In this lesson, we go over the QuikSigma Cause and Effect Matrix and how you can work with your team to move your Six Sigma project forward.
One of the primary things that we did in the process map was to collect a catalog of all the input variables that might influence the outcome of our process. Well, all input variables are not equally important and there are good mathematical reasons why you can concentrate on the important few and give relatively less attention to the unimportant many. In the cause-and-effect matrix, we’re going to start winnowing out the less important variables and that’s wonderful, because of all the thousands of things that you might pay attention to, we’re going to take most of them off your plate so that you don’t have to deal with it.
It’s only going to take us probably an hour or two to do that and we’re going to eliminate maybe eighty five percent of the possible tasks that we might undertake. So this is a great economy. Here in the cause-and-effect matrix, you’ll see that the CTQs have been carried forward from the process man, as have all of the key process input variables. This matrix, by the way, is very much the same as the house of quality, out of the quality function deployment. What remains for us to do is give each of these CTQs a weight from 1-10. I’m just going to pick some arbitrary numbers here for purposes of demonstration. Yes, you can have two of these have the same 2number. That’s all right. Then in all of the cells down here, we put one of four inputs. If I enter a 9, which I’m going to do here, that indicates that this input variable has a strong influence on good taste.
If I enter a three here, it means that the cook’s skill has a moderate influence on the correct weight. If I enter a one, it will mean that the cook’s skill has a very small influence on good texture. Well obviously that’s wrong, but we’re just entering numbers here for an example. If I leave it blank, it means that that variable has no influence on that CTQ. So when we click calculate, what will happen is that QuikSigma will multiply the cell times its header, cell times header, cell times header, and then add across and give you a score. What that really amounts to is this. The variables that have a strong influence on the important CTQs will rise to the top. Those that have less influence will sink to the bottom, so we’ll be able to cut off many of the variables that we really don’t need to carry forward.
Now, just so you understand the context in the FMEA, we’re going to eliminate more variables. If we skip the step that we’ve got here, we would end up with this big long, long, hard to do FMEA. If you invest the time to do this C&E upfront, it’s a breeze because it gets rid of about 90% of what you might have to do.
Now, I’ve gone back and populated this matrix with some more realistic numbers. What I need to do is click calculate and that will bring up some totals over here and as it happens, I’ve already been in fiddling with this and I’ve already done this, but how we get these little flags set is this. We just simply go up here and lasso those and then we right-click and we set the flag to disable it. What we’re saying then is these variables up here 1, 2, 3, 4, 5, 6, are the ones that we’re going to pay attention to and the rest of these don’t have enough influence to merit our attention right at the moment. So we’ve got cook’s skill, recipe, baking time, baking temperature, type of flower, and and mixing time as important variables that will now be carried forward to the FMEA.
One other cautionary note, filling out the matrix is definitely a team activity. You want people from multiple functions there so that you can get all views. The other thing that might help you is this, if you really get stuck, can you enter another number? Yes. There’s not much point in it, but if it gets you on to the next item, that’s fine. If you need to put a five or six in there, you can. If there is any doubt, if there’s discussion and you need to and you just can’t resolve that need to move on, put the higher number in. If somebody says it’s a three and somebody else says it’s a nine, you can’t resolve. Put the nine in. What that will do is it will increase the odds of it being carried forward to the FMEA and if it’s truly a week variable it’ll fall out there.