In this post, we go over another complete Six Sigma project example to help you understand how to use our patented state of the art QuikSigma program to help you solve the big problems plaguing your business. You can purchase your QuikSigma licence here. If you have any questions about this video tutorial, or specific questions about how to use your QuikSigma software program, leave a comment below!
First and foremost, Six Sigma is about results. Six Sigma is not about doing more work, it’s about doing the work we already have, better, and with less effort. It’s not about training. Training is only there to help us get the results. And finally, it’s not hard. Some people have over complicated Six Sigma, made it more difficult than it is. At its core, it’s very simple, it’s very direct, and very powerful.
In this video, we’re going to look at a complete project from start to finish. This is not a complicated project, this is to show you just how simple a project can be. We’re going to use the example that we created for the videos on chartering the project, I’ve made a couple of minor changes to it, but we’ve got our team lined up. The problem basically is that we’re getting 70,000 customer service requests a week and every time they talk to a technician it costs us money. We’d like to drive more of that business to self-service on the web because it’s much much less expensive. And right now we’re running at about eleven percent. We’d like to get that up to sixteen percent. We estimate that will save us a little over nine hundred thousand dollars, and our
justification is here.
By the way, these are not quite actual numbers, but this is very similar to a project that I once helped coach. So we got our charter done, it’s checked off, we’ll sign it off, and that’s done, and we go to measurement systems analysis. Here we’ve got one of those special cases. We’d much rather know in advance that our measurement system is good enough to detect change. Unfortunately, there just isn’t any way, that I can think of, to do that on this project. So what am I going to do? Give up the project? No. The rationalization is this, in the end if you can detect change, your measurement system is good enough.
So, we go down and in our baseline situation, we’re going to expand, let me get the right one up here, we’re going to expand on our statement that we’re running at about eleven percent self-service. Well how do we get that number? Every one hundredth customer on the web we just pop up a little survey say would you answer a couple of questions for us? And the question is, did you come here looking for technical help? Yes or no. Did you find your answer? Yes or no, and then another very simple question that we ask about, how did you find the ease of navigation? Give us a score 1-10. That’s our measurement and we’ve got tabs down here, you’ll see how we use those in the future.
But this is our baseline condition. Were at eleven percent. Process appears to be stable and predictable, but completely unsatisfactory. The next step is to go down to the process map and so the title is getting technical answers via the web and the CTQs are that the answers are easy to get, they’re correct, and in the customers eyes, it’s best if they’re not expensive. So we break this down into logical blocks of activity. One is where the customer identifies the problem. Then he searches for the answer and then he implements the answer and we see if it works. And where we got our KPOVs and here the input variables, that I could think of, that influence this. Customer level of knowledge and the type of problem. Mmm, yeah, those are variables. I doubt that i can do much about them, but I’m going to list them. Does the customer have web access? The organization of the website, the correctness, the understandability of the site documents, whether the documents cover all the needed topics, and the accessibility of voice support.
Down here again I’m going to list technical level of competence and the accuracy of the answer given. Carry those over into the C&E matrix. So I’ve listed everything that i can think of, and down here I’ve given the answer is easy to get. A 9 for a weight. The answer being correct as the most important, probably, I gave that at 10, and the answers inexpensive for the customer, maybe that’s the 7. Now, I filled in the matrix. I’ve given it nines, threes, ones, or blanks and then I click calculate. It sorts for me, and tells me that the important variables to pay attention to are the understandability of the site documents, that getting all the needed topics covered, the correctness of the documents, and how well organized the site is.
All of those go forward to the FMEA. So here are my four items that I decided to carry forward, and I examine those in detail. Severity is usually an eight or a nine because in this game the worst thing that can happen is you got a disgruntled customer who’s not getting the answer that he needs. So I go through and fill that out and I fill out my proposed action items. The ones that need to be carried forward, I carry forward, many of them the answer is none required.
We’re still pushing things off our plate, getting rid of the less important things, so we can pay attention to the very important things. So then the things that I have flagged, let me move that over for you, oops doesn’t want to go. C’mon. The things that I have flagged then carry forward and I’ve got my RPM number carried forward. My action automatically carries forward. I can state my reasoning. Who did I give that to? What the due date is? And as i get these done i can just simply check them off. This is my laundry list of things that need to be done. Now out of this, I’ve selected all three of these to carry forward to my control plan.
So here’s my control plan. These two columns will be populated. Here’s a very typical error indicator. We’re going to keep a control chart. If it goes out of limited and unfavorable direction, we’re going to investigate and find out what went wrong and fix it. Now every project does an experiment. In the simple case, one like this, the experiment is usually the old system versus the new system and what we are obliged to do, is we are obliged to show, that we made an improvement that cannot be easily accounted for by random chance. So all I’m going to do then, as this is my own system data, I’m going to get my new system data, and put that down below here. Then I’m going to split my chart. This is how we were running. This is how we’re running now. You really don’t need a whole lot of statistics to tell you what’s going on. This is better. Now we’ve also got an a weekly average of our ease of use of navigation score. We have the the question on a scale of one to ten was the navigation easy? We look for the shift. This is the old system. This is the new system. Most of these points fall outside the upper limit. Very easy to assert and win that you have made a difference. IF you want to be a show-off, you can put it in a t-test, and you can look at p-values, but i guarantee you, the behavior chart version that I just showed you, communicates better and is more convincing. That’s as simple as it needs to be.
So let’s count up the the tools we used. Every project should start with a good charter and we did that. When we got to measurement systems analysis not much we can do. Under baseline we used to control charts or process behavior charts, and I also quietly slipped capability and on one of those and didn’t really call your attention to it. We use the process map, C&E, FMEA, action plan, we use the t-test and I pointed out we didn’t really need to do that. That was just kind of showing off, and as every good project should, we ended up with a control plan. So what’s that? Nine tools that we used. One of which we were not required to do. Yeah, and in real life this project actually did save about that much money.