OEE Explained: What It Is and How to Improve It

Your machine ran for 8 hours. You made parts for 8 hours. So utilization is 100%, right?

Not quite.

That machine spent 45 minutes on setups. It ran 10% slower than rated speed because "that's just how we run it." And 3% of the parts needed rework. Your 100% suddenly looks more like 60%.

This is why Overall Equipment Effectiveness (OEE) exists. It's a single metric that captures what's really happening with your equipment: how often it runs, how fast it runs, and how many good parts it produces.

Most manufacturers guess at efficiency. OEE forces you to measure it. And the results are often uncomfortable.

This article explains what OEE actually measures, how to calculate it, and practical steps to improve it. If you want to run your own numbers as you read, try our free OEE calculator.

What OEE Actually Measures

OEE = Availability x Performance x Quality

That's it. Three percentages multiplied together. The result tells you what percentage of planned production time resulted in good parts at full speed.

Here's why the multiplication matters.

Say your machine is available 90% of the time. It runs at 95% of ideal speed. And 99% of parts are good. You might think you're at 95% efficiency. Take the average, right?

Wrong. OEE is 90% x 95% x 99% = 84.6%.

Small losses compound. A machine that looks "pretty good" on each individual metric can have significant hidden losses when you combine them.

This is the power of OEE. It exposes losses you didn't know you had. The "Six Big Losses" framework categorizes these into availability losses (downtime, setups), performance losses (slow cycles, minor stops), and quality losses (defects, rework).

But before we get into improvement, you need to understand what each component measures.

The Three Components of OEE

Availability: Are You Running?

Availability answers a simple question: Of the time you planned to produce, how much did you actually run?

Formula: Run Time / Planned Production Time

Planned production time is your scheduled shift minus any planned downtime (lunch breaks, scheduled maintenance). Run time is what's left after unplanned downtime.

What counts as an availability loss:

Breakdowns and equipment failures
Setup and changeover time
Material shortages that stop production
Waiting for operators, tools, or information

The common killer here is untracked downtime. The machine was down for 20 minutes, but nobody logged it. It "just happens." Those 20-minute gaps across a shift add up to hours of lost production that never appears on any report.

Performance: How Fast?

Performance measures whether your equipment runs at its ideal speed when it's running.

Formula: (Ideal Cycle Time x Total Count) / Run Time

Ideal cycle time is the fastest possible time to produce one part. Total count is everything produced during run time. If your machine can theoretically produce 60 parts per hour but only produced 50, your performance is 83%.

What counts as a performance loss:

Running slower than rated speed
Minor stoppages (jams, idling, small adjustments)
Operator inefficiency or process variation

The common killer here is the "that's just how we run it" mindset. The machine is rated for 1,000 RPM but everyone runs it at 800 because "it works better." Maybe it does. Maybe you're leaving 20% of capacity on the table. OEE makes you ask the question.

Quality: Are They Good?

Quality measures how many parts meet specifications on the first pass.

Formula: Good Count / Total Count

This is the most straightforward component. You made 100 parts. 97 were good. Quality is 97%.

What counts as a quality loss:

Scrap that gets thrown away
Defects caught during inspection
Parts that need rework to meet spec
Startup rejects while dialing in the process

The common killer here is not counting rework. A part needed 10 extra minutes of hand finishing, but it shipped, so it's "good." OEE should capture that. The part wasn't right the first time. That's a quality loss.

How to Calculate OEE (With Example)

Ready to calculate your own OEE? Use our free OEE calculator to plug in your numbers. Or follow along with this example.

The scenario: An 8-hour shift on a CNC machine.

Step 1: Calculate Availability

Planned production time: 480 minutes (8 hours)
Planned breaks: 30 minutes
Available time: 450 minutes
Downtime events: 45 minutes (setup, breakdown, material wait)
Actual run time: 405 minutes

Availability = 405 / 450 = 90%

Step 2: Calculate Performance

Ideal cycle time: 2 minutes per part
Total parts produced: 180 parts
Run time: 405 minutes

Performance = (2 x 180) / 405 = 360 / 405 = 88.9%

Step 3: Calculate Quality

Total parts: 180
Good parts: 171 (9 scrapped or reworked)

Quality = 171 / 180 = 95%

Step 4: Calculate OEE

OEE = 90% x 88.9% x 95% = 76%

This was a "decent" shift. The machine ran most of the day. It made parts at a reasonable pace. Scrap was under 5%. And yet, nearly a quarter of potential production was lost.

That's 108 minutes of an 8-hour shift that produced nothing of value. Multiply that by 250 working days per year. That's 450 hours. Now multiply by your machine's hourly rate.

The math gets uncomfortable quickly.

What's a Good OEE Score?

The "world-class" benchmark is 85% OEE, typically achieved with:

90% Availability
95% Performance
99% Quality

For context, typical manufacturing runs around 60% OEE. Most shops that have never measured OEE find themselves between 40% and 65%.

