Is "Stat-Free" Six Sigma possible?
by Matthew Moore on 27th March 2007I published an article yesterday about a new book which claims to provide "Stat-Free Six SIgma". The article, including the corresponding press release from the publishers, is here: http://www.onesixsigma.com/article/new-book-claims-stat-free-six-sigma
Given that I doubt anyone has read the book yet, is modifying Six Sigma to reduce the use of statistics a dangerous practice that goes against the ethos of the methodology? Is it even possible to do so and still call it 'Six Sigma'? Or is this simply an evolution in a world with an ever-increasing need for quicker results and shorter project timescales?
Let us know your thoughts.
Comments
Service Providers
hi ı believe there is a
hi
ı believe there is a major commercial dispute that sig sixma is getting mixed with lean manufacturing tools. although when we look at the roots of six sigma and see that it reasons to minimize fluctuation or noise in a system and aims to make it more smoother and expectable, there are many consulting activities that are carried very similar to lean tools. my belief is six sigma is a powerful tool that can be used as a remedy in the later steps of any disease. considering that lean can help us to collect apples on the lower brenches of a tree, lean comes later to facilitate gains that are hanging on the higher end of the tree. before isoloting all the parameters and minimize the noise any model to test the relevance will be more diffcult and time consuming to identify root causes. lean can sweep all easier to moderate reasons that disturp the efficiency of the overall system afterwards six sigma modeling can help a deeper analysis to map the system and test all the action plans.
It may be closer to the original Shewhart/Deming way ...
Shewhart has invented Statistical Process precisely because he was against the use of traditional Statistical Tests.
I was just about to write an article aboutt that:
Why Shewhart denounced the misabuse of Student's t-distribution in Science and Manufacturing
I don't have time to finish it yet (wait next week end) but for now on you can read an introduction to this:
"Shewhart/Deming Statistical Process Control vs Six Sigma"
I'm happy that you made me discover their book and I should write an article about them on my Statistical Process Control Blog.
Stat-free Six Sigma
I would say that DMAIC and Six Sigma are separable, i.e. the DMAIC process can be applied to a non-six sigma project. It is in my opinion that six sigma automatically implicates statistical measurement - how can you aim for a 6 sigma level process without statistical measurement.
So I would say stat-free DMAIC is possible but stat-free Six Sigma is not.
Simon Brice
Stat Free?
I facilitate six sigma and practice it within international projects within the company i work for.
In my view its the statistics that separate the facts from the fiction.
To practice stat free six sigma i think firstly you have to had experience theoretically and practically in understanding and applying statistical six sigma.
Once you have this experience and knowledge then you can minimise the statistics as appropriate. I would not be comfortable in removing all statistical testing.
Its the stats that sets six sigma apart from other methodologies.
Stat Free Six Sigma
You have raised good questions. Six Sigma has become overly a statistical tool without realizing how it was started, what made it work, and what was its intent. Six Sigma has four components...Intent, Methodology, Tools, and Measurements. You can see that using statistical tools without internalizing intent of aggressive goal requiring innovation would not produce desired solutions. The purpose is to practice the intent and methodology with simple tools to maximize benefits. As to being called Six Sigma, the original definition implied Six Sigma as a methodology to achieve virtual perfection, rather than applying some statistical tool.
Praveen Gupta
Co-author, Stat Free Six Sigma
Stat-free
As usual, Amazon's 'Open Inside,' offers nothing to explain his approach. Why not email Praveen and ask him?
Oh my god!!!!!!!!!!
Teresa- how the hell are you? Wow, what a nice surprise! Why are you readng stuff on this website? Give me details about yourself. jet
Stat Free Six Sigma
I wholly agree with your comments. On too many occasions SS experts are becoming increasingly 'geeky' with statistical analysis; often losing the why, what, when purpose of collecting the data in the first place. SS without statistics is like football without a ball. Statistical data collection and analysis is vital to correct SS implementation; after all, giving the organsiation enduring successes is our target. I am intrigued about the book though and will obtain a copy to read.
Certainly not 6S
I read your comment and found myself wanting to respond as well, but certainly not about 6S.
Congrats on career move!
Teresa
U of K
Rugger Hugger
Stat Free 6S
I will be reading this with interest. In the meantime...maybe it is because I had a few moments on my hand near the end of a day and because I have been around the block a few times in this field; that I find myself wanting to respond. There are so many definitions of 6S that settling on one seems impossible. All I know is that some years ago, I decided that it truly does not matter. We changed the "I" in DMAIC to "E" to push organizations to "Engineer" the most appropriate solution(s). That also taught people how to use the appropriate tools for the problem or opportunity at hand...we call this situational surveying. A number of our most significant problems have been solved with the 7 basic quality tools. And, some of my most recognized projects did use more stats...data partioning, multiple regression, etc...and all in front office financial services projects. BTW- 2 of them won Rath's and Strong's Best Project Competition. In the end the customer does not care! Stats or not...did you fix a problem worth fixing and did the solution get at the root cause or did you apply bandaids. We use a pretty simple thought when asked about this optimization stuff like 6S...what is it? "Good Data, combined with Common Sense applied with a Practical View of the world around us...yields better decisions." We can do something about the first "X"...but there is still no program to teach the second or third "X". So in the end, data is but one of the key variables...stats or no stats won't matter without the other two.
Jon Theuerkauf
MD- Credit Suisse
* Global Head- Operational Excellence
* Global Co-Head- Cost Management Initiative
Stat Free Six Sigma
The main aspeect of stat-free six sigma is to practice the intent of Six sigma. Without knowing the intent it really does not matter how much statistics any expert uses. With the intent understood, the methodology is dominated much by the Define phase which is practically non-statistical. Well defined problem is half solved, and poorly defined problem is rarely solved. Then, in the Measure, Analyze and Improve phases,,,DPU and DPMO are non-statistical measures. In the Analyze phase root cause analysis is non-statistical, and in the improve phase comparative experiment, some of techniques for evaluating means and variation have been converted into information that can be easily remembered and applied.
What is more important is statistical thinking not the statistics and complex statistical tools.
I hope it clarifies. Otherwise, it may be just worth buying the book to learn more about it:)
Praveen
Stat Free Six Sigma
I haven't read the book yet, but would certainly like to understand the approach. To be honest however, I don't see how you can articulate 6 Standard Deviations, (3 on either side of the mean), which is where the term came from, without using statistics. I agree that many of my collegues are getting bogged down in analysis and forgetting there is a purpose for the analysis...but the fact remains that analysis and statistical hypothesis testing is part of the approach.
Jeff Book
Director Six Sigma
Procise GmbH