Q: You wrote we should track a maximum of 10 to 15 failure categories, but I want to define 85. How else can I ever know what part is breaking?

Arno Koch •    At first glance that seems to make sense. But OEE is not a breakdown registration system, OEE is a loss detection system.

Let me explain the consequences: Imagine after 3 months registering OEE data, it shows the main loss is in availability. So now you want to know what you need to do to get your availability up. So you take one of the seven tools and draw a pareto diagram of all your time-events. Lets assume each of the 85 breakdown items occurred at least once. What you will see, is a pareto with an immense long tail, not giving a clear clue where the main losses are located.

Register process steps instead of components

In the other approach where you would have registered failures on, lets say, 5 process parts of the equipment (like ‘incoming conveyer, pre-heater, moulder, compressor, outgoing conveyer’) it would reveal what part of the equipment is restraining the process most. Then you start temporarily to focus on that part. A simple registration card (maybe even showing the 85 breakdown items) for some weeks will give detailed insight what’s going on.

Sometimes it is a matter of some simple maintenance, sometimes a Small Group Activity can solve the problem for once and forever. By this circle of focused improvement the equipment will become better and better.

This example also shows another disadvantage of the ‘breakdown registration approach’; if you handle each breakdown in the best way, thus eliminating it forever (either by reengineering or taking it into PM), it will not reoccur, but other breakdowns may start to occur. So after a while the breakdown registration system will not reflect the reality anymore. In the other approach this is less likely to happen.