Lean Six Sigma Green Belt – Six Sigma Control Phase

1. Introduction to Control Phase

Hello, and welcome to this session of Lean Six Sigma Greenbuilt control fees. This is the last phase of DMAIC and we are into the control phase. As usual, we know that this session is driven using Minitab. In this particular session, we will look into control charts using Minitap. We will look into the basic concepts, theoretical concepts, and then we are going to discuss the case studies on which control chart to use, given a particular situation.

Why did I put this image here? We are looking probably it is a fly, right? I wish to say that grassrophila fly eggs are measured in the canned orange juice, citrus juices. Is that true? Oh, yes. How do they monitor that? Using something called as control charts, we will get into the depth of the control chart, who will discuss about various concepts there. Just to give you a brief about this, let me explain. This control chart which is present here. Upper specification limit and lower specification limit are provided by the customer end user.

And these specification limits will not change. They will remain the same. And your control limits, upper control limit and lower control limit. These limits are calculated based on the data points which get generated in these black dots. These are your data points, right? And here we have the mean. Let us understand this using cricket ball example, right? And later on you can relate to how insect fragments are measured in citrus juices. How is the maximum content of your Drosseful hour fly eggs within citrus juices? Can citrus juices is measured? All right.

Let us understand this example using cricket. Cricket is a game which is played and it’s extremely popular in India and it’s popular across the globe as well, right? So cricket, if you all are not aware, I assume that it is played with the ball, right? Which is called as ICC cricket ball. Right? And ICC stands for International Cricket Council. This International Cricket Council provides you the limits saying that, hey, no cricket ball circumference should be greater than 229 mm. No cricket ball circumference should be lower than 224 millimeters.

Right? That’s the unit there. Why do they provide these kind of limits? Because if they do not provide sue, manufacturers might manufacture a cricket ball whose Circumference might be 250. If it is 250, you end up playing cricket. Not with a cricket ball. You end up playing cricket probably with a volleyball, right? A ball which is bigger in size in comparison to your cricket ball. Why do we have these lower specification limits here?

Because if International Cricket Council does not give these limits to you, you might manufacture a ball which is the size of a table tennis ball and say, hey, here is the cricket ball, go play cricket with that. Right? So if you are not going to provide these kind of specification limits to the manufacturers of the cricket ball, there is no guideline, there is no way they can manufacture the cricket balls. Hence, we have these kind of limits set. There might be control charts with only your upper specification limit. As in case of these fly eggs, right, there should not be more than five Drosphila and other fly eggs per 250 ML of canned orange juice.

250 ML of canned orange juice. That is measured according to Food and Drug Administration US. FDA. All right. So that is upper specification limit which is set. Lower specification limit would obviously be zero. You can have such situations also or in 250 ML of canned citrus juice, there should not be more than one. My God. My God is an insect, right? A way of looking insects, not more than one. These kind of limits are set to ensure that my bad once again.

So we can just have the upper specification limit, as in the case of your Rossoville fly eggs or Megat, which is measured in 250 ML of citrus can juice. Or on the other side you might have only the lower specification limit without the upper specification limit as in the case of productivity. If you were to measure productivity, do I have an upper limit on that? I most often than not do not like to have an upper limit on my productivity.

I do not want to keep a tab on that higher the productivity, better it is for me probably right, probably. So in some cases you have only lower specification limits. In some cases you can have only upper specification limits. And in some cases, or most often than not, you will have both upper specification limit and the lower specification limit. Let us take another example.

Say I want to monitor my effort variance, right? And say my organization gives me this goal which is upper specification limit is 10% and lower specification limit. Is -10% what does that mean? This means that for each task, say task one, I have the planned effort planned effort could be say for task one, say it is 100 hours. For task two, say it is 120 hours things like that this plus or -10% upper and lower specification limits say that this task should not take more than 110 hours, right? Or it should not take less than 90 hours. How did we get this 110 it is plus 10% hundred plus 10% of hundred would take the count to 110.

So you start monitoring against this upper specification limit, right? For task one how did we get this 90 here it is 100 hours -10% which would be or -10% of 100 which would be 90. I need to complete this task one. Within these limits though the planned effort is 100 hours. I’m not going to be alarmed right? I will not be confused, I will not take any action. If the actual effort turns out to be any number between these limits. If it is greater than 110, say you were supposed to complete this task one in 100 hours. However, you have taken 115 hours.

That is an alarm for me. I’ve exceeded the upper specification limit. I would go back and analyze what went wrong. Why did I spend so much of original effort as opposed to the planned effort of 100 hours? And then I try to identify the root cause using fishbone and five y analysis and try to address the root cause. The scenario is okay. If I overspend, I’m going to do that. However, think about the lower specification limit, which is 90 hours.

