What Is OEE (Overall Equipment Effectiveness)?
OEE (Overall Equipment Effectiveness) measures how much of your planned production time is genuinely productive: producing good parts, at full speed. A perfect 100% score means every scheduled minute made a good part at the machine’s design rate.
Manufacturers use OEE because it catches lost capacity that downtime logs miss on their own, including slow cycles, minor stops, and scrap. Published benchmarks put the typical discrete manufacturer at about 55% to 65% OEE1, and most manufacturers score closer to 60%2. World-class discrete manufacturing is widely cited at 85%.
Enter your own shift numbers and see your score in seconds.
Proven. Trusted. Certified. Awarded.
The OEE Formula
A shorter version reaches the same score without isolating which factor caused the loss:
OEE = (Good Count × Ideal Cycle Time) ÷ Planned Production Time.3
The Compound Effect
The three factors are multiplied, not averaged, and that is what makes OEE unforgiving. A machine holding a strong 90% on all three factors does not score 90% overall. It scores 72.9% (90% × 90% × 90%), because three small losses compound into one large one. Averaging the same three numbers would still show 90% and hide the loss entirely.
Availability, Performance, and Quality
Availability
Availability measures how much of the scheduled time the machine actually ran. Planned breaks come out of planned production time before the clock starts, so they don’t count against the machine. Unplanned downtime does.
A press brake scheduled for 420 minutes loses 48 minutes to a machine fault and a tooling delay, leaving 372 minutes of run time.
Performance
Performance measures how fast the machine ran against its ideal cycle time. It catches slow cycles and short stops that never show up as recorded downtime.
A welding cell with a 90-second ideal cycle runs for 372 minutes. At its design rate, it could produce 248 parts; it produces 226 instead.
A 40-second jam may never reach a downtime log, but it still costs the shift a part.
Quality
Quality measures how much output meets specifications without scrap or rework. A reworked part still counts against quality, since the machine didn’t make it right the first time.
That same shift produces 226 parts, with nine sent to rework, leaving 217 good parts.
A Worked OEE Calculation
Here is one shift on a CNC (computer numerical control) machining center, taken through the full arithmetic. The shift runs 480 minutes. After 45 minutes of planned breaks, 435 minutes remain as planned production time. The machine loses 52 minutes to unplanned stops (a tool change overrun and a spindle fault), leaving 383 minutes of run time. At a 2.5-minute ideal cycle, it produces 138 parts, six of which are scrapped or reworked, leaving 132 good parts.
| Step | Calculation | Result |
|---|---|---|
| Shift length | 8 hours | 480 minutes |
| Planned breaks | Scheduled into the shift | 45 minutes |
| Planned production time | 480 − 45 | 435 minutes |
| Unplanned stops | Tool change overrun + spindle fault | 52 minutes |
| Run time | 435 − 52 | 383 minutes |
| Availability | 383 ÷ 435 | 88.0% |
| Ideal cycle time | Machine’s design rate | 2.5 min/part |
| Total count | Machine counter | 138 parts |
| Performance | (2.5 × 138) ÷ 383 | 90.1% |
| Scrapped or reworked | Machine counter | 6 parts |
| Good count | 138 − 6 | 132 parts |
| Quality | 132 ÷ 138 | 95.7% |
| OEE | 88.0% × 90.1% × 95.7% | 75.9% |
At the ideal cycle time, the 435-minute planned production time could have yielded 174 parts. The shift produced 138, a shortfall of 36 parts before quality even enters the picture. The 52 minutes of unplanned downtime accounts for about 21 of those (52 ÷ 2.5); slower cycles account for the remaining 15. Six more of the 138 parts produced needed scrap or rework, leaving 132 good parts.
We’d flag Availability first in this example, since downtime is the largest source of lost parts here. That won’t hold for every shift. Run your own numbers to find which factor is costing you the most.
See your Availability, Performance, and Quality broken out, not just one final number.
What Is a Good OEE Score?
85% is the widely cited world-class reference for discrete manufacturing (OEE.com), a direction of travel rather than a universal target. Only the top 5% to 10% of plants sustain 80% to 85%1, and the right target for a given plant depends on its industry, product mix, changeover frequency, and its own baseline.
