Measuring the Overall Equipment Effectiveness (OEE) is one thing, but before you measure the OEE you should know when and where you actually need the OEE to improve your industry. This post describes what the OEE is good for and what it’s not.
There are countless OEEs measured in industry. Unfortunately, there is usually not much done with these apart from using them to evaluate management performance. In fact, this is usually a good thing since the OEE can be enormously misleading. Quick pop question … What is a better OEE: 60% or 90%? While many in industry will immediately answer that 90% is better than 60%, the truth depends on the circumstances.
After the definition of OEE, how OEE is measured, and the top three methods on how to fudge your OEE, we will now look at what the OEE is good for and what it’s not. Because, in some cases, a high OEE may be the worst thing you can do.
What the OEE is Good for … and What it’s Not
The OEE measures something akin to the utilization of a machine. The higher the OEE, the higher the output of parts from a machine. Hence, according to the Western mantra in industry that “only a running machine is a good machine,” a high OEE would be good, right? Lean manufacturing, however, has a different philosophy:
Produce only what is needed, when it is needed and in the amount needed. (Taiichi Ohno, Father of the Toyota Production System)
Hence, a high OEE without the corresponding customer demand would lead to overproduction. And, out of all the wastes in industry, overproduction is the worst of them all. Overproduction leads to all kinds of secondary waste, and lean manufacturing is usually most well-known for its lack of inventory.
Of course, this doesn’t mean that a low OEE is better than a high one. As mentioned above, it depends. But before we go into details regarding when a high OEE is good, I would like to point out that measuring the OEE makes sense only if you want to change the OEE, namely to have a high OEE. For processes where the OEE doesn’t matter, there is no reason to waste energy on measuring the OEE. Hence, you should measure the OEE only where the OEE matters, and you should measure the details on the losses only if you want to gather data for an improvement project.
Having said that, there are two approaches where a high OEE is useful: for bottlenecks and possibly for production lines.
The bottleneck is the process that slows down your entire system. Hence, a high OEE at the bottleneck translates into a high output of the entire system. If this is matched with a high customer demand, then measuring and possibly improving the OEE is relevant.
First, assume a machine that is NOT the bottleneck but before the bottleneck itself: A high OEE will lead to a jam of parts before the bottleneck, so in this case a high OEE is bad. Now assume the system below with five processes, of which the middle one, process C, is the bottleneck. If the goal is a high OEE on process A, the likely result is a pile of material between process A and process C.
This similarly applies to a high OEE in a process after the bottleneck. If you want to have a high OEE in process D, chances are you won’t get it, since process E will always wait for parts from the bottleneck C. Hence, process E cannot have a high OEE due to the bottleneck being somewhere else. Only an improvement in the OEE of process C brings an improvement of the overall system!
Please note that the above example assumes that all processes are working the same shift pattern. Of course, different shift patterns cause changes in the OEE, but as long as process C is the bottleneck, all other OEEs are not relevant. Please also note that process C does not necessarily have the highest OEE of the processes, since process C may be the bottleneck precisely because of many losses in availability, speed, and quality. So for process C—and only process C—the OEE is of interest.
Measuring Production Lines
There is a possible variation on the OEE for the bottleneck. It is possible to measure the OEE for the entire line. The OEE is the number of parts produced divided by the theoretically possible number of parts that could have been produced. This theoretically possible number of parts can be determined through the slowest cycle time in the system.
Please note: Adjust for number of parts needed for the final product if necessary (i.e., if a car needs four wheels but only one engine, the wheel production has to be four times as fast as the engine production). Also note that while you can measure the OEE for a production line, it is tricky to measure the details of the losses due to the interactions within the system. In my experience, it is usually nearly impossible to accurately say why the system produced less parts than theoretically possible. Hence, the line OEE gives you only a measurement of productivity, but not a tool for improving the line.
In any case, you can measure the OEE for a production line. In this case, the same applies as for measuring the OEE in processes: The OEE is relevant only if the line is the bottleneck. Furthermore, in theory the OEE of a line can help you determine if you can reduce the number of shifts while still keeping output constant. However, if you want to reduce the number of shifts, the measurement of parts produced per shift is a much more useful and easy measurement than the OEE of the line.
The OPE: Measuring People
After all the details on how and where to measure the OEE for processes, you may wonder, Can you also do this for workers? Can you also measure the OEE for your people? The answer is a typical lean answer: Yes, but …
First of all, in this case it is renamed from OEE to OPE. The OEE was the Overall Equipment Effectiveness (or Efficiency). The OPE was initially the Overall People Effectiveness (or Efficiency); however, this touched some sensitive nerves, as it indicates that employers treat people the same as machines, whereas people are not machines, but … well … people. Hence, while keeping the acronym OPE, it was renamed to Overall Process Effectiveness (or Efficiency ) or sometimes also Overall Performance Efficiency. Here it is merely assumed that the process or performance includes manual labor.
Now, in theory you can measure the OPE just as you can measure the OEE, by observation and note taking (probably less through digital monitoring ;-)). In practice, however, as I said above, people are not machines, but … well … people. And people do not like to be measured, especially by their supervisors and on their job. Hence, starting an OPE measurement has a high risk of also starting trouble. Your workers may refuse to cooperate, management loses (even more) respect, and even if you get your measurements done they may be worthless since the workers staged a dog-and-pony show for you. Never underestimate how much an employee can make you believe what he wants you to believe.
So you should measure the OPE only if there is a high level of trust and understanding on the workers’ side. Try to explain them why you and they need this data. Assure them that this is not to find out whom to fire. Convince them that this is not a way to squeeze even more work out of them. Involve their unions. Involve the workers in the project. Have the workers take the measurements themselves. And then hope that they accept and believe your motives, because otherwise your data will be garbage. Finally, stand by your promises and do not screw your workers over after you got your OPE, or loose all respect and trust from your workers.
In short, avoid measuring the OPE unless you absolutely have to. This also applies to the OEE, but much more so to the OPE.
This concludes my series of posts on the OEE. I hope it was useful for you and helped you avoid problems and improve your industry. Now, go out and improve your Industry!
See also my other posts on OEE:
- Definition of OEE
- How to measure the OEE
- Top Three Methods on how to Fudge Your OEE
- What the OEE is Good for … and What it’s Not