Taguchi with multiple responsesA user is interested in conducting a Taguchi experiment with multiple responses. I know little about Taguchi designs, but I do know that JMP's Taguchi Array supports only a single response. A quick search turned up some work on multiple-response Taguchi designs, but this doesn't seem too common. What might we be able to suggest to this user?
Re: Taguchi with multiple responsesWe might suggest that they try living in the 21st Century. 
Sorry, I don't have any better suggestions, only facetious ones. I know almost noting about Taguchi designs. They were already a relic by the time I started learning about DOE.
Re: Taguchi with multiple responsesI would try to get them to articulate what questions they wish to answer by using Taguchi. With that information it is very likely we have a more efficient and effective alternative. It may also be that they consider Taguchi as a synonym for the more efficient and effective alternative.
Re: Taguchi with multiple responsesTaguchi orthogonal arrays were my entry point to modern experiment designs. At the time I learned about them, I had found limitations in classical designs and was looking for a better alternative. A colleague and I decided to give them a try. The results were no better than playing darts blindfolded. Frustrating for both of us as the experiment was quite costly in both time and expense. I spent several weeks/months trying to figure out what went wrong and whether there were better alternatives. In my research I found a few harsh critiques of Taguchi and then stumbled on optimal experiment designs and a well versed mentor to learn from. I found optimal designs that were adaptable to the experimental variables and constraints I encountered regularly. Long story short, I lost interest in figuring out how to make Taguchi designs work and/or proving Taguchi wrong and made a very profitable move to modern experiment designs. Incidentally, one of my favorite Doug Montgomery quotes from a Discovery keynote talk is "Taguchi got the question right". I like this as Taguchi made significant contributions even if his experiment designs are inferior to modern designs. Echoing @malcolm_moore1 , might be worth learning where the customer is in their journey and what problem they are trying to solve and guide them to modern designs in JMP.
Re: Taguchi with multiple responsesThanks, all. Sounds like there just isn't a good reason to use a Taguchi design over modern approaches. Having read up a little more on the design, I can see why it is considered inefficient, though "Taguchi got the question right" seems appropriate, too. I'll see what additional information I can get from this customer and find whether the problem might be better addressed by an alternative design.