Here. If I complete a task which was planned for 800 hours, if I complete this task in 80 hours, effectively speaking, I’ve saved 20 hours, right? I should have completed the task in 100 hours. That was a planned effort. However, I’ve taken only 80 hours. Why do I need to even monitor that? Why should even I be concerned about that? Why should I be alarmed about this?

I should be alarmed because probably I might have compromised in the testing, or my estimations might be wrong. My plan effort might be a wrong estimation. Or maybe I have missed out on a particular feature or functionality. Anything is possible, right? So you need to analyze it to identify whether this is really a bad situation or is it a good situation wherein you have come up with some reusable component.

You have done some process improvement, which is ensure that your 100 hours of plan task was completed in 80 hours. Don’t you want to document the best practice? You want to do that, right? Because we want to replicate the best practices. That is the reason why we have upper specification limit and the lower specification limit in most cases. And remember, your specification limits are provided by the customer. That is fine, but what are these numbers control limits? Control limits are statistically calculated based on the data points that you enter. These are the data points that you enter. Let us go back to the cricket ball example now and understand the control limits.

We know that the upper specification limit for manufacturing the cricket ball is 229 mm and the lower specification limit is 224, right? These are the specification limits given to me by my customer. Say I’ve manufactured the first cricket ball and its Circumference, say, is 226 mm. You’ll just jot it down on the control chart. If this is 224, if this line is 224, and if this line is 229, then 226 would fall somewhere here. Say the next cricket ball that you have manufactured, Circumference, is 227. 5 m, mainly meters, right? Then you plot another point on this control chart. Say the next cricket ball that you have manufactured. See the Circumference of that was 226. 5 mm.

You plot the third data point, so on and so forth. As in how you manufacture the ICC cricket ball, as in how you get the Circumference of that. You jot it down on the control chat and record it. The moment you have n number of data points, these control limits are automatically statistically calculated. Right? The moment you plot another data point. Here again, your control limits might tend to change because they are statistically calculated. They are dynamic in nature.

All right? So this is how it gets calculated. Now we’ll do a deep dive. Also, I’ve just given you a highlevel overview of what control charts are because we would be spending some time analyzing that. And we’ll also understand some of the terminologies within this. Okay, going back to our equation, y is equal to f of x, right? In the defined phase, we have defined what is the output which we want to achieve.

In the measure phase, we have measured the current status of your output, what you want to achieve. And then in analyze phase, we have identified all the potential x’s. And then we have shortlisted the critical inputs, critical excess which would impact your output. And then we have improved those inputs. Critical inputs. Now, our job is to monitor and control the improvements which we have brought in. For that, we are going to get into the control fees wherein we monitor and control both our inputs and also our outputs. You can monitor both of those. Control both of those. All right. Now let us get into the azure.

2. Control Critical Inputs

How to control the critical inputs. First things first, we need to come up with the overall strategy for improving the process performance over time. Right? We need to come up with a strategy for maintaining the improved process. You have improved the process does not mean you are going to leave it aside. Remember, we have discussed in a major phase that the improvements that you bring in now, no matter even if it is six sigma process, if it is left unattended, or if you do not do any continuous optimization or improvement on that, your sigma level is tend to grip by 1. 5 sigma, right? Say you have a process after improvement it is at six sigma.

Assume if you do not do any improvements on that, it’s going to drop down and how much will it get degraded by 1. 5 sigma? And this is statistically proven, right? Hence, we need to always be on tools. We need to improve the improved process. We need to maintain the improvement and also try to improve it further.

That is our objective. What are the objectives of the control system in that case? We want to identify the specific actions and tools required for sustaining the process improvements. Against that is the first thing. What should I do? What actions am I supposed to take? What tools am I supposed to use for sustaining the process improvements which we have brought in with such a great difficulty? Right? We have literally spent a lot of effort in analyzing the data, crunching the data, bringing business insights from the data, right? If you have spent so much of effort, it is worth spending some time in coming up with the specific actions and tools which are required to sustain the improvements of gains which we have brought in. We also want to ensure that a process stays in control. I’ve implemented a solution. Does it mean I’m going to take a backseat?

Does not mean I’ll be laid back. Instead, I’ll ensure that this improved process stays within the control always. That is extremely important. If there is any out of control situation, we should be able to quickly detect that and take an appropriate action. Right? We should be able to identify what is a special thing which has happened because of which a process was out of control, special causes.

Anything which falls outside your control limits or specification limits for that case would be your special case. Because you can assign a special cost to that and why it went out of control. All right. Here is a typical control plan and these are the various components of a control plan. You would have a data collection plan. We need to have mistake proofing system which we are going to discuss in detail.

 We need to address the risks, not just mitigate the risk, but we also have to look into various other mechanisms of coming up with the risk response strategies for both your positive and negative risks, by the way. Right? You know that risk is any answer at any point which might occur in the future which might have negative or positive impact on that. If mistake proofing is not possible, then we’ll come up with statistical process control as part of your control chat.

