To DOE or Not to DOE

I get asked by my colleagues from time to time for advice on setting up a DOE, and what I usually find is that they’re nowhere near ready to effectively execute one. I guess DOE is the “sexy” weapon in the six sigma arsenal (although I’d put it head-to-head with ordinal logistic regression or even a kappa study any day), but if not used properly it can be a waste of time at best and misleading at worst. In the DMAIC process, DOE happens somewhere in the “I” phase when the process inputs have been well-defined and understood and you’re beginning to define your y=f(x) transfer function. If you’ve skipped “D”, “M”, and “A” altogether and want to jump straight into DOE, it’s a good bet you won’t get the most out of it. Here are some questions to ask yourself if you’ve been bitten by the DOE bug:

 1.  What problem are you trying to solve? Make sure you understand why you’re running the experiment in the first place. Remember that six sigma deals with normal variation, not special cause variation. So if you’re trying to find the root cause of a defect that has suddenly appeared, you might want to take a different tack.

 2.  Do you really trust your measurement system? It’s amazing how often people will simply gloss over this one, even when they know there are inconsistencies due to gage or operator variation. Keep in mind that the discrimination of the measurement system will have a lot to do with the interpretation of your DOE results.

 3.  Do you know what factors to test and at what levels to set them? This seems to be the most common hang-up that I’ve encountered. Typically there is a lot of hypothesis testing that needs to be done prior to thinking about a DOE. If you don’t know what your significant inputs are and have at least pretty good idea of where to set them for an experiment, then chances are you need more analysis. ANOVA and t-tests are good for this. Also consider a fractional factorial screening DOE before going all out.

 4.  What are you going to do with the data? This goes back to number 1 a little bit. Make sure you understand how to interpret the results you get. Running the experiment is just the first step. Whether you plot the results by hand or dump the whole thing into Minitab, you’re going to get a bunch of data. In order to use it for intelligent decision-making you need to know what it means.

 DOE’s are time-consuming and often expensive to set up, run, and analyze. Take a little time to make sure it’s the right tool before committing to it.

Learning the Layers of Lean

Last week I had a neat opportunity to visit a company that was recently acquired by our own. The purpose of the week-long visit was to teach them about how we do lean conversion and to help them develop a roadmap for their facility. They say that if you really want to learn something you should try teaching it, and boy have I found that to be true. So being a learner first, I found last week’s trip infinitely valuable.

 I think the most striking thing that was reinforced for me was how much of a layered approach you have to take with deploying lean. It’s like a house. You can’t put up the walls until you have a solid foundation. You can’t even start the second floor until the first floor is completely framed. And don’t even think about starting on the roof before everything else is in place.

 Just as with building a house you can’t make 1-piece flow work (or even really understand) it without having established standard work. You really have to have flow working before implementing pull systems. You can’t even think about heijunka until you’ve at least somewhat mastered SMED. And forget about SMED if you don’t have good 5S. Starting with 5S and standard work, you have to carefully lay the foundation for lean conversion. Then really work it before moving to the next level.

 From a teaching perspective this can be a little tricky. You want people to see the vision and understand the great benefits of a lean system, but you can’t just dump it all on them at once and expect them to embrace it.  That would be like trying to explain concepts of algebra to someone who had not yet mastered arithmetic. The key is patience and knowing what to work on and what to defer. A clear view of the road immediately ahead will help keep the view from becoming murky.

 It’s both frustrating and exhilarating at the same time. Frustrating because I  am eager to share all that I know with anyone who will listen, and I know this can be counterproductive. Exhilarating because I know that my own lean journey has many lessons and milestones yet to be revealed.

The Hitchhiker's Guide to Lean

I haven’t written a book report since I was in school, but I picked up this book a while back and thought I’d share by impressions. I am of course in no way qualified to offer a critique, so these are just my general thoughts.

Oh No, Not Another Book About Lean!

Admittedly this was my first thought. But I’ve long admired Mr. Flinchbaugh’s blog and his seemingly honest, straightforward, and technically sound approach to lean; so I figured I’d give it a try. The first section is a basic lean overview, but it does offer an interesting distillation of lean into a few basic principles. Even several years into a lean journey of my own, I found it to be a good perspective on the journey and its key guideposts. I found it especially helpful to be reminded of the importance of “high agreement”.

Advice for Lean Leaders

The book also touches on one of the most important topics of lean transformation, leadership. Although I do not have nearly the depth or breath of experience in leading lean that Mr. Flinchbaugh has, my experiences align almost perfectly with the book’s principles. I think this is a great synopsis of effective leadership, and I think that everyone in a lean leadership position should hear this message.

Charting Your Course

I’ve personally seen the enormous galvanizing power of both a roadmap and, as the book puts it, an operating system. These are crucial tools in getting going effectively, and I think chapters 4 and 5 do a good job of not letting us forget that.

Number Crunching

Chapter 6 reminds us that cost accounting can be a real impediment to lean transformation. I like that the subject is broached and that we can have the conversation. And I hope we see more progress in the area of lean accounting in the near future.

Material Movement

I like that this chapter doesn’t focus solely on kanban; but rather it talks about more fundamental keys to making material flow, like eliminating waste and making things visible. I was pleasantly surprised by this approach and learned a lot.

You Want Fries With That?

I was especially interested in the chapter on service, since I’ve been spending more and more time trying to improve transactional processes. This was the first time I’ve seen a clear and concise breakdown of the elements necessary to effectively manage a lean transformation in a non-manufacturing environment. I found it very helpful.

Look in the Mirror

I get laughed at all the time for trying to find innovative ways to use lean in my personal life. So the chapter on “Personal Lean” hit very close to my heart. There’s great stuff in there, including some very helpful tools. I like that someone else recognizes the tremendous value in walking the walk. I think this goes hand-in-hand with good leadership.

Case Studies

I always enjoy these, and there are some great ones at the end of the book.

Summary

I think Mr. Flinchbaugh has written a solid and comprehensive book on lean, and I would recommend it to anyone who is contemplating or who has set out on a lean journey of their own. No one can truly show the way, but in the pages of this book you’ll find the knowledge necessary to plot your own course.