But here's the thing: don't obsess over the absolute number.

OEE is most valuable as a trend over time. If you're at 55% today and 62% next quarter, you're improving. That matters more than whether you've hit some arbitrary benchmark.

Also, comparing OEE across different processes is misleading. A high-volume production line making the same part all day will have different OEE dynamics than a job shop running 50 different parts per week. The setups alone will crater availability.

Use OEE to track improvement on the same machine over time. Use it to identify where losses are hiding. Don't use it to compare your shop to a factory with completely different operations.

How to Improve OEE (Practical Steps)

OEE tells you where you're losing time. Improving it means attacking those losses systematically.

Start with the biggest loss. This is the Pareto principle in action. Usually, 20% of the loss categories cause 80% of the problem. Fix those first.

Improving Availability

Availability losses come from downtime. Fix downtime.

Faster changeovers: If setups take 45 minutes, ask why. Can you pre-stage tools? Standardize the process? SMED (Single-Minute Exchange of Die) methodology can cut changeover times dramatically.

Better maintenance: Unplanned breakdowns kill availability. Preventive maintenance schedules catch problems before they stop production. IoT-connected machines can even predict failures before they happen.

Material staging: Machines waiting for material is availability loss. Ensure the next job's material is ready before the current job finishes.

Improving Performance

Performance losses come from running slower than possible.

Know your ideal cycle time: Many shops don't actually know the theoretical maximum speed. Find it. Document it. Compare actual to ideal.

Track minor stops: A 30-second jam doesn't feel like a problem. When it happens 20 times per shift, that's 10 minutes of performance loss. Track it.

Question the "we've always done it this way" speeds: Maybe there's a good reason to run slower. Maybe there isn't. OEE forces the conversation.

Improving Quality

Quality losses come from defects and rework.

First-pass yield focus: Don't just track scrap. Track parts that needed extra work to be acceptable. That's hidden quality loss.

Root cause analysis: When defects happen, ask why five times. Fix the root cause, not just the symptom.

Process capability: Is the process capable of making good parts consistently, or are you fighting variation every day?

Manual vs. Automated OEE Tracking

You can track OEE with a clipboard and a spreadsheet. You can also track it with IoT sensors and real-time software. Both work. Neither is perfect.

Manual Tracking

How it works: Operators log downtime events, part counts, and scrap on paper or in a spreadsheet. Someone calculates OEE periodically.

Pros:

No technology investment
Start immediately
Forces operators to think about losses

Cons:

Data accuracy depends on operator discipline
Easy to forget or fudge numbers
Insights come days later, not in real time

Manual tracking is where most shops should start. It costs nothing and proves whether OEE matters to you before investing in automation.

Automated Tracking

How it works: Sensors on machines detect run/stop status, cycle counts, and sometimes quality events. Software calculates OEE automatically.

Pros:

Real-time data without operator input
More accurate than human memory
Historical trends for analysis
Dashboards that show what's happening now

Cons:

Initial investment in sensors and software
Connectivity challenges with older machines
Still needs operators to log downtime reasons

The best approach is often a hybrid. Automate what machines can tell you (run time, cycle counts). Have operators log what machines can't (why it stopped, quality issues).

When OEE Matters (And When It Doesn't)

Here's an honest take: you don't need OEE on every machine.

OEE matters when:

The machine is a bottleneck constraining your output
Equipment is expensive and utilization directly impacts ROI
You're running close to capacity and need to find more
You're trying to justify capital investment in new equipment

OEE might be overkill when:

Every machine has ample capacity
You're not constrained by equipment
The overhead of tracking exceeds the value of the data

A job shop with 10 machines, all running at 50% utilization because demand is low, doesn't need OEE on all 10. That's not where the business problem is.

But if one machine is the constraint, and everything waits on it, then OEE on that machine is critical. Every percentage point of improvement translates directly to throughput.

Start with your constraints. Expand tracking as it proves valuable.

The Bottom Line

OEE combines availability, performance, and quality into a single metric that shows how effectively your equipment produces good parts.

The formula is simple: Availability x Performance x Quality = OEE.

The insight is powerful: small losses in each area compound into significant hidden waste. A machine that seems fine might be losing 30-40% of its potential.

Most shops that measure OEE for the first time are surprised. The numbers are lower than expected. But that's the point. You can't improve what you don't measure.

Start simple. Calculate OEE for your most critical machine. Use our OEE calculator to run the numbers right now. Track it weekly. Identify the biggest loss category. Fix it. Repeat.

OEE without action is just math. OEE with consistent improvement effort is how you find capacity you didn't know you had.

Ready to track OEE automatically?

Workcell calculates OEE in real time with IoT integration. Connect your machines, see availability and performance automatically, and track quality through our production module. No more spreadsheets. No more guessing.

Book a demo and we'll show you what OEE tracking looks like with your actual equipment.