That is the next best solution. The first best solution is mistake proofing. Do not let people make mistakes even if they want to. The next known solution is your statistical process control using your control chats. If that is also not possible, probably you’ll resort to mechanism of conducting the audits or inspection. By the way, that is the least preferred process ownership. Who is supposed to own the process owner, right? What is the response or reaction plan if something goes wrong?

And then we also document the process. So these are the various components of your Control Plan. Now, let us understand few more things about this Control Plan. Here we go. Control Plan is not a replacement for your detailed instructions. If you have standard operating procedures for your various departments, control Plan is not a replacement for that.

It’s just a supplement. It is documented description of systems which are used in minimizing the process and service variation. It documents all the systems it documents your processes to ensure that the variation in the process of the service which you’re providing is to the bare minimum. It describes the actions that are required at each phase of the process, right? It describes the actions for each of your inputs, each of your outputs and periodic requirements to assure that all process outputs will be in the state of control.

All Cricut balls that are manufacturing should meet the specification limits. The circumference of all cricket balls that are manufactured should be within 229 to 224. That is all about that. Control Plan is a living document. What do you mean by a living document? Is it going to work now? No, living document means this is not a static document which is written once and does not get updated anytime, right? Control plan gets updated throughout the project of Six Sigma.

Even after that, if you have any changes, if you have any good suggestions, good improvements which you bring in even after the implementation of Six Sigma, you’re free to do that. It contains few of the things such as what process is being controlled, what measures are being monitored, right? Are there any trigger points for each measure? If he has, what are those? What actions are you going to take when there is a trigger point? And last but not the least, who is responsible for taking action? Who will take the action? That is most important part here. Next thing is your control plan. Checklist. You need to have a checklist in place and ensure that your Control Plan is complete.

How do you do that? And what are the key components of your Control Plan? Here the first thing is process maps, right? Because now the process has changed. So the change the process map, the various process steps, information flow and the important variables should be part of your checklist. Key service variables. Key variables which you have identified by giving more importance to your customer, by giving importance to the desired target value and the specification range. You need to document that, right? That’s the second checkpoint. We all have discussed already about capability indices, right? Cpcpk, right. Long term, short term, z values. We have discussed about all these things. These trends shots are required to track the variation reduction progress.

You document the critical input variables, right? Which are identified with the relevant targets against each of those, right? Statistically determined control limits. When you plot the control chart, you would get the statistically calculated control limits. And what are the control strategies that you have defined there? You also come up with a reaction plan. If processes out of specification, out of control, what would you do? How would you react?

Are you going to do an analysis on that to identify the root cause? Yes, you’ll do that, right? So you document that. It’s as simple as that. Have a checklist in place. We have already done the measurement system analysis even before we started the analyze phase, as part of a major phase. However, now the process has changed. You have brought in the improvements, post improvements. Is your measurement system still capable of measuring accurately? You do that.

You need to have a checklist there. How do you do sampling? Are you going to use a simple random sampling or will you go with strategic sampling? How are you going to do the inspection? Are you going to inspect the entire population? Maybe? If very few things are manufactured, you inspect the entire population, maybe, right? But if there are a lot of things which are getting manufactured, for example, if you are manufacturing thousand shirts per day, inspecting all the thousand shirts does not make sense. It’s not worthy of investing so much effort there. So what do you want to inspect in that case? Probably just 10% of the total 1000 shirts which are manufactured.

You need to have a checkpoint against that and testing plans, right. How are you going to test everything? Also? How often are you going to do the inspection? On a daily basis? Or would you do the inspection, sampling or testing on a weekly basis? Right. Where would you do that? At the customer site or at the manufacturing site? Or in your go down in your logistics, in your inventory store? Where are you going to do that and to whom are the results going to be reported? There should be someone, right? Some key stakeholder to whom you send the reports to.

You need to have that information also handy here. Standard operating procedures are going to identify the actions, responsibilities, roles and responsibilities. Primarily, are you going to do any maintenance on your current machines or things like that. If she has, what are the maintenance schedules and service segregation requirements? How are you going to segregate the service from your different operations, different projects and things like that? Training materials are extremely important, which are going to describe the aspects of process operation and responsibilities. Who is going to train?

How are you going to train? What is the process improvement that you have brought? And you need to train the people also, right? On the changed process. That is extremely creative. Process improvement efforts, which are fully documented and available for reference. That is also another important thing. And then you have your Control Plan, which is reviewed and updated on a quarterly basis. Or in few places they updated on a monthly basis.

In few organizations they updated on half yearly basis or yearly basis, depending on the amount of change that is brought into the process. Quarterly seems to be an app and a safer bit, right? And it resides in the operating area. You do not put your Control plant somewhere in your dryer, somewhere in your cabin, in it. It should be available for everyone for reference. Hence you put it in the operating area. That is all about your control plan. And needless to say, this is something which Knives Six Sigma Consultants would miss out upon. You need to document that a good Six Sigma.