OEE for maintenance managers

Monday morning meeting. The OEE figure for line 2 is on the screen in red. The ops manager looks at you.

You know what happened. The case packer jammed every few minutes on Thursday nights. The capper was turned down because it drops lids at full speed. And the big stop on Saturday was two hours of fault-finding, then three hours waiting for someone who knew the servo.

None of that is on the slide. Just one number, and the feeling that it's your fault.

OEE is a useful number. But most maintenance managers get handed it without the breakdown underneath. Without that, you can't tell which losses are yours, and you can't argue for what would fix them.

Three numbers multiplied

OEE is availability times performance times quality.

Availability is run time against planned time. Breakdowns, changeovers and waiting for materials all eat into it.

Performance is how fast the line ran while it was running, against its ideal speed. Small stops and slow running eat into this one.

Quality is good output against total output. Scrap and rework on restart sit here.

Here's a worked example. Say a line is planned for 8 hours and loses 1 hour to stops. Availability is 7 out of 8, or 87.5%. While it's running, it makes 90% of its ideal rate. And 98% of what it makes is good. Multiply them and OEE comes out at about 77%.

Each loss looks small on its own. Together they take nearly a quarter of the shift.

Which losses are yours

Breakdowns are the obvious one. But maintenance owns more of the slide than most people think.

Look at performance. A line that's been turned down because it jams at full speed is a maintenance problem. So is a filler running slow because a valve sticks. Nobody logs those as breakdowns. They just quietly become the new normal.

Then there are the micro-stops. The case packer that jams every few minutes. The sensor that drops out and needs a reset. Operators clear them in seconds, so they never reach the breakdown log. Over a week they can add up to more lost time than the big stop everyone talks about. Our post on micro-stops covers how to catch them.

Quality matters too. A line that scraps the first twenty minutes after every restart has a maintenance story behind it.

Split the stop time

Here's the bit that changes the conversation. Break every big stop into pieces:

• time to respond: from the stop to someone at the machine
• time to find the fault
• time waiting: for a part, a permit, or the right person
• time to fix
• time to restart and get back to rate

On a lot of stops, the fix itself is a small part of the total. The rest is waiting. And a lot of that waiting is for the right skills. A mechanical fitter on nights can spend two hours on a drive fault that a controls engineer would clear in twenty minutes. That time sits on the slide as "breakdown", and nobody asks why.

Once you split it, the fix becomes obvious. If response time is the problem, look at your cover. If waiting for skills is the problem, look at who's on shift and who you can call. Our post on mean time to repair goes further into this, and what an hour of downtime really costs you puts a price on each hour.

Where OEE can push you wrong

Here's our view. OEE is a production number. Used carelessly, it pushes maintenance in two bad directions.

The first is planned work. Every hour of PPM counts against availability if it's booked into planned time. So the pressure builds to skip it. HSE's advice runs the other way: "try to use downtime for maintenance" (HSE, maintenance of work equipment). Agree with production how planned stops are counted before the argument starts.

The second is safety. OEE rewards a line that never stops. That puts pressure on operators to clear jams without isolating, and on fitters to work around faults. HSE warns that injuries from "the use, maintenance or clearing blockages at packaging machines can be severe or even fatal" (HSE, packaging machinery). If your micro-stops are being cleared by hand on a running machine, the OEE figure is hiding a risk.

The PUWER code of practice has a line worth keeping for these meetings. When the law says kit must be maintained in an "efficient" state, it relates to health and safety. In the code's words, "It is not concerned with productivity" (L22, paragraph 70). The code also points out that a production fault usually shows up quickly, while a fault in a safety-critical system "could remain undetected" unless maintenance checks for it (L22, paragraph 71). OEE will tell you when the capper drops lids. It will never tell you the guard interlock has stopped working.

Make the number work for you

Used well, OEE is the best case you'll get for more resource.

Take your losses for a month. Rank them by time. Put the top five on one page, each with its cause and a fix. "Line 2 lost six hours waiting for a controls engineer on nights" is a far better argument than "we need more people". Our post on maintenance KPIs that mean something covers what to report alongside OEE.

And when the big losses are waiting time, look first at cover on the shifts where you're thin, and at an engineer you can call when the fault is beyond the people on site. That can close the gap faster than a new hire.

What to do on Monday

1. Ask for the OEE figure split into availability, performance and quality for each line.
2. For every stop over 30 minutes last month, split the time into respond, find, wait, fix and restart.
3. Agree with production how planned maintenance stops are counted.
4. Take your top five losses, with a fix for each, to the next meeting.

* Paragraph numbers refer to L22, fourth edition (2014, amended 2018), correct as of 21 October 2026. Check HSE's maintenance guidance for the latest.