When a line's OEE drops, the production manager needs to explain why.

OEE (Overall Equipment Effectiveness) combines stops, speed and scrap into a number, and therefore usually summarizes the plant's efficiency in the production meeting. For the same reason, it shows how much was lost, but not where or why the loss began. This source of loss is the root cause of OEE. Without it, the team may treat the wrong problem.

If the same stops have been at the top of the stop list for three months, the team knows the symptom, but probably hasn't gotten to the cause yet. In many factories, reaching this origin is still slow. The operator writes down the stop on paper and only enters the reason at the end of the shift. The supervisor fills out the production in a spreadsheet, and the team only sees the OEE at the next day's meeting. Without crossing the data, the same failure remains unexplained: there are cases in which it is only understood when someone realizes that it only happens in one model or color of the product. Meanwhile, the loss returns every time the condition repeats.

Following the number every day alone does not resolve this delay in explaining the loss. According to the study "2026 Overall Equipment Effectiveness", by PMMI, an association of processing and packaging machinery manufacturers, digital tools allow food factories to use OEE on a daily basis, and not just in the monthly report. In the same study, interviewees point out root cause analysis among the points that can improve. The research was reported by Snack and Bakery on 10/16/2026 [click to read more].

This delay decreases when the team always follows the same path. In practice, to discover the root cause of OEE you need to:

1. open the number into the three categories that form it: availability, performance and quality
2. check whether the calculation and classification of losses are correct
3. read period events by duration, shift, product, setup and process
4. check whether the goal achieved resolved the loss or just compensated it

Each step depends on the previous one: there is no point reading the events of a poorly calculated number.

What are the OEE losses?

OEE losses are divided into three families: the stopped machine, the slow machine and the bad part. Each family reduces a rate from the formula OEE = availability × performance × quality, in which performance also appears as performance. A line that breaks and another that runs slowly may have the same OEE and require different investigations.

On a daily basis, those who analyze the indicator talk to maintenance about the line that breaks, check the speed, adjustment and material supply of the line that runs slowly and discuss scrap with quality. Those who only receive the consolidated OEE do not see whether the loss came from a stopped machine, a slow machine or a bad part, and therefore do not know who to talk to first.

The OEE calculation separates these families. Imagine a shift with 500 minutes planned to produce. The machine stopped due to breakdown, setup and idleness, and ran for 405 minutes. In these 405 minutes, at the ideal speed, she would make 460 pieces. He did 414, because he went slower and had small stops. Of the 414, 396 were good and 18 were rejected.

The three rates in the example shift

OEE multiplies the three rates: 0.81 × 0.90 × 0.9565 = 69.7%. No rate is below 80%, and even so the OEE ends up close to 70%, because in the multiplication each rate is applied to what is left of the previous one, and the losses accumulate. The biggest loss is in availability, but the 10% of performance is 46 parts that are usually left out of the stop list: a slow machine does not appear as a stopped machine, and the small stop is often left unregistered.

Converted into time, the 396 good pieces are equivalent to about 349 minutes at the ideal pace, 460 pieces in 405 minutes. The other 151 of the 500 planned minutes, around 30% of the shift in the example, became a loss: 95 minutes of machine down and the equivalent of 56 minutes at low speed and scrap.

Why doesn't the OEE dashboard show the cause of the loss?

The account shows what rate the loss is at, but not at what point in the process it began. This point is the root cause of OEE. The panel only shows the symptom.

In Sequor's experience with OEE, the downtime that the system records is not always the source of the problem. One station stops because the previous one did not deliver parts. What appears to be a loss of speed is a lack of material on the line. Therefore, before looking for the cause in the events, it is worth checking whether the number itself has been correctly determined.

Which investigation errors hide the root cause of the OEE?

The error is sometimes in the factory indicator itself. When OEE is implemented with a generic rule, the same for all processes, it can show 80% on a line where the correct calculation would give 70%, and the error also happens in the opposite direction. Does your line's rule consider its process or is it the same as others?

In addition to the rule, three calculation errors change the OEE or its reading:

• the OEE of the shift calculated as the average of the OEE of each hour
• the microstop classified in the wrong family
• the loss of the previous season counted in the next season

OEE of the shift calculated by the average of hours

An hour with 20 minutes available, for lunch or shift change, weighs less than a full hour. In the simple average, the two count the same, and a very good or very bad short hour pulls the shift to a number that the line did not deliver. The right thing to do is add up the minutes and pieces for the entire shift and calculate the sum.

Microstop in the wrong family

Microstop is a very short interruption, which the line then resumes. It reduces the parts produced in the time the machine runs, which is why it improves performance. The long stop becomes available. If the same interrupt is thrown in davailability in one shift and performance in another, the total OEE does not change, but the comparison between shifts loses meaning.

Loss from the previous station counted in the next station

If station 3 stops because station 2 did not deliver a part and this time is recorded as a loss for station 3, the team investigates a machine that was ready to produce. Each loss must be counted at the station that caused it.

How to find the root cause of downtime when Pareto is not enough?

With the reliable number, the search moves on to events. If the largest bar of your Pareto stops is called "other", the chart shows how the stop was recorded, not where the loss began. To find the root cause of stalls beyond Pareto:

• read all the events of the period
• order the stops by duration, not just the number of occurrences
• cross events by shift, product, setup, process and raw material
• compare similar situations that had different results

Pareto remains the starting point, but has two limits. The first is registration. In free text, noted at the end of the shift, the biggest losses tend to be "interval" or "other", which do not identify machine or process. And an hour downtime may involve just a few minutes of maintenance and the rest of waiting. If the reason recorded is just "maintenance", the wait does not appear in the analysis. The alternative is to register on the spot: in stop control, the operator chooses the reason from a list.

