Lean Six Sigma Green Belt – Six Sigma Control Phase Part 4

7. SPC – Control Charts – Part 4, C Chart

Your quality team inspects manufactured shoes against the checklist. There are seven opportunities for error in each transaction. 90 shoes are inspected every day and the number of errors were recorded. Number of errors are recorded. Comment about the process. Is a process in control or not? First of all, which control chart are we going to use here? Go back here. It’s a defects, right? Because they’re discussing about the opportunities for error and the number of errors are recorded.

So there’s no mention of defective here. So we know that they are defects. Is a sample size fixed there? Oh, yes. Why? Because you daily pick up nine issues, so the sample size is fixed. Obviously, if the sample size is fixed, then you go with which? CChat? All right, you go with CChat.

So let us go back to the magic box. Let us go to the shoe manufacturing. Daily you pick up 30 and you try to identify the errors within that. Hey, here there are 90 shoes and how can there be one outfit defects. Errors. That is because each shoe has seven opportunities for error. So if I have 90 hues, it will be 19 into seven opportunities for error. It will be 630 opportunities for error in total, right? 19 into seven. Out of 630 opportunities for error, there are one out for errors. That is what it means. Yeah. So don’t get confused, folks.

Let us do the C chat now. Go to stat control charts. Yeah, go to attribute chart. We are still in attributes. And click on CChat there. What do you want to plot? I want to plot the number of errors and go to see chart options there. Go to test for attribute tests. You have four tests, so do a drop down. Perform all tests for special causes. Click on OK and okay. Now let us see whether the process is in control or not. Wow. We got the control limits. Lower and upper control limit. You have the count or average of the count here we see that there are two outliers, right?

There is one outlier here. There is another outlier there. Both of these are violating test number one. Test number one. So what do you do next? You need to identify what is the test number one. How do you do that? Right. Just click on this icon there to go back to the previous edit window.

 Go to see chart options. Go to test. Here we have the four tests and we know that there are two data points which are violating the first test. What is the first test all about? Out there is one point which is greater than three standard deviations. There’s one point greater than three standard deviations from the center line. Wow. So that’s violating that. You analyze, you try to identify the root cause and is it a genuine case of being an outlier or not? You try to identify that here.

8. SPC – Control Charts – Part 5, X bar S Chart

Here we have a case study of soft rings. What does this case study say? A soft rings company wants to assess the stability of bottle filling process. The fill height of bottles at the same temperature are inspected every week. Comment about the process. So you are measuring what height of the bottles fill height. Height is what? Continuous variable. Right. Height is a continuous variable. All right. How many bottles do we have? Or how many did I pick? We’ll come to know once we see the case study there. So we know that it’s a variable data. And let us see the sample size there in the Minitap file and comment on what we are going to do here. We have picked up ten bottles, right? We have picked up ten bottles. So the sample size is greater than or equal to ten here.

So you go with x bar S chart. X bar s chart. Right? So let us do that. Go back to your Mini tab. So our frank’s example go to stat control charts. Go to variable charts. For subgroup we need to do x bar S. Click on that are all observations for the chart in one column is select all those subgroup sizes. The weeks, I believe, right? Go to x bar s there. Click on test. Click a drop down. We want to perform all tests. What is the difference that you have noticed here? For your attribute test attribute charts, you have four tests. And for your variable you have eight tests. You have eight tests. Four more got added. The value of K can be changed to suit your requirements if you go with Western Electric rules. That is your project management professional, right? In that the control charts use a rule of seven. Seven data points falling on one side of the mean of the center line.

But here we are following what? Nielsen rules. Nielsen rules. Hence you see the difference. Okay, let me click on a K and a K. Let’s see what happens here. There was an error. So now it says that the options which you have selected are absolutely wrong. So there is a reason why you have done it in that way. Because Minitab is not a dumb tool where you put whatever values you wish to, and it’s going to throw you a control chart. For that you have to be judicial. And Minitab knows a few things, right? Even if you go wrong, you can make the corrections. So let us go back to the edit window, right? What did we do? Let me go back to the worksheet. Also here is a worksheet for your full height. If you go to stat control chats tab groups, and if you click on x bar S there, this is the option that we have selected. Right?

