Key takeaways
"Why is my OEE low" is the wrong question to act on directly, because OEE is a product of three separate numbers. A 55 percent score tells you there is a problem, not where it is. The fix is to stop staring at the headline number and break it apart.
OEE is Availability times Performance times Quality. A low score means at least one of those three is low. Before changing anything on the floor, look at all three and rank them. The lowest one is where your loss lives, and it is almost never spread evenly. See how OEE is calculated if you need the formula first.
Availability loss means the machine is not running when it should be. The usual causes:
Symptom: the machine sits idle or stopped for long stretches. If availability is your lowest factor, downtime is the fight, and reason-coded stop data is where to look first.
Performance loss means the machine runs, but slower than it should. This is the sneakiest loss because the line looks busy.
Symptom: the machine is running most of the time but the count is short. It helps to be clear on availability loss versus performance loss, because teams often blame downtime for what is actually a speed problem.
Quality loss means the machine makes product you cannot sell.
Symptom: good units are well below total units produced. Quality loss is usually concentrated at startups or after changeovers, so that is the first place to look.
Once you know the dominant factor, do not fix the whole category. Within it, one or two causes usually account for most of the loss. A quick Pareto analysis on the reason-coded data points straight at them. Fixing the top cause of the top loss moves the number far more than spreading effort across everything.
The most common reason OEE stays low is that it is averaged . A plant-level or shift-level average blends a healthy line with a broken one and shows a mediocre middle, so nobody knows where to act. Always diagnose at the machine and the loss level.
When the data is live and reason-coded rather than averaged in a spreadsheet, the biggest leak is obvious. That is what Fabrico is built to show.
See the approach on a connected OEE platform , then work through how to improve OEE once you know the target, or book a demo .
Before you pick the lowest factor, turn each loss into minutes, because the three factors multiply. Take a 480 minute shift with availability at 80%, performance at 79%, and quality at 98%, which gives an OEE of about 62%. Performance is the lowest percentage, but availability loses 96 minutes (480 × 0.20), performance loses about 81 minutes (480 × 0.80 × 0.21), and quality loses about 6 minutes (480 × 0.80 × 0.79 × 0.02). When two factors are close, the minutes tell you where the most time is going.
The often quoted world-class OEE of 85% comes from Seiichi Nakajima, who introduced OEE and the Six Big Losses as part of TPM in Japan in the early 1970s. It rests on about 90% availability, 95% performance, and 99% quality (0.90 × 0.95 × 0.99 is about 85%). OEE.com reports that most manufacturers score closer to 60%, and that it sees more plants below 45% than above 85%. Judge your score against your own baseline and line type, not against 85% alone.
One cause worth ruling out first is a wrong ideal cycle time. If it is set faster than the machine can physically run, performance looks permanently poor. If it is set too slow, performance can read above 100% and hide real losses. Check the ideal cycle time against the machine's rated speed for each product, and check that breaks and planned maintenance are treated the same way in planned production time on every line, before you change anything on the floor.
It is relative to your line type, but the useful move is not to judge the headline number. Split it into availability, performance, and quality and see which factor is dragging it down.
The lowest of the three factors, and within it the single biggest cause. Chasing a small loss while a big one bleeds is the usual waste.
Usually performance loss: micro-stops and reduced speed. The machine runs, so it looks healthy, but the count falls behind the ideal cycle time.
Yes. Manual logging misses micro-stops and mislabels stops, which distorts the split. Automatic, reason-coded capture is the fix.
An average blends good and bad machines into a meaningless middle. Diagnose per machine and per loss, never at the plant average.
Stop chasing downtime by hand. Fabrico logs each stop against the machine's history, so a repeating fault shows up before it becomes a hard line stop, and the team can open a prioritized work order with the stop itself captured on video wherever a camera covers the line.
Related: Overall Equipment Effectiveness (OEE), the complete guide to the formula, the three factors and what a realistic score looks like.
Connexion
Blog