Overall Equipment Effectiveness is three fractions multiplied together. Availability times performance times quality. It is deliberately simple, and that simplicity is both why it spread through manufacturing and why so many plants report a number that means nothing.
- 85% world-class OEE benchmark1
- 60% typical for plants new to measuring it1
- $400k/hr average U.S. unplanned downtime cost2
The 85% benchmark is not arbitrary. Seiichi Nakajima, who originated Total Productive Maintenance, set it in his 1984 book as the minimum a plant should aim for, built from 90% availability, 95% performance and 99.9% quality. Plants that have never measured OEE commonly land nearer 60%.
The Three Terms
- Availability: the time the machine actually ran, divided by the time it was scheduled to run. Breakdowns and changeovers live here.
- Performance: how fast it ran when it was running, against its ideal cycle time. Small stops and slow running live here.
- Quality: good parts divided by total parts. Scrap and rework live here.
Multiply the three and you get a single percentage. An OEE of 85% is generally considered strong for a discrete manufacturer. Many plants that have never measured it land somewhere around 60% and are genuinely surprised.
The Ways the Number Gets Gamed
Because OEE is a product of three fractions, it is trivially inflated by shrinking the denominator of the first one. If planned downtime is excluded generously enough, availability approaches 100% and the metric stops describing reality.
The most common version of this is treating changeovers as planned and therefore excluded. A plant that changes over eleven times a shift and excludes all of it can report excellent OEE while running badly.
The second common distortion is the ideal cycle time. If it is set to what the machine achieves on a good day rather than what it was designed to achieve, performance flatters itself permanently.
What the Number Is Actually For
OEE is poor for comparing two different machines, and worse for comparing two plants. Its value is entirely in the trend of one asset against itself, and in which of the three terms is dragging.
That decomposition is the useful part. An availability problem is a maintenance and changeover problem. A performance problem is usually small stops, which are chronically underreported because nobody logs a ninety-second clear. A quality problem is a process problem. The same headline number can mean three completely different things.
Where the Parts Supply Chain Shows Up
Availability is where a parts strategy becomes visible. A wear part with a known service life that arrives before the shutdown is planned downtime. The same part sourced after it fails is a breakdown, and a breakdown on a machine whose builder is in Europe can be a multi-week hole in the availability number.
This is the argument for knowing your consumables and their lead times. Press pads, filter elements, seals, contact rollers and cutting tooling all have predictable lives. Ordering them on a schedule rather than an emergency is the cheapest availability improvement most plants have available.