OEE (Overall Equipment Effectiveness)

OEE (Overall Equipment Effectiveness) is the standard manufacturing metric for answering a simple question: of all the time a machine could be producing good parts, what percentage does it actually use? It is expressed as a single number between 0% and 100%, and it was born within total productive maintenance (TPM) in Japanese industry in the 1970s.

Its power lies in the fact that it does not look at just one thing, but at three at once. OEE multiplies availability (was the machine running when it should have been?), performance (did it run at the speed it should have?) and quality (did the parts come out right the first time?). That is why equipment that looks "very busy" can have a low OEE: losing speed or generating scrap is enough to make the number drop. In the Argentine and Latin American industrial context, OEE is the gauge that drives plant decisions, and it is part of what gets measured and acted on when a Manufacturing operation wants to produce more without buying new machines.

How OEE is calculated

OEE is the product of three factors, each also expressed as a percentage. The formula is OEE = Availability × Performance × Quality. It is worth looking at them one by one, because each factor "catches" a different type of loss.

• Availability = Run time / Planned production time. It captures unplanned stops (breakdowns, failures, long setups) and any planned stops you choose to include. If a line was supposed to run for 8 hours and was down for 1, its availability is 87.5%.
• Performance = (Parts produced × Ideal cycle time) / Run time. It captures minor stops and reduced speed, that is, when the machine is running but below its rated speed. It is the most underestimated factor.
• Quality = Good parts / Total parts produced. It captures scrap and rework: parts that had to be discarded or reprocessed do not count as valid production.

A numerical example makes it all clear. Take a packaging line at a consumer goods company in Greater Buenos Aires: availability 90%, performance 95% and quality 99%. OEE is 0.90 × 0.95 × 0.99 = 84.6%. It looks high, but the catch is that the percentages multiply: three factors that are "good" on their own give a result that shows how much real margin is still left on the table.

Why it matters and what counts as "good"

OEE matters because it turns productivity into a single number you can compare over time and across lines. Without it, plants argue over impressions ("machine 3 is acting up") instead of data. As a reference, an OEE of 85% is often cited as world class: it is the target Seiichi Nakajima set when he introduced TPM in Japan in the 1970s, and according to Vorne (OEE.com) most manufacturing plants are closer to 60%. These values are industry benchmarks, not universal targets: what really moves the needle is your own trend and the detail of which factor is losing ground.

The losses OEE is meant to catch are usually grouped into the six big losses of TPM: breakdowns, setups and adjustments, minor stops, reduced speed, process defects and startup losses. Each one hits a different factor in the calculation, which is why OEE is not used just to grade the plant, but to direct where to attack first.

OEE vs other plant metrics

OEE does not stand alone. It is complemented by maintenance reliability metrics. It is worth not confusing them:

A common mistake is to look only at overall OEE and lose sight of which of the three factors is dragging it down: two plants with the same 70% OEE can have opposite problems (one with many breakdowns, the other with a lot of scrap) and therefore completely different action plans. Another frequent mistake is to compare OEE across very different equipment or across plants with different measurement criteria: if one includes planned stops in availability and the other does not, the numbers are no longer comparable. That is why it is key to define and lock the calculation method before you start measuring.

Common mistakes when implementing it

• Inflating performance by using an "ideal" cycle time that is too generous, which gives an artificially high OEE.
• Not recording minor stops, which are the hardest loss to spot by eye and the one that erodes performance the most.
• Using OEE to pressure operators instead of to find the root cause: the metric is a compass for improvement, not a tool for controlling people.
• Measuring by hand in spreadsheets when the volume of data calls for automatic capture from the PLC or the plant system.

In Latin American practice, many companies start by measuring OEE at their bottleneck (the machine that sets the pace for the whole line) before extending it, because there every point recovered translates directly into more sellable units. Once the data is reliable, it feeds finer analysis for predictive maintenance and planning, and it becomes the foundation on which the continuous improvement targets of the entire operation are built.