I have been blind

Now I like to think I am quite an objective & freethinking person and don’t always follow the herd when I think something is wrong. I’m not a complete contrarian but am willing to “go it alone” when I feel something is important. So it is a great disappointment to me to say I have only recently discovered DoE. Let me explain the circumstance of enlightenment first.

Over about a month, one of our processes fires two letters and an outbound telephone call to our customers to achieve a particular goal. The process is about 50% effective and we were discussing options for improvement. People talked about changing the wording on the letters or the call date. During the meeting, memories rushed-in of me in BB training, adjusting the settings on a catapult and measuring the distance the ball travelled and I slowly said, “we…could…design…an…Experiment”. I nearly pulled it off but didn’t quite have the confidence or conviction to convince people that DoE would fit the bill.

As I do, I dived into the DoE material, we had a classic 3-factor, 2-level, factorial experiment and I didn’t see it! How could I have got myself into a situation where I have ignored one of the fundamental tools of Lean Six Sigma? How could I have been so blind?

Since my earliest days on BB training when we covered DoE, the picture was painted of a tool used in manufacturing that really does not transfer into a transactional environment. The exercise was “manufacturing”, the examples were manufacturing. The books I have give manufacturing DoE examples; one of the more transactional books completely ignores it and most give a passing reference. My coaches have never really talked about using DoE and when they did they talked about it being difficult to apply in a transactional environment.

Handpicked Content:   The Next Next Big Thing

I never really challenged the orthodoxy here and feel I have really missed out. Am making serious amends and it’s a strict study diet of confounding, blocking, attribute response, response surface design and loss function for me.

Comments 6

  1. Dave B

    This still amazes me, I cannot grasp the concept of why DoE is always labelled as a manufacturing tool only. When I made the move from manufacturing to transactional (can I go back?) I harped on about this stuff and was told that it does not apply in this environment… Hmmmm

    Take a incident centre for instance, do an observational study data recording exercise with a team, segment the team and using the X’s – run the results through MT and using an array based test – start messing!!

    Guess as you state, the transferred mind set is historically manufacturing….

  2. Mike Carnell


    If you read Mario Perez-Wilson’s book titled “Six Sigma” it is in a story format but pretty closely tells the story of working on the FMU-139 bomb fuse line at Motorola (by the way Mario’s book is excellent but it is in a manufacturing context. If you want another good DoE book check out the stuff from Bob Launsby). We had a Production Manager that had a sign that you could see hanging behind him when he sat at his desk that read “Now that you have told me why it won’t work, tell me why it will.” Great comment. I hope you don’t take this personally but when you get in an industry that is not manufacturing people tend to look sideways at everything trying to figure out why this doesn’t apply to them – Deming’s Obstacles to Improvement “we’re different.”

    Interesting where they got labeled a “manufacturing” tool. If you do some research you will find that a lot of the early stuff came from agriculture i.e. split plot designs. Lots of work from your part of the world as well. If you want to see the transaction environment, try checking out conjoint analysis (make sure you understand what it is regardless of what it is called) used in marketing. The transaction stuff isn’t new. It just takes understanding what you want to know and what tools create that piece of information.

    Best of luck with your DoE. Software makes them relatively easy to analyze. They can be difficult to control. Your design is relatively small so it shouldn’t be a huge issue but it can go south in a hurry. If you want to try something interesting do an FMEA on the DoE before you do it and see how many holes you plug before you run it. You will probably save some money.


  3. Robin Barnwell

    Thanks Mike, as aways packed with useful advice.

    I’ll add the conjoint analysis to my study time. Bob Launsby’s web-site also looks promising.

    I’m wondering if I have reached a point where I can now say I really understand DMAIC and am now ready to push into DfSS via DoE?


  4. Mike Carnell


    That is a good question about understanding DMAIC. I learn stuff all the time. I think that means I understood it on one level but each time you learn something I think it takes you to another level of understanding. The interesting part is putting pieces together to make an improvement system. Everytime you get a new deployment the pieces fit differently.

    I am not sure if any of Bob’s books will take you to DFSS but if you can get him into a conversation he can make the link for you pretty quickly. He lives in Colorado Springs so the beer is cold and the scenery is great. Worth the trip. You know you are some kind of Six Sigma geek when you sit in some micro brewery discussing Six Sigma and leave believing you had a good time.

    If DFSS is where you want to go check out the guys at Statistical Design Institute (SDI) these guys are good. They have a couple old timers around like George and Tom who have been doing this from the Six Sigma Reseach Institue days at Motorola (they were TI guys). They have another guy named Jesse who is younger so they have the old and new mixed pretty well. I have had them to South Africa and they put the connection together really well for our MBB’s.


  5. Mike Carnell

    I just got an invitation to meet with a guy in London. That may be on the books before year end. Are you near London?


  6. Robin Barnwell

    Hi Mike

    Thanks for this. I feel I understand something once I have learned it, used it, and relearned it. I like to read around a subject and find applications.

    You have some great contacts here. It would be a struggle to justify a trip to the US for training purposes, but if anyone were interested in opening a UK office that would be a different story. I can’t offer mountains and a microbrewery but I do have some great homebrewed cider.


Leave a Reply