Vishwadeep Khatri

You are absolutely right Inderjeet. If we have an option we mostly prefer continuous data, especially in a critically important process.

Someone asked me recently if the Lean Six Sigma methods are really standardized globally. This seemed to be an incomplete question. I wanted to know about the intent. The revised question I got was - Can people get different results if they are using the same information and the same assumptions? My answer was - yes, and no. Methods are standardized to a large extent. People may or not get same results depending on the chosen method. The answer is yes for many statistical analyses for given data sets. Once a statistical method is chosen, the results obtained by different users making use of same data-sets are likely to be same. Let us consider some examples showing how such statistical methods are standardized by ISO. The list below highlights a few important ones. (there are many more) Let us assume you are trying to assess the average turn-around-time for a process using some sample data collected for a limited time period. There is an international standard for this and it is named ISO 2602:1980 If you wish compare average performance before and after a particular change, you may like to use ISO 3301:1975 If you want to know whether data capturing mechanism is reliable and shall provide trustworthy data, there is ISO 10012:2003. For detecting and treating out-liers in your data, there is ISO 16269-4:2010 For risk assessment and mitigation, there is ISO 31000:2009 Not all methods provide standard results The answer to the question is NO, if it comes to qualitative methods like Quality Function Deployment, Design FMEA and the like. No two teams are likely to get same results if they use the same methods. While using ideation techniques, the objective is not to get same standard answer but to generate multiple alternatives. You may notice that even such methods are standardized. For example QFD finds a mention under ISO 9000:2000 and Design FMEA is covered under TS 16949. Lean Six Sigma methods are largely standardized. However, methods being standardized does not necessarily mean that we are likely to get same results. The beauty of lean six sigma lies in diversity of available methods. It makes sense to learn more of these so as to consider the most suitable one's for a given situation.

ISO standards mostly deal with "WHAT" and not the "HOW". For example, ISO 9001 is a quality management system that prescribes what should a company do to meet the requirements, get certified and to be able to showcase to their customers that they have a quality assurance system in place. If the same company has to implement Lean Six Sigma, it shall have to carry out Lean Six Sigma projects. Each such project shall delve into "What specifically needs improvement" and "How can the improvement be brought" with the help of world's bet tried and tested methods. ISO and Lean Six Sigma are not comparable. ISO provides generic frameworks, requirements and guidelines, while LSS projects work with sharp focus on a specific problem or improvement area to bring business benefits. With ISO, companies get certified while with use of Lean Six Sigma, they seek better business results. While ISO does not guarantee business results, Lean Six Sigma has grown because of its ability to improve results. With ISO, if certification is the focus, it may not matter how it is achieved as long as it is sustained. But with Lean Six Sigma, success and sustenance is dependent on how well it is understood and driven.

Hi Guest, Thanks for bringing more views. The discussion I started was specifically about manually controlled traffic in a high density peak traffic situation. It is here that one has to wait and the situation creates a valid challenge. Point 2 raised by you does not apply. Point 3 suggests that traffic shall continue to crawl if you have small batches. Point 4 does not apply to the situation. Kindly review your response.

Well said, Nagendra. Should absolutely take into account the upstream and downstream flow. If signals are too close, small batches are the better bet. Regards, VK

Hi, Rajesh, thanks for sharing this. Sequencing of signals should be workable in several key areas. Regards, VK

As I wait at a traffic signal to turn green in peak hours, I have noticed something peculiar. Even the last vehicle crossing the intersection before signal change is seen to be crawling. In my opinion this should not be the case. The commuter sequence at the intersection mostly seems to go like this. You wait at the signal for the light to turn green. As it turns green, there is a time lag before your vehicle starts moving. As the turtle pace starts to improve, the signal is again red and you stop but do not turn off the engine as it is likely to turn green again soon. While there are automated traffic signal management systems in other parts of the world (that sense traffic density and react), India does not have them yet. For traffic management, each intersection has its own timer activated at different frequencies which might be altered based on time of the day. Traffic police person mans the important intersections and also decides if lights/ timers should be used at all. While I am not exposed much to the art and science of manual traffic management, my thinking for peak times with manual management without auto-timers goes like this. As traffic starts moving in one direction, it should be allowed to reach a speed of 30 kmph or so before the signals change. When the speed reaches 30 kmph, more number of vehicles pass the intersection per second. As the signal is switched, the moving traffic in the new direction should again be allowed to reach a 30 kmph speed. The traffic flow rate is finally dependent on average speed at the intersection. Even for a single commuter's experience, It is better to wait longer, and then move once and clear the signal in one motion rather than crawl and stop multiple times. People shall save fuel in this approach as they can turn the vehicle off and relax for some time. The views are based on normal circumstances. There might be several other factors to be taken into account like traffic congestion upstream or downstream, VIP movement etc. If you put on your lean six sigma hat, the above theory is similar to batch processing being recommended against single piece flow (in this case). Allowing small duration intermittent flows does not help the cause as the changeover time has its adverse impact. The simplest mistake proofing solution to ensure continuous flow is to eliminate signal switch and have flyovers everywhere. While flyovers may take their own sweet time to appear or may not be feasible at some places, there should be a good manual traffic management approach in use at important interrsections that create bottlenecks. Traffic policemen at intersection may have their own approach towards signal switching which I am unaware about, but I do observe often that the speed of moving traffic at the time of signal change is very low. I invite your views and participation on this subject. Do you also feel the same way as I do for manually controlled traffic intersections? Does this situation seem to have any parallels with organizational processes? Benchmark Six Sigma is the best in class provider of Lean Six Sigma and Project Management training and consulting solutions. Have a look at why benchmark six sigma to know more about us.

