowns fall.

Be conservative

Count only benefits the business will actually realize, say how each is calculated, and ask finance to agree the method. A believable estimate of $80,000 beats an inflated $300,000 that nobody trusts. The lab after the next lesson calculates extra units, freed hours, value and payback.

Some OEE improvements cost almost nothing: changeover carts, settings sheets, cleaning standards. Others need money: quick-release tooling, new sensors, a monitoring system, a rebuild of worn equipment or a maintenance contractor. For those, build a short business case.

What the business case contains

Payback example

A $40,000 changeover upgrade on a bottleneck is expected to lift OEE from 55% to 65%, worth $120,000 a year in contribution margin. Payback is $40,000 ÷ ($120,000 ÷ 12) = 4 months. Even at half the expected benefit, payback is 8 months, which makes the decision robust. The Insights article OEE project payback walks through a larger worked example.

Funding

Larger productivity projects may qualify for government programmes or financing. Check the Funding page on this site for programmes that support productivity and technology adoption, and confirm eligibility directly with each programme before relying on it.

Close the loop

After implementation, measure the OEE gain and the realized benefit against the business case. Sharing honest results, good or bad, makes the next business case more credible.

OEE gains fade if nothing holds them: settings drift, cleaning standards slip, new operators learn shortcuts and a key person leaves. Sustaining needs structure.

Standards

• Standard work for operation, changeover, start-up and cleaning, with photos, posted at the machine.
• Settings sheets for each product, kept current and controlled.
• Clean, inspect and lubricate checklists with owners and frequencies.

Routines

• Daily tiered reviews of OEE and the top losses, as covered earlier.
• Weekly review of open actions and the loss Pareto.
• Monthly review of trends, MTBF and MTTR and the improvement pipeline.
• Regular audits of standards, with findings closed out.

Skills

Cross-train operators on changeovers and basic care. Train team leads to run the daily review and basic problem solving. Make sure maintenance knowledge is written down, not held by one person.

Watch the definitions

When OEE suddenly improves, check that nothing changed in how it is calculated: a new ideal rate, a new exclusion from planned time or a change in what counts as good. Keep a log of definition changes with dates.

Spread what works

Once the bottleneck is stable at a better level, a new bottleneck may appear elsewhere. Apply the same approach there, reusing the forms, codes and standards that worked.

A focused 90-day programme on one bottleneck is a practical way to prove OEE's value and build the habits to spread it.

Measures to report

• Weekly OEE with availability, performance and quality.
• Minutes lost by top reason.
• Actions opened and closed.
• Output, overtime and scrap cost on the bottleneck.

Signs it is working

Operators raise problems and suggest fixes, maintenance plans more work and fights fewer fires, the daily review is short and practical, and the biggest loss at day 90 is different from the one at day 30.

Related courses

The Operational Excellence course places OEE in a wider improvement programme; the Lean Six Sigma Foundations course gives you DMAIC for stubborn quality losses; and the 5S course builds the clean, organized workplace that autonomous maintenance depends on.