The second limit is counting. Two three-minute stops and one forty-minute stops appear as three occurrences of equal weight, and the order by duration changes priorities.

With registration and order corrected, the cross shows what Pareto hides. Imagine a stop that appears every shift and is never explained. When events are separated by product, it focuses on a model or color. Comparison with models that run without it points out what changes in that condition: a setup adjustment, a raw material, a speed. That's where the investigation has to begin.

Do the 5 whys and Ishikawa still work?

They serve, after reading the events, which shows which problem to attack. The process team then uses the 5 whys or the Ishikawa (fishbone) diagram to arrive at the cause of the process.

When the OEE target is reached, has the loss been resolved?

Not always. The target can be reached at the end of the month at the cost of overtime, maintenance and redesigned setup, and a month with a higher OEE may give less results than another with a lower OEE. This effort is outside the OEE account, and the loss compensated without treating the origin may return. Therefore, the last step of the analysis is to ask what it cost to reach the goal.

How does technology change in the search for the root cause of OEE?

With one line, you can check the calculation and read the events in a spreadsheet. With multiple lines and shifts, the volume of stops grows, and the team's time goes to gathering and tidying up data instead of acting on the loss. For analysis to become routine, technology helps in five points:

• Registration at the time of the event. The stop comes from the signal from the machine or from a note at the station, with a list reason, and the alarm from the PLC (programmable logic controller) is stored together.
• A calculation rule. The same definition of available time and microstops on all lines and shifts.
• Events ready to cross. Shift, product, setup, machine and duration are in the same record.
• Warning during the shift. When the downtime passes a defined threshold, the system warns production and maintenance while the loss occurs, and the time between the problem and action decreases.
• Machine signals. Vibration, current and temperature analyzed together can show the change before breakage or scrap.

With this, in addition to knowing why the machine stopped, the team starts to ask if it was possible to notice it before.

Where does Sequor come in?

Without structured analysis, getting to the root cause of OEE depends on a spreadsheet, days of analysis and someone who knows statistics.Sequor shortens this path on two fronts, with different roles: MES records what happened and calculates the OEE, and OEE Intelligence analyzes this data to show where and why the loss started and how to prevent it before it happens.

Reliable recording, at the time of the stop. Sequor MES identifies the stop in real time, records the reason in a list of up to four levels and calculates the OEE automatically. Andons, screens visible from the line, show active stops during the shift. The gain is a number that the team trusts, without waiting for the end of the turn to be recorded.

From loss to cause, without a spreadsheet. Sequor OEE Intelligence cross-references these events so the team can get to the root cause of OEE. It shows the three rates per line, machine and shift, with target and deviation alert, and opens the loss at four levels, for example from availability to the failure of a sensor. The gain is time: the team stops exporting data and creating spreadsheets and starts discussing the cause, with less risk of wasting effort on the wrong problem.

Sequor OEE Intelligence transforms OEE from an indicator that records what has already happened into information to decide what to attack first, before the loss is repeated.

OEE Intelligence works on top of the MES that the factory already uses, as long as it records production, stoppages and scrap. It also needs ERP and automatic machine data collection.

When the cause is equipment failure, Sequor Dynamics comes in, which reads machine signals to anticipate the risk of failure. OEE Intelligence and Dynamics are part of Sequor's Data & AI solutions.

With the stoppage registered immediately and the loss open to the reason, the manager arrives at the production meeting already knowing the cause and what to attack first.

Frequently asked questions

What is the root cause of OEE?

It is the point in the process where loss begins. The panel tells you at which rate the loss fell: availability, performance or quality. The root cause tells which machine, step or conditionThe loss of production generated this loss.

Microstop is in availability or performance?

In performance. What separates microstops from stops is duration, and this limit is defined for each process. The important thing is that it is the same in all shifts and lines that will be compared.

Is there a world-class OEE?

There is no number that applies to every factory, because each process has its own losses. OEE was presented by Seiichi Nakajima as an indicator of TPM (total productive maintenance). The most useful comparison is with the history of the line itself, always determined using the same rule.

Does Stopping Pareto Show the Root Cause?

It shows where to look first, as long as the reason comes from a list and the bars add minutes, not just occurrences. The cause appears in the next step, when the events are crossed with the production context.

Where to start when OEE is low?

For one line. Check the calculation rule, organize the stop reason in a list and read the events by shift and product. When the reading gives results, move on to the other lines.

Conclusion

Low OEE is the symptom the team sees in the production meeting. The root cause is the point where the loss begins. In the example shift, around 30% of the planned time became a waste, and while the origin does not appear, part of it tends to return.

To go from symptom to cause, the team follows four steps: open the number at three rates, check the calculation, read the events by duration and context and ask what it cost to reach the goal. Before looking at the month's number, it's worth asking which loss was repeated and if anyone already knows why.

Next step

Download the e-book KPI's in the industry, a visual guide for manufacturing leaders, with the OEE calculation explained rate by rate and the roadmap for interpreting the result. At the next production meeting, do the math on a line with last month's numbers and see which of the three rates is the biggest loss.

Want to see how Sequor OEE Intelligence breaks down your line loss to the cause? Speak to a Sequor expert.