Another location for healthcare case studies is http://globalnews.benchmarksixsigma.com/category/green-belt-qualified/example-green-belt-projects-health-care/

Good question, Inderjeet. Generally, DPM (defects per million) is a preferred approach as PPM (Parts per million) corresponds to defectives per million. Defects data is considered more useful and powerful for process owners. At the same time customers may be concerned about PPM, as it is the defectives that affect them. If you look around in automotive component manufacturing, many companies initially target 100 PPM for a part that has not yet reached a high process maturity level. It is important to visualize that DPM ensures a stricter and safer approach and 100 DPM is more difficult to achieve than 100 PPM.

I agree with Suri. Pareto analysis should be used to capture and target top few RPN's. A caution point here is - If a threshold is fixed, there is tendency of some people manipulating data to keep values within threshold so that no action is needed.

SIPOC - a) Process capability - d) Most important factor - d) Main objective of analyze - a)

Companies like to differentiate GB and BB using one or more of the following criteria. Cross functionality Business impact Complexity. While many Six Sigma experts prefer the third criterion, the second one becomes the driver at many places. Hope this helps.

We may look at it this way. Do the following have a role to play in an HR professional's profile? Data driven decision making Problem solving Process improvement. If any of the above three seems relevant, there is a case for usage of Lean Six Sigma.

Hi Dhiraj, You may like to try the following. 1. Process mapping with VA/NVA analysis. 2. Analysis of pain points based on data on delays or errors. Hope this helps. VK

Hi Arik, Sigma level cannot be calculated in this situation. This is the disadvantage of discrete data. Unless you get defects/defectives, you cannot assess the z value. Best Regards, VK

Hi Priya, New entries and amendments may be taken separately if they have reasonably different process steps and pain points. To assess process capability, you may use a missed entry or amendment as a process defect. As this data is discrete, you may not find a need to calculate standard deviation.

Hi Mayank, This kind of analyses are used in mixture designs while doing DoE. You may like to try exploring that area. Hope this helps.

Hi Shiva, Such carefully selected samples by marketing teams may not represent the population. A good way to try would be to request random customer selected sample. Well, once you have already purchased data and want to check whether the initial sample represented the population, you may use one proportion or a 2 proportion test to check for correctness. Hope this helps.

Hi Nagaraj, Thanks for posting this question. I think this is so because Quality Assurance is the function that poses as customer's advocate. QMS and EMS implementation also consider and assure different customer groups about no adverse impact and commitment to VoC.

Hi Kamal, The most suitable program is the Benchmark Six Sigma Green Belt (Probably followed by the Black Belt). Both these programs have global recognition and validity. You may like to have a look at Mr. Manikandan's video (He is from Dr. Reddy's Design team) who found several benefits on completion of our programs. Do let me know if you have further queries - http://forum.benchmarksixsigma.com/page/media.html/_/manufacturing/manikandan-r89 Best Regards, VK

Dear Wani, There is one example of problem solving in credit card company which is shared here - http://www.benchmarksixsigma.com/lean-six-sigma-in-financial-services-finance-mutual-fund-credit-card Best Regards, VK

You may like to have a look at the overview of DMAIC with various examples - shown here - http://globalnews.benchmarksixsigma.com/category/green-belt-qualified/example-green-belt-projects-services/ If required - Do provide your mobile number. I shall call you to bring more clarity for your course of action.

Hi Salil, Can you please bring more clarity on "questionnaire which can show tangible results to achieve improvement". As mentioned earlier, the first step is identification of a problem area which we shall call the project Y. Please let me know - Has this been identified? Also, please look at the checklist shown at the end of Green Belt material. This shows the entire road map of 15 steps in DMAIC. Do keep posting your questions. I am right here. ​Best of luck.

I understand this situation. It can be quiet intimidating. No, you do not need to be an expert in the products but you need inputs to identify a product or process where there are more issues. You may try the following. Look for customer data - complaints, feedbacks, comparisons, unresolved queries, comments from lost customers, repeat customers, or any such information available. Look for workable metric among these. If reasons are known, try doing a pareto analysis for reasons. Speak to experts in marketing, sales, product development, version release, installation folks for 2 or 3 of the most selling or highest value products. Are there repeat issues faced by customers? Do they want some issues addressed to ensure first time right? If you are able to find main areas that can be supported by key people, you shall be on the right track. Hope this helps.