All observations for a chart are in one column. How did you even believe that? We have column 1234-5678, 910, right? So we have different columns. Observations for a subgroup are in a row of columns or in one row of the column. Yeah. So now you select all the bottles from one till ten. Go to Xbors. Go to test. We want to do all the tests here. Click on OK and okay, it’s coming up. It’s populating the control chart for us now. Here we go. Both are good. This is for your mean X bar and this is for your standard deviation. You’re looking into both of these in tandem. First you look into your X bar chart, a sample mean chart. You see that there is one outlier here and it is violating test number five. Now we need to find out what the test number five is all about. Let us go back and check what the test number five sees. Go to test. The fifth test says k out of k plus one points is greater than two standard deviation from the center line. What? Is K here? Two.

So two out of three points are greater than two standard deviations from the center line. Two out of three points are greater than two standard deviations from the center line. So in that way you identify what is an outlier, which test is it violating, and how to counter that. All right, now we have one last test which we need to perform the loan application. One last case study which we have to crack now. Loan Application the loan application processing time here itself.

The thought should click you right. It should strike you and say, hey, we are speaking about continuous data, variable data here. Loan application processing time is recorded for one loan application every day to determine whether the process is in control or not. Comment about the process. There is only one sample, right? One loan application per day. Only one sample. And we know that it’s variable data. Variable data and sample size is equal to one. So you perform IMR chat individual moving rainchat.

Let us go. Let us go there. Let us open the loan Application case study and solve it using your on a daily basis. We pick up one application and we check the loan processing time. Probably this is in minutes. Hopefully it’s not in hours or days, right? It would be in minutes, right? That’s the assumption. So now let us plot the chart, which is individual Moving Rain chat. So we are now doing variable charts for individuals, which is IMR Chat. We click on that, we select the loan processing time. We go to IMR options, go to Tests. Select all the tests all eighties.

Remember, for attribute data we had four tests. For continuous variable data, we have eightis. Click on OK. Click on OK. Let us see what the output turns out to be. Wow. Individual Value Chat does not have any outliers here, right? But your moving range chat has an outlier here. What is that? Outlier there are two points, three and three. They’re violating the test three and three there. Yeah. So what do we do now? We need to go back and check what the test three means. Look at that.

K points in a row all increasing or decreasing. So it’s a six. K is six. So six points in a row all increasing or decreasing. Is that right? Let us count. 12345. Yes. Six data points are continuously decreasing. Hence it is an outlier. The next one is also decreasing. Another outlier. In that way, you try to identify what is an outlier. Why is that an outlier? Yeah. And there we go. This is the way we monitor our process on an ongoing basis.

9. Control Charts Extension & Solution Documentation

Control charts. Interpretation is not limited to observing violation of the control limits only, right? While a point outside the control limit is a good indicator of non randomness, other patterns are possible. Also you can see a cyclic trend decrease, increase, decrease and increase, decrease and increase. That is a cyclic trend and periodic signal. That is also an indication of probably an outlier or probably something going wrong. Minutes to work days of week, right? How much time do you take to arrive at office on Monday? You take more time. Why is that too? Because after Sunday, right? A lot of boozing. You actually wake up a little late. Not my beer friend.

On all Mondays you go late to the office. Time taken to reach the office is more maybe there can be a mixture of two or more sources. You can see probably this kind of a cyclic crane versus you can see a shift also, which is abrupt change. Look at that. Here is a process. Suddenly you see an abrupt change and then a dip in share market. You can see this kind of situation also. So if you see one or two of these trends, it’s mixed years, then we also see a trend gradual change. Look at that. Gradually it’s increasing gradually, right? That is called as a trend. Increasing or decreasing gradually is also a trend that you need to focus upon. That also can be a violation, right? We don’t know until unless we analyze, then comes a stratification. If the variability in the process is too small, it might be difficult for you to identify the variability, right? So if you stratify the data further, probably it might help. We don’t know.

All right, so we need to analyze control charts with thorough importance. Out of control points should be explained and excluded. If you feel that that is an outlier, identify that it is an outlier, analyze it, conclude that it is truly an outlier and then exclude it from your data while you plot the control chance. Leave it as is in the graph. If you cannot explain it. If you feel that no, I think that is business as usual. I’m really not sure why that outlier has occurred. Let it be in the graph and proceed with the further monitoring of the process. Interpreting the Tests we have done that, right? We know that one to four tests are for your attribute data and for your variable data. You have eight tests. That is the importance here. All tests are used for mean of continuous data if you have attribute data or for the variationrelated charts like your R chart, S chart or moving green charts, right?

