1. What is OEE?
OEE (Overall Equipment Effectiveness) is the true level of availability of a plant — true because it combines all losses in a single key figure: availability losses, performance losses and quality losses.
It is the holistic indicator for measuring plant productivity — can be used wherever machine-driven production takes place: filling, machining, food production, interlinked lines.
Calculation example: A plant could produce 1,000 pieces per shift.
- Standstill (repair): −210 pieces
- Speed losses: −90 pieces
- Scrap: −30 pieces
- Result: 670 pieces → OEE = 67%
2. The three OEE factors
Availability level: Proportion of time in which the plant is actually running — measured by the planned production time. Downtimes due to faults, missing parts, missing employees, set-up, cleaning or maintenance all count as a loss, whether planned or unplanned.
Performance level: While the plant is running: How much could it have produced vs. how much did it produce? Causes of losses include poor packaging materials, incorrect adjustments or employee training.
Quality level: Only the good parts of the actual output go into the OEE. Committee is withdrawn.
The total OEE is calculated as: Level of availability × level of efficiency × level of quality
3. Who uses OEE — and for what?
Plant employees: See how the system is running in real time and can immediately initiate deviating measures.
Process engineers: Plan improvement measures, measure their impact (e.g. did a set-up workshop actually increase OEE?).
Management: Reconcile planning and production, allocate financial resources and identify where there are still reserves in the plants.
4. Why is OEE so low in practice?
According to a McKinsey study in the food sector, the average OEE is 62% — meaning that 38% of planned time is lost as a loss. With 100% OEE (theoretically), output could be increased by a further 50% without additional equipment or employees.
Three main reasons for this:
- Missing data: It is unclear when and why the plant was up and on which shift losses occurred. Too little granularity for well-founded analyses.
- Lack of analysis tools: Even when data is available, there is often a lack of methodological support to evaluate it.
- Undirected optimizations: Work is being done on what you know — often preventive maintenance or set-up time. Performance losses are often the biggest lever and are overlooked.
5. Practical implementation with oee.ai
oee.ai (start-up since 2016, located in Aachen) enables OEE recording without having to interfere with plant technology.
Hardware: A fully encapsulated sensor (also waterproof) that collects system data and transmits it to a server in Frankfurt in real time. Installation: approx. 5 minutes
Interference detection: Using a standard tablet, employees enter the reason for the fault from a customizable catalog. The employee's domain knowledge is essential — the system knows that it's there, but not why. Only the employee knows the real reason.
Presentation: Runs in the browser — no additional IT equipment required. Can be displayed on a television, tablet, smartwatch or shop floor board.
6. Evaluations & Advanced Analytics
The interface provides:
- Daily OEE broken down by availability, performance and quality
- Quantity history with marked downtimes and reasons for faults
- Pareto charts for availability, performance, and quality losses — including MTTR (Mean Time to Repair) and MTTF (Mean Time to Failure)
- Heatmaps: focal points of disturbances over weeks and shifts
- Box plots: Deployment fluctuation — frequent short problems vs. rare long failures