Design of Experiments (DOE) in the Service Industry?

Six Sigma – iSixSigma Forums Industries Hospitality Design of Experiments (DOE) in the Service Industry?

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
  • #55341

    Rory O’Brien

    I am new to Six Sigma and have been tasked with doing an assignment on DOE in the service industry. The thing is, I have hit a mental block. I cannot think of any levels or factors that are measurable?! In manufacturing, this is easy as you can set equipment to high/low, etc and have a multitude of factors that can be affected by the levels.

    How can this be applied to the service industry such as a hotel or restaurant?


    Robert Butler

    The first thing to do is decide what it is that you want to investigate. Your post is too generic for anyone to offer anything of substance.

    What kind of a service industry?

    What aspect of that industry are you thinking about investigating?

    How do you know that the aspect you have chosen is really an issue?


    Something to consider. Before I do anything with respect to service improvement I make sure I, and everyone I’m working with, really understands the problem and what it is that we want to improve. This usually means a lot of work on my end. Specifically, meeting with people involved in various parts of the provided service and, with their help, assembling a fishbone diagram of their part of the service. Once I have all of the pieces I call a meeting. Before the meeting starts, I put the fishbone diagrams on the wall of the conference room in the order everyone understands to be the process flow.

    with the players in the process present, I talk everyone through the process. I make sure everyone understands it is ok to interrupt me at any point in my description to offer corrections/additions/etc. For me, more often than not, this exercise not only identifies the REAL problem but offers variables/current practices that one might want to consider in either in the framework of a design or as something that is obviously in need of correction/modification without my having to do anything more.

    In those cases where we have run a design it has always been a simple blocked main effects design. The variables have been things like new vs. old way of work documentation (testing proposed streamlining of paperwork to avoid duplication, presenting fill-in-the-blanks in the order they occur in practice as opposed to whatever it is that we are currently doing, trying different types of staffing with respect to addressing customer problems, etc).

    In my experience, building the fishbone and talking the process usually results in the identification of obvious problems in need of correction. More often than not the type of correction needed is already known and all that remains to do is make appropriate changes in the process.


    Rory O’Brien

    Hi Robert,
    Thanks for your detailed response. I should have been clearer of what I meant in my original post. Basically we have been asked to complete a DOE assignment in the service industry. We have not been given any problems but rather we are to make up our own types of a very simplistic design experiment.

    My only problem is that I cannot think of how DOE can be applied to a service industry or hospitality industry. My initial idea was to perform a DOE in a barista or coffee shop testing the different brands of coffee beans, etc.



    If you were actually doing DMAIC you would have defined, measured, and analyzed a problem before considering a DOE. And you would have a testable hypothesis about how to improve a specific problem. The fact that it’s in a service industry makes little difference.

    If I’m correct and you are just grasping at straws to do an exercise you will need to make some assumptions about what you would have found during Define/Measure/Analyze. For instance, assuming that coffee is always prepared the same way so the brand of coffee bean is the only variable, and assuming that customers have a specific and consistent idea of how a good cup of coffee should taste, how would you design an experiment to determine which bean is better?

    It’s been stated many times in this forum that if you expect us to help you with an exercise you should explain what you think is the solution and why you have questions about it rather than ask us to just give you an answer. The more work you do on your own the more likely we are to be able to help.


    Norbert Feher

    This is my favourite and after going through the list You will have a couple of ideas:-) :



    @Nfeher, thank you for an interesting link.

    @RoryOBrien I agree with the other comments. DoE helps solve problems and you must have a problem before even deciding if DoE is the right tool for it.

    I have always found the minitab blog a very interesting place for using statistics in many diverse applications- magic to weight loss- it is really fun. maybe the ideas there would help you:

    All the best.

Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.