Only these four would be applicable. So if you’re plotting IMR chat, if you are plotting IMR chat, for example, for this moving range, only these four tests would be applicable. However, for your individual chat, all these eight would be applicable. Though you are combining these two even within that continuous part is applicable for few variable part is applicable for variation thing. For variation things, footage are applicable. All right? That is how you interpret data. Once you’re done with that, document the solution and the benefits that you have attained. Now, appropriate documentation of solution is important because people want to review on whether you have truly brought out the business value by implementing Six Sigma or not. And also you want to replicate the best things which you have done in the past, right. For these reasons, you need to document the benefits and solution the results of the improvement, the financial benefits that you say your process would attain, right. Or your organization would attain as part of your process improvement.

You mentioned that because it has to be monitored for running anomaly. So the moment you implement Six Sigma project, you are not going to realize the benefits immediately. However, it might take an hour or so for monitoring and checking whether your promises have been actually realized or not. People have to constantly monitor that. And for that reason, you need to look into your financial benefits and improvement and monitor them on an ongoing basis for one year. Finally, you need to hand over the project to the project owner or the process owner, right? Hand it over. Hand over the control plan which you have prepared to the process owner. Get a formal sign off by the sponsor saying that hey, I have successfully implemented the Six Sigma project. You’ll realize the benefits within a year or so and here is my report. Get a formal sign off, right?

And process owner, you need to get a sign off from him also him or her process owner on the results of the project. He also has to say, yeah, understand you have implemented the Six Sigma project and I agree with the results. Last but not the least, identify all the best practices and replication opportunities you have done the Six Sigma project. Maybe in a particular department, you want to probably implement that in other departments also. So good to document, identify the best practices and you have to start replicating everything and also advise the process owner on the same that you have identified the best practices. And then he needs to look into the replication opportunities. Finally, we come to the outputs.

The key output of control phase is control plan. However, there are a lot of other outputs also, right? But in control plan you have the documentation of the project findings. Along with that, you also have training and communication plan, which is another output of the control phase. And you transfer the improved process to the process owner. Validation of business financial benefits right, would be done by a financial team or by a sponsor. They’re going to monitor it on an ongoing basis. Here is the last summary. We created a control plan to monitor and sustain the improved performance. A control plan would typically contain your audit plans, risk response plans, statistical process measures, visual controls and poke AOkay measures, right? It’s in that previous order.

First, try to mistake proof. If not, try to use your visual controls. To some extent it’s going to address the issue or use your process control for ongoing come up with your risk response plan. And the last but not the least, or I would say it is the least also audits and inspections least preferred, but sometimes you tend to do that. We also discuss about the control charts which detects the presence of the special cost and it alerts us. Control charts don’t ask capability or conformance to specification, right? They only look into the process stability. Capability is measured by your Cpcpk values. Remember Sigma values there in the measure phase. Finally, controlled plan was shared with the process owner thoroughly disguised and the process is handed over to the process owner.

Now the process owner diligently follows the control plan to sustain the performance. And with that, we close the Six Sigma project we have worked in to define, measure, analyze, improvement, control and we are done with the project. Cheers to you guys. What next? Thank you by the way for patiently listening to my session. Hope it was interesting and hope it was helpful for you all. What next? Tune into our Lean Six Sigma Black belt. You learn about the advanced concepts, right? In the major phase, you have learned only about CP and CPK. Here you learn about PP and PPK also there you have learned about Sigma.

Here you’ll learn about Sigma entitlement Z, basically z value, entitlement z goal and things like that. In measurement system analysis, we have looked into attribute agreement analysis in greenbuilt. In Black Belt we are going to look into gauge RnR, right? And there are a lot of control limits, how to calculate the control limits, how to do logistic regression, how to do multilinear regression. We learn about a bunch of advanced techniques as part of your Lean Six Sigma Black Belt. Once you are certified on lean Six, sigma Black belt move on. Do not wait. Do your mask black Belt which has DMADV concepts, define, measure, analyze, design and validate DMADV, right? Design for Six Sigma Process.

That’s it about mass black belt. If you want to further enhance or hone your skills and statistics, look into our courses on business analytics. We provide a plethora of opportunities by catering to almost all the business analytics courses, right? From statistical analysis to your forecasting, to your data mining, supervised and unsupervised data visualization, financial analytics, simulation, data optimization all these are taught to you, right? So all the best for your future session. Thank you so much for attending the session now. See you once again in lean six Sigma black belt. Thank you.