OEE & line performance

OEE (overall equipment effectiveness) asks a simple question: of the time a line was scheduled to run, how much turned into good product at full speed? It multiplies three factors:

Take one 8-hour shift (480 minutes) on a filling line rated at 100 bottles an hour, one every 0.6 minutes:

• Availability: 48 minutes were lost to a breakdown and a changeover, so it ran 432 of 480 minutes: 90%.
• Performance: at full speed, 432 minutes would give 720 bottles. It made 648 because of minor stops and slow running: 90%.
• Quality: 616 of the 648 bottles were good, the rest underfilled or mislabelled: 95%.

OEE = 0.90 × 0.90 × 0.95 ≈ 77%. Put simply, the line produced 77% of the good output it could have made in that shift.

The breakdown is more useful than the headline. Low availability points to breakdowns and long changeovers; low performance to small stoppages and slow speeds; low quality to defects and rework. These are often grouped as the “six big losses” of total productive maintenance (TPM). An often-quoted world-class benchmark is about 85%, but a line’s own trend matters more than comparisons across plants.

Don’t confuse OEE with utilization. A line can be busy 95% of the time and still have poor OEE if it runs slowly and makes scrap.

OEE as a time model

OEE is easiest to understand as time slipping away in stages:

• Planned production time: the time the line was scheduled to run.
• Run time: planned time minus stops (breakdowns, changeovers). Availability = run time ÷ planned time.
• Net run time: the time the actual output would take at the ideal rate. Performance = net run time ÷ run time.
• Fully productive time: the time the good output would take at the ideal rate. Quality = good units ÷ total units.

OEE = fully productive time ÷ planned production time, which is the same as multiplying the three factors.

The six big losses

OEE, utilization and TEEP

OEE measures how well the planned time is used. TEEP (total effective equipment performance) also counts the time the line wasn’t planned to run: TEEP = OEE × (planned time ÷ calendar time). If the line runs two shifts out of three, TEEP = 77% × 16/24 ≈ 51%. TEEP shows how much hidden capacity exists before buying new equipment: often a third shift or weekend work before new capital.

The ideal rate decides everything

Performance compares actual output with an ideal cycle time. Choose it badly and OEE means nothing:

• Use the equipment’s true best rate (design speed or the best demonstrated rate), not a slow “standard” rate. With a padded standard, performance can exceed 100% while real losses stay hidden.
• Use the same rule across lines and keep it stable, or trends become meaningless.

How OEE gets inflated

• Excluding stops from planned time: if changeovers, cleaning or meetings are counted as “unplanned time”, availability rises without any improvement.
• Counting reworked units as good: quality looks better while rework consumes capacity.
• Manual data: short stops of a minute or two are rarely written down, so performance losses vanish. Automatic counting on the line usually reveals far more small stops than anyone recorded.

Where to use it

OEE matters most at the bottleneck, where every lost minute is lost plant output. Pushing OEE up on non-bottleneck machines may only produce more work in process. Comparing OEE between plants or companies is risky because definitions differ; a line’s own trend, with a fixed definition, is far more useful. The often-quoted world-class figure of about 85% traces back to Nakajima’s work on TPM, built from roughly 90% availability, 95% performance and 99.9% quality.

The improvement sequence

1. Measure losses reliably, ideally automatically.
2. Attack the largest loss on the bottleneck first: often changeovers (SMED) or breakdowns (maintenance).
3. Standardize the fix, then move to the next loss.
4. Involve operators: autonomous maintenance and quick problem-solving at the line fix many small stops.