Overall equipment effectiveness (OEE) is a key performance indicator that reveals an asset's overall productivity. The best part? When an asset is under-performing, OEE helps you find out why. That's because the data you used to calculate OEE is the same data you'll use to find the source of your problem.
There are two ways to calculate this important manufacturing KPI. Let's look at the more basic equation first.
For this one, calculate OEE as a ratio of fully productive time to planned production time. Here, the definition of fully productive time is when you're producing only good parts, as quickly as possible, without any stop time.
What's stop time? It's the sum of your planned and unplanned stops. Planned stops are for things like changeovers, where you're setting up or adjusting the asset. Unplanned stops are for failures and unscheduled maintenance. If an asset is offline for scheduled maintenance, don't include it in the stop time.
To calculate OEE, put everything together using the following equation:
quick and easy OEE
This is a perfectly acceptable way to calculate OEE, and it gives you a pretty good number for both bench-marking and base-lining, which we cover a bit lower down. But there's a weakness to this equation, which is that when you get a low OEE, you're not sure why. That's the bad news. The good news is there's a better way to calculate OEE, and it's better exactly because when OEE is low, you can figure out why.
For this one, you need to know your three loss-related factors: availability, performance, and quality. After that, the math is straightforward. OEE is the product of multiplying the three factors.
loss-factors and OEE
But we're getting ahead of ourselves. Before we can plug them into the equation, we need to know what these factors are and how to find them.
Ratio of run time to planned production time. Take how long you ran the asset and compare it to how long you'd planned to run it.
For example, an asset was scheduled to run for eight hours. That's your planned production time. At the end of the shift, it had only run for seven and a half hours. That's the run time. Where did we get the seven and a half hours? We took the planned production time (eight hours) and subtracted the stop time (in this completely made-up example, 30 minutes).
Availability = run time / planned production time
Ratio of total count to ideal run rate. Take how many widgets the asset made and compare it to the maximum number of widgets it could have made.
For example, when everything is running as perfectly as it can, your asset produces 1000 widgets every eight hours. But when you ran that asset for eight hours, it only produced 950. It might not have been from a major failure. Performance is affected by slow cycles, when some widgets take longer than expected to complete. And it's also affected by small stops, pauses so short they don't get counted in stop time. But small stops do add up and affect performance.
Performance = total count / ideal run rate
Ratio of good parts to total count. Take the number of parts that pass quality control and compare it to the total number of parts produced. It's a bit more specific than that. Good parts pass the first time they are inspected and require no additional reworking. Other parts might pass eventually, but they're not included in your good parts count.
For example, and again we'll use the classic widget, of the 950 widgets you produced in eight hours, 945 of them were good.
Quality = good parts / total count
Once we have our three loss-related factors, we can quickly calculate the OEE. But why is this second equation better than the first? It has to do with how you use OEE, covered in the next sections.
Overall equipment effectiveness is an indicator of a manufacturing process's efficiency. Once you've calculated it, you can use it as a benchmark or a baseline.
When comparing OEEs, there's always going to be some variation based on the type of manufacturing, but here are some general numbers.
100% Perfect. You're producing as quickly as possible, without stops, without mistakes.
85% World-class. For a lot of manufacturers, this is the long-term goal.
65% Typical. The good news is there's lots of room for improvement.
40% Not unheard of. Often when companies first start tracking OEE, this is where they find themselves. Again, the good news is there's a lot of room for improvement.
You can also use it for internal benchmarking. For example, you can compare the OEEs of two similar assets. Or, you can look at an asset's OEEs from different shifts.
You can also use OEE to track assets over time. As you fine-tune your processes, OEE should increase. If it's not, the numbers you used to calculate OEE can give you some solid clues about why. Again, here's why using the second way of calculating OEE is better. To get the OEE, you need your loss-related factors. And it's those factors that help you figure out what's going wrong.
when you know the loss-factors, you can see what's holding you back
If availability is dragging down the OEE, it could be because the asset is not on the right preventive maintenance schedule. It's failing too often, which means your preventive maintenance work orders are too far apart. Longer downtime can also be related to inventory control. If techs have to wait on critical parts before they can make repairs, you need to set the inventory par levels higher to ensure they have the parts they need when they need them. A work order software ensures your PM work orders are scheduled in a timely manner with the right frequency.
If OEE is having performance issues, you might be doing the wrong type of maintenance or the right type but incorrectly. For example, you might be inspecting and realigning the belts on the correct time- or meter-based schedule, but you're not realigning them properly. It's like you're having breakfast every morning at the same time, and it's the right time, seven-thirty am, but instead of coffee and toast, you're eating ice cream. Right schedule, wrong task.
Quality issues also lower OEE, but are usually not something the maintenance department can fix. The asset might not be set up properly, or there could be operator error. The problem might even be from the feed-stock.
OEE is just one of the important KPIs for maintenance. There are a bunch more, but it's not a case of the more, the merrier. You need to find the ones that work best for you, that help you reach your goals by giving you information you can use. Imagine you're on a road trip. It's helpful to know things like your your average speed and mileage. You could also track the number of tire revolutions, but what would be the point? Not every type of data has value. For a good introduction to some of the more common ones for manufacturing, check out What are MTTR, MTBF, and MTTF Metrics?