In practical terms, a good OEE score is one measured consistently, benchmarked against the right comparison, and improving over time.
| Reference point | Value |
|---|---|
| Perfect production | 100% |
| World-class OEE, discrete (OEE.com) | 85% |
| Standard factor breakdown behind 85%2 | 90% Availability / 95% Performance / 99% Quality |
| Sustained 80% to 85% | Top 5% to 10% of plants |
| Advanced | 70% to 80% |
| Typical discrete manufacturer | 55% to 65% |
| Below average | 40% to 55% |
| First automatic measurement | Often under 40% |
Published sources give these reference ranges. Most discrete manufacturers score 55% to 70% with automatic measurement, 70% to 80% is advanced, and 80% to 85% is sustained by the top 5% to 10% of plants1. An analysis of 3,500+ machines on one monitoring platform found OEE scores clustering between 55% and 60%, with about 6% of organizations reaching 85%, and the analysts estimate the true share is closer to 3%4.
How OEE Varies by Industry
OEE varies by industry because product mix, changeover frequency, and automation level differ from plant to plant. Two identical lines show the effect: one that makes a single product and one that makes ten will not score the same, because the second loses far more time to changeovers2. A pharma packaging line at 70% may outperform an automotive press at 80%, since validated processes and mandatory cleaning cycles shape what each line can reach1. The ranges below show how the typical score, the target, and the main loss driver shift by industry1.
| Industry | Typical OEE | World-class target | Main loss driver |
|---|---|---|---|
| Automotive (stamping, welding, assembly) | 65% to 80% | 85% or higher | Setup time, tool wear, variant complexity |
| Metal processing (CNC, forming) | 50% to 70% | 75% or higher | High-mix, low-volume work and setup time |
| Electronics and assembly | 55% to 75% | 80% or higher | Manual intervention, test station bottlenecks |
The Six Big Losses
Every point lost from OEE traces back to one of six categories, each tied to a specific factor.5 Measure all six and prioritize whichever is largest in your own operation; no single one is always the biggest.
| Loss | Factor | On the Floor |
|---|---|---|
| Equipment failure | Availability | An unplanned breakdown |
| Setup and adjustments | Availability | Changeovers and tooling swaps |
| Idling and minor stops | Performance | Short jams that never reach a log |
| Reduced speed | Performance | Cycles running below design rate |
| Startup and process defects | Quality | Scrap during startup |
| Production defects | Quality | Scrap during a stable run |
In our experience, the two Performance losses are often a bigger surprise than equipment failure. A machine that never breaks down can still be quietly losing more output to short stops than the one everyone worries about.
How to Improve OEE
Improvement follows a sequence, and skipping ahead rarely works.
Measure accurately first.
Agree on planned production time and a validated ideal cycle time, then capture run time, stop events, total count, good count, and reason codes consistently across shifts. In our experience, the first automated measurement often reads lower than the manual one, because a clipboard catches breakdowns but misses the 90-second jams. That drop is normal and expected. The plant is seeing its real baseline for the first time.
Find the dominant loss.
Rank Availability, Performance, and Quality using the Six Big Losses framework to see which one is costing the most.
Attack one constraint at a time.
Fixing everything at once makes it impossible to tell what worked.
Re-measure using the same rules.
Changing the measurement method mid-project produces a number that can look like progress but proves nothing.
Published case studies show how much OEE can change in a single plant, line, or machine:
A Pakistani auto manufacturing plant raised OEE from 33% to 48% over six months in 2022.6
One Spanish injection-molding machine averaged about 64% OEE from January to September 2025 and about 89% from January to April 2026.7
A Jordanian plastics line study reports OEE rising from 47.3% to 86% after its recommendations.8
Each result reflects one site, so treat these as examples.
OEE Software and Continuous Measurement
A spreadsheet can calculate OEE from production data, but by itself, it cannot automatically capture what is happening on the factory floor. Manual tracking on a clipboard tends to miss the short stops and minor speed losses that make up a real share of the loss picture.
