What is OEE? (Plus 5 Reasons OEE is an Epic Fail)

What is OEE (Overall Equipment Effectiveness)?

OEE is, by definition, an abstract manufacturing metric. To oversimplify, OEE takes quality, availability and speed and mashes them all together to give you a number that represents performance.   The exact definition is OEE (%) = Availability rate × Performance rate × Quality rate.  Each factor breaks down like this:

Before you use OEE, allow me to offer some advice.

Logic vs Abstraction

Some years ago, back at Washington State (Go Cougs), my philosophy professor loved debate. He was younger than most and would debate anything and everything as long as you followed his rules.  His debate rules were simple but inviolable: I had to argue a minimum of three premises, present a logical conclusion, and IF he accepted my premises, he also had to accept my conclusion.

I loved it, because it brought logic to arguments that are based on ideologies and emotion. I’ve used it frequently, and I’ve found it improves my understanding of the counter argument.  In some cases, I’ve actually revised my position because of the exercise.

The Origin Story of OEE

Throughout much of my post college journey, I’ve been involved in manufacturing.  I’ve been in automation, analytics and performance of manufacturing and machines, and I’ve been in various leadership roles on the production and business side of manufacturing.

Early on, about the time when PC based HMI’s became more common, time based historians from the continuous process industry began to be deployed to discrete and batch based manufacturing. Some well meaning PhD decided that since historian based analysis of Continuous Process manufacturing (CP) helped performance in that industry, it was only natural that it would help other manufacturing types as well.

Don’t get me started on that one, we’ll be here till next week. In a nutshell, discrete and batch went from having no data at all to drowning in data.  To this day, most manufacturers don’t analyze the information consistently, and even fewer manufacturers use the information to improve performance on a day to day basis.

Unlike CP, discrete/batch manufacturing are by nature stop/start processes and some have the additional challenge of multiple products running on the same line (batch manufacturing).   In addition, CP has 100 years of gradually improved methods for processing like SPC.  Discrete and batch are incredibly diverse, and methods of manufacture vary greatly, and in many cases include some form of human assistance.

In addition, in CP most sensor data is, well, continuous. Meaning:

1. Data is collected every x seconds / minutes/ whatever
2. You can use the data because the process never stops.

So go ahead and log that temperature sensor, and then use SPC to improve output and quality.  This CAN’T work in discrete/batch, because you may start and stop 30 times a day.  Your production might do some rework, change products, or do a Clean-in-Place.  Somehow, you have to tie historian data to production events.  I won’t go into details, but THIS IS HARD TO DO.

Turning data into information

After the data is stuck in the right spots, you must face the most daunting challenge of all.  How do I turn data into information people can use to improve performance.

This is no small challenge, and it requires three things:
1. A tool that can show and alert on information in real time
2. A method for displaying the information that is cause specific and actionable
3. Prioritization of the most important issues in a way that requires attention and devalues less significant issues

Big ask.

OEE wasn’t intended for the plant floor

It’s not as if industry hasn’t tried. We have been inundated with production metrics and manufacturing KPIs.  OEE was invented to quantify production line performance. OEE was a logical idea, but OEE is designed for the boardroom.  OEE has NO PLACE on the plant floor.  I don’t believe it belongs in the boardroom either, but that’s a tale for another day.

OEE is, by definition, an abstract metric.  As I said above, Overall Equipment Effectiveness takes quality, availability and speed and mashes them all together.  So, maybe you’ve expended the human and financial capital to capture the necessary raw data to show your OEE.   Ok, You run a report and It says Line 1 is at 61% OEE and Line 2 is at 56% OEE.

Cool.

Now what?

Well, say the academics and six sigma pundits and all of the other theory spewers: now you can improve your number!

Why doesn’t OEE work?

So my question to all of you out there is: how do you go from an abstract metric to better performance?  Well, according to most, you need to form local teams to take ownership of each production assets OEE, dig into it, discover what the problems are and then build other teams to address each problem that was identified.  Really?  Not only would that take YEARS, but it’s doomed to fail.  Why?  Because of the following three reasons:

1. No Time. Manufacturers and machine builders run skinny. Every person on staff is tasked with a day job.  Your faithful Continuous Improvement staffer can’t build a team out of people that don’t have any availability! In addition, the time between meeting one and the “fixes” the team recommends are put into place is measured in months.  AT BEST.  Yet batch and discrete production have additional changes that occur DAILY.
2. No method. These teams aren’t usually lead by manufacturing performance experts. In some cases they are lead by continuous improvement, and in others they are led by folks that know how to make the product well.  This means that every team will come up with their own method for improvement, and it will be tied to the biases of the group. The corollary is the method won’t consist of best practices for performance improvement.
3. No specific and actionable information. These teams are saddled with an OEE: an abstract method that says you need to improve. It does not, and never will, provide specific and actionable information. In fact, it invites the team to make their OWN INTERPRETATION of what is happening.  To have any long term success, it also requires validation of each new interpretation.

I’m not picking on Continuous Improvement groups; CI is usually the only path to improvement.  I am saying they need more than OEE to do their job effectively.  They need to be able to put real time, actionable information that is specific and prioritized in the hands of people who can make changes. What is OEE?  Well, OEE is certainly not a tool for Continuous Improvement teams.

5 Reasons OEE is an Epic Fail

1. OEE does not show you when machines are starved for materials
2. OEE does not tell you when machine adjustment and maintenance is required
3. OEE does not provide a specific machine or production line problems
4. OEE does not show specific downstream blocks or problems
5. Summary:  OEE provides no specific, actionable loss information.

How do you improve manufacturing performance?

Let’s win the damn shift.  For that matter, let’s win the hour!

Focus your people on what they can do, not how poorly they are doing.  Give them a tool that is both method for action and actionable information.

And whatever you do, don’t put OEE on the plant floor and point it at your operations staff as a tool. If you want to use it, keep it for the management team and away from your action staff.

Did I win the argument?

So, let’s circle back around to the main theme.   If you accept my premises are true then you accept my conclusion. (My game, my rules :):

• Premise 1.  Machines require materials, adjustment and maintenance to perform at rated output.
• Premise 2.  Machine performance is based on minimizing and quickly solving input, output and internal losses.
• Premise 3.  Losses are specific, actionable causes which can be addressed in a timely fashion only if identified.
• Premise 4.  OEE provides no specific, actionable loss information.
• Conclusion. OEE should not be used on the plant floor to improve real time performance.

Food for thought.

________________________

More info on our products:  Wave7