When NOT to use a DOE
Our video today is about when NOT to use a DOE. But before we talk about when NOT to use a DOE, let’s make sure you have a general idea of what a DOE is.
Video: DOE, Warning! When NOT to Use!
First off, what IS a DOE?
DOEs are awesome tools in the right time and place! If you’re not familiar with the term DOE, it stands for Design Of Experiment. It’s a special experiment designed to analyze variations that may affect your given condition. In our case, it’s typically to identify whether or not we’ve found the root cause of the problem.
In case that sounds complicated, let’s simplify it a bit. Let’s say you have a water leak around a screw in your roof. Sometimes it leaks, sometimes it doesn’t. Now you can run a DOE by screwing in a whole bunch of screws a few different ways until you figure out what the actual problem is. This is an example of a GREAT time to run a DOE!
You can find a lot of good information on the internet that’ll go into further detail on how to run a DOE, if that’s what you’re looking for. But for today I want to explain why I think DOEs are one of the most commonly misused tools in problem-solving!
Misuse of DOEsUnfortunately in my experience a DOE is one of the most commonly misused tools in problem-solving! Click To Tweet
There are 2 main reasons for this:
- Most people don’t realize that there’s usually only 1 main root cause to their problem.
DOEs can be spectacular if your project actually does have more than 1 main root cause! But are you ready for a surprise? Even if you really do have more than 1 root cause, you can often confirm this without even using a DOE! (I’ll tell you how in just a minute!)
- A lot of people think there’s only 1 way to prove if you have the root cause or not: and that’s to run a DOE.
Maybe you love DOEs and you feel extra smart when you run them, but I’m about to give you some first-rate reasons to think twice next time before running your DOE.
- Proving your root cause without a DOE will probably be FASTER!!! When you have a big problem costing your company a lot of money, the faster you can prove your root cause and start implementing a solution, the better!
- Proving your root cause without a DOE is likely to save you money!!! DOEs can be very expensive!
So why would anyone in their right mind throw away a bunch of money running some slow, unnecessary tests when there’s often a faster, better way to get the job done without running a DOE?!
And that, my faithful readers, is why we are talking about: When NOT to run a DOE!
Here’s how you know when NOT to run a DOE:
DO NOT run a DOE if you only have 1 root cause candidate!
Really people, this should almost go without saying. Why would anyone think it’s a good idea to run an experiment to analyze different variables… when you’ve already narrowed your relevant variables down to… ONE?
If you really only have 1 valid root cause candidate, just go ahead and make the change. Chart it over time. It is absolutely pointless to squander valuable time and money running a DOE when all you have is 1 potential root cause.
And what I never, ever want to see someone do is to add more possible root causes to a DOE than they actually need, just to justify running it! If 1 is all you need, then forget the DOE and move on!
DO NOT run a DOE if you can run a different type of test that might be more beneficial.
This video discusses some alternatives to running a DOE that can actually be more productive sometimes. For example, if you are able to turn the problem on and off 3 times, then do a process test instead. Compare good processes to bad processes.
DO NOT run a DOE if you can make a chart showing how that root cause is going over time.
Bad Habit of Problem-Solvers
I’m going to stop for a minute here and bring up a bad habit that many problem-solvers fall into when running DOEs. That is adding unnecessary variables.
“They will add more possible root causes to the DOE than they actually need. Sometimes they do that because they really don’t know what to run, they’re just guessing, and they don’t want to spend the time actually solving the problem the right way.”
A lot of times these guys will run a five variable DOE! The fact is, really they only had 2 or at the most 3 variables that they needed to run. For whatever reason, they just think it might be a good idea to throw in two extras to see what would happen. In other words, they are throwing away time and money on something completely irrelevant to the problem they’re trying to address!
So you’ve gone through this list of reasons NOT to run a DOE and you are still convinced that your project would benefit from a DOE? Here’s my recommendation:
Bring the variables down to two, maybe three at the most, and then run your DOE.
How to Recognize a GREAT Problem-Solver:
Great problem-solvers think smarter, not harder.“A great problem-solver should be able to find the small amount of variables that actually make a difference before running a DOE.” Click To Tweet
I’m going to save you from making a huge mistake!
Some people think that running a huge DOE will make them look smart! But that’s not the case. Running more than 2-3 variables on a DOE will instead make a problem-solver seem inexperienced or lazy. So be sure to always take the time necessary to understand which 2-3 variables are actually relevant. And then by all means, run your DOE!
If you want to solve your projects effectively, efficiently, and lightning fast – and you’re considering using a DOE – remember these steps:
- Evaluate if a DOE is actually necessary or if something else would be even more valuable.
- Make sure you have more than one variable that could be the root cause.
- Once you’ve concluded that a DOE is actually going to benefit your project:
- First go learn more about the root cause.
- Learn more about the project.
- Be sure to narrow down the variables to only 2 or 3 at the most.
If you have experience with running DOEs I would love to hear about it! Comment below to let me know if this video was helpful to you. And let me know what you’d like to hear about next!
Be sure to check out my other videos like: