If you work in production, you’ve probably already heard of OEE. Maybe you already calculate it, maybe you see it in monthly reports, or maybe it’s one of your department’s targets. But how often, looking at that number, have you felt it tells you little about what to do to actually improve it?
OEE is one of the most cited KPIs in manufacturing. It’s also one of the most misunderstood—not because it’s complicated, but because a number alone isn’t enough. You need to understand what’s behind it.
In this article, we start with the basics and move to what matters most: how to use real factory data to stop just measuring OEE and start improving it.
What is OEE (Overall Equipment Effectiveness)
OEE stands for Overall Equipment Effectiveness. It is an index that measures how effectively a plant or machine is being used compared to its theoretical maximum potential.
In other words: of all the time that machine could be producing, how much is actually used productively and without defects?
OEE is expressed as a percentage, ranging ideally from 0 to 100%. An OEE of 100% means the machine is always producing, at maximum speed, without ever generating defects. In reality, no plant reaches this, and it’s not necessary to do so. But understanding where you stand relative to this benchmark is the first step toward improvement.
In discrete manufacturing, a world-class OEE is considered to be 85%. The average for factories ranges between 40% and 60%. The gap—often 25–30 percentage points—represents a huge amount of latent production capacity that goes unused.
The OEE formula: three factors, one multiplication
OEE is calculated by multiplying three indices: Availability, Performance, and Quality.
OEE = Availability × Performance × Quality
Each of these three factors captures a different category of production loss. Let’s look at them one by one.
Availability
Availability = Actual operating time ÷ Planned production time
It measures how much time the machine is actually producing compared to how much it should be. Availability losses are caused by unplanned stoppages: breakdowns, waiting for materials, prolonged setups, or unexpected maintenance. A machine that should run for 8 hours but stops for 2 hours due to a breakdown has an availability of 75%.
Performance
Performance = (Units produced ÷ Operating time) ÷ Ideal speed
It measures whether the machine, when running, operates at the speed it was designed for. Performance losses are caused by micro-stoppages (short stops that accumulate), reduced cycle speed, or process inefficiencies. A machine that could produce 100 units per hour but produces 80 has a performance of 80%.
Quality
Quality = Conforming units ÷ Total units produced
It measures how many of the produced units actually meet the specifications on the first pass (first pass yield). Quality losses include scrap, rework, and defective production during startup. A machine that produces 100 units but rejects 5 has a quality of 95%.
A concrete example
Let’s take a machine with 80% availability, 90% performance, and 95% quality:
OEE = 0,80 × 0,90 × 0,95 = 68,4%
Three values that all seem acceptable individually produce an OEE that is nearly 17 points below the world-class level. This is the multiplicative effect of OEE: inefficiencies amplify one another. This is why the index is so useful and why improving just one factor is not enough.
The Six Big Losses: Where Efficiency Goes
Behind the three OEE factors lie what the TPM (Total Productive Maintenance) methodology calls the Six Big Losses. Understanding them is essential because every improvement effort must begin by identifying which specific loss needs to be reduced.
- Breakdowns (Impact: Availability) — unplanned stops due to equipment failure or malfunctions.
- Setup and Adjustments (Impact: Availability) — production changeover times, tooling, and warm-up periods.
- Minor Stoppages (Impact: Performance) — short and frequent stops, often not recorded manually.
- Reduced Speed (Impact: Performance) — equipment running, but below the ideal cycle time.
- Startup Rejects (Impact: Quality) — non-compliant parts produced during the startup or production changeover phase.
- Production Rejects (Impact: Quality) — scrap and rework occurring during steady-state production.
Most factories that measure OEE manually fail to distinguish between minor stoppages and reduced speed. Both impact performance, but the causes and solutions are completely different.
The problem with manually calculated OEE
Many manufacturing companies still calculate OEE by collecting data manually: operators note down stops on a sheet, the department head enters them into Excel at the end of the shift, and someone aggregates them into a weekly report. The result often arrives the next day or on the Monday of the following week.
This approach has three structural limitations that make the data nearly unusable for improvement:
- Latency: by the time the data is available, the problem that generated it is already in the past. It is impossible to intervene in real time.
- Insufficient granularity: minor stoppages, often the most significant loss category, are almost never recorded because the operator does not note them down (they last 30 seconds, then the machine restarts).
- Subjectivity: the categorization of causes depends on who fills out the form. The same stoppage is classified differently by different operators.
The result? An OEE that measures but does not explain. It is often used more as a performance indicator in executive reports than as an operational lever for improvement.
How to improve OEE with real-time factory data
The shift in quality occurs when OEE is calculated automatically, using data generated directly by the machines. Not from the operator taking notes, but from the system detecting the data.
Would you like me to translate any further sections of this text on OEE?
This changes everything. Not only the accuracy of the data, but the use that can be made of it.
Real-time data: from historical OEE to live OEE
With machine interconnection, every cycle, every stoppage, and every speed variation is recorded automatically. OEE is no longer a number from the day before: it is a live indicator, visible on operational dashboards, allowing department heads to intervene during the shift.
From average to detail: OEE by machine, shift, operator.
When data is collected automatically, it is possible to break down OEE at a granular level: not just the average OEE of the department, but the OEE of every single machine, every shift, and every product type. This granularity is what transforms the KPI into a diagnostic tool.
Example: if the OEE of the night shift is systematically 8 points lower than the day shift, there is something to investigate (operator training, maintenance conditions, different setups). An aggregated OEE would never reveal this.
Would you like me to translate the next section or focus on a specific OEE component?
Analysis of causes: from losses to actions.
The system that automatically records downtime can also categorize and aggregate it with the respective causes. The result is immediate: 20% of the causes generate 80% of the losses. You intervene on those specifically, rather than on everything indiscriminately.
The role of the MES
An MES software with a machine monitoring module is the tool that enables this level of analysis. Integrated with PLCs and machine control systems, it automatically collects cycle data, calculates OEE in real time, and provides production managers with the information needed to make decisions.
It is not a solution reserved for large plants. Today, thanks to IoT gateways and standard connectors, it is possible to connect even legacy machines without replacing them, and start measuring OEE automatically in a relatively short time.
From 55% to 80% OEE: what happens in between
Raising OEE from 55% to 80% is not a one-off operation. It is a journey that typically develops in phases, each of which requires data, analysis, and targeted interventions.
Phase 1 — Measure accurately. Before taking action, you need reliable data. 70% of companies starting an OEE improvement project discover that the actual value is significantly different from what was previously thought. Often, it is lower.
Phase 2 — Identify dominant losses. With automatic data and Pareto analysis, the 2–3 causes that alone explain most of the gap are identified. They are almost always different from what was assumed.
Phase 3 — Targeted intervention. Reducing setup times using the SMED methodology, improving preventive maintenance on the machines with the most failures, and standardizing startup procedures to reduce initial scrap. Every intervention is data-driven, not based on intuition.
Phase 4 — Monitor and Stabilize. Improving OEE is not a one-time project. It is a continuous cycle: measure, analyze, intervene, verify. The digital system that made measurement possible also becomes the tool for monitoring the stability of results over time.
An increase of 10 OEE points on a production line often equates to an additional production capacity of 15–20%, without investing in new machinery. It is the most immediate and least expensive lever for increasing productivity.