A spreadsheet fed by accurate manual logs may be enough for a stable line with simple changeovers. In our experience, this call usually comes down to needing:
- Machine-level data broken out by product and shift
- Short stops that a manual log won’t show
- Same-day numbers instead of next-morning ones
Before buying software, ask a few direct questions:
- Can it read your machines without a retrofit you can’t justify?
- Does it count parts in a way that identifies which machine lost them?
- Can operators log reason codes as they happen?
- Can a supervisor see a shift’s numbers live?
- Does it integrate with your ERP system?
We built Prophecy IoT® because we kept installing CloudSuite Industrial for plants that had no way to answer those questions in the moment. An operator sees a stop reason on the screen as it happens, instead of a supervisor finding out at the end of the shift what went wrong and when.
Prophecy IoT® gives manufacturers real-time visibility into what is happening on the factory floor by capturing production and machine data as it happens. It helps teams identify downtime, track stop reasons, monitor OEE and production performance, and uncover the small stops and speed losses that manual tracking misses.
By connecting factory-floor activity with Infor CloudSuite Industrial (SyteLine), Prophecy turns real-time production data into information that operators and supervisors can use to see where time and capacity are lost and respond while the shift is still running.
The OEE Research Series
This page covers the fundamentals. Each of these goes deeper into one part of the picture:
Availability, Performance, and Quality: Which Factor Costs Manufacturers the Most?
A factor-by-factor look at where production lines lose the most OEE, drawn from published case studies.
The Six Big Losses in OEE
What each of the six losses costs manufacturers, with 2026 data.
OEE Improvement Benchmarks
How much manufacturers actually gained across ten published before-and-after cases.
OEE Benchmark by Industry
How typical OEE scores compare across manufacturing sectors.
See How Your Plant Is Really Running
We’ve partnered with discrete manufacturers since 1984, and one thing we’ve learned is that industry benchmarks can only take you so far. Your own production data tells the real story. Calculate your OEE to uncover where production time is being lost and where your greatest opportunity for improvement may be.
Frequently Asked Questions
Overall Equipment Effectiveness: a metric combining Availability, Performance, and Quality into one score that shows how much planned production time produced good parts at full speed.
OEE equals Availability multiplied by Performance multiplied by Quality. A shorter version, Good Count times Ideal Cycle Time, divided by Planned Production Time, reaches the same score without showing which factor caused the loss.
A machine or line is producing at the widely cited world-class reference for discrete manufacturing, with minimal loss from downtime, slow cycles, or scrap. Treat it as a benchmark to work toward, not a target every plant should expect to reach.
It depends on the industry and the plant’s own baseline. Most discrete manufacturers score between 55% and 70%, and 70% to 80% counts as advanced1, so 70% is a solid score. Whether that 70% is measured consistently and improving over time counts for more than the number itself.
Not under a correctly measured formula. A score above 100% almost always points to an input error, most often an ideal cycle time set too slow for what the machine can actually do.
No. Planned breaks and scheduled maintenance come out before planned production time is set. Only unplanned downtime during scheduled run time reduces the score.
All four measure equipment performance against a different amount of time.
- Utilization measures how much available time the equipment ran, without a Performance or Quality component.
- OOE (Overall Operations Effectiveness) applies the same three-factor OEE math, measured against all available time.
- TEEP (Total Effective Equipment Performance) applies that same math across all 168 hours in a week, which makes it a measure of unused capacity rather than day-to-day performance.
- OEE measures against planned production time, which is why it’s the metric most plants use to judge how well they ran the time they intended to run.
Continuously, if the goal is to catch and correct losses while a shift is still running. At minimum, calculate it every shift; a monthly rollup can obscure which specific shift, machine, or product is driving the loss.
Yes, and it earns its keep when paired with other numbers. “OEE went from 68% to 71%” gives a plant manager little to act on. “Changeover time dropped 14 minutes per job” gives the team something repeatable. OEE works best paired with throughput, cycle time, first-pass yield, and on-time delivery; see manufacturing KPIs for how it fits with the rest.
