OEE (Overall Equipment Effectiveness) is the most common production metric of how effectively equipment uses planned production time. It yields a single percentage, but its real value is in showing where the loss is.
The formula
OEE = Availability × Performance × Quality
| Component | Formula | What it reflects |
|---|---|---|
| Availability | Run time / Planned production time | Downtime losses (breakdowns, setup, waiting for material) |
| Performance | (Ideal cycle time × Total count) / Run time | Speed losses (slow running, minor stops) |
| Quality | Good count / Total count | Quality losses (defects, rework) |
A worked example
The numbers below are made up purely to show the calculation:
- Planned production time: 480 min
- Total unplanned downtime: 60 min → run time 420 min
- Ideal cycle time: 1 min/part
- Total produced: 380 parts, of which 361 are defect-free
- Availability = 420 / 480 = 87.5%
- Performance = (1 × 380) / 420 = 90.5%
- Quality = 361 / 380 = 95.0%
- OEE = 0.875 × 0.905 × 0.95 ≈ 75.2%
In this example the biggest losses are in downtime (12.5%) and speed (9.5%); quality loss is the smallest. Improvement effort should therefore go first to downtime reasons. To try it with your own shift data, use the OEE Calculator in the IT Hub.
The six big losses
| Category | Loss |
|---|---|
| Availability | Breakdowns |
| Setup and changeover | |
| Performance | Minor stops and idling |
| Reduced speed | |
| Quality | Production defects and rework |
| Startup/warm-up losses |
The most common mistakes in measuring OEE
- Setting the ideal cycle time inconsistently. Performance is calculated against the ideal time; “ideal” is sometimes the machine’s design speed and sometimes the best historical value. Which one is used must be clear and the same for everyone.
- Leaving the classification of planned stops unclear. Whether stops such as breaks, planned maintenance and product changeovers count in the OEE calculation must be written down; otherwise comparison between lines becomes meaningless.
- Relying on manually collected data. Manually entered downtime reasons are often incomplete and late. State and counter data collected automatically from the machine is more reliable.
- Treating 100% as the target. OEE is a tool for making losses visible. Targets should be set realistically by product and process; a single frequently quoted “world-class” figure does not apply to every process.
- Comparing different machines directly. OEE values of different products and processes may not be on the same scale; the trend of the same machine over time is often more instructive.
- Looking only at the single number and ignoring the components. Two lines can have the same OEE and suffer entirely different losses.
Data for reliable OEE
The OEE calculation rests on three basic data items: machine state (running/stopped and the reason), the production counter (total and good count) and the planned time. These need to be collected automatically from the PLC/SCADA, time-stamped and with context (line, product, shift), which is a job for data modeling and contextualization. So the calculation is the same in every dashboard, it is recommended to define it once in the semantic layer.
The analytics side
In a platform such as GoodData Analytics, OEE can be followed with breakdowns by line, shift, product and downtime reason; trends and loss distributions can be shared on dashboards. For our approach see the Industrial Analytics & AI page. As downtime data accumulates, it also becomes possible to move to advanced scenarios such as predictive maintenance.
