iSixSigma

Criticisms of SS

Six Sigma – iSixSigma Forums Old Forums General Criticisms of SS

Viewing 28 posts - 1 through 28 (of 28 total)
  • Author
    Posts
  • #43967

    Marlon Brando
    Participant

    In  a  new  article  I  read the  following points  of  view:
    *The  term  SS has  its  roots in  quality  tools that  can easily be  misapplied by  a  naive  user and  to  the  controversial 1.5 sigma  shift.
    *SS has  been  broadly  critized   for  clinging to  the  concepts of  “attribute” and   “variable” data,rather  than  the  much  more  widely  accepted “nominal”,”ordinal”,”interval,and  “ratio’  model.
    *** SS can  be  considered just a  collection  of  tools  and  mehods,rather  than   a  methodology,itself..
    **** SS has  been  described   as  a  collection of  superficial  changes  that  ignore  many  of  the  major  factors affecting  quality  and  productivity.
    I  appreciate  any  comments  or  contribution,as  I’m  preparing  a  research  paper , focusing  on  the SS myths & advantages. 

    0
    #140140

    Hans
    Participant

    *The  term  SS has  its  roots in  quality  tools that  can easily be  misapplied by  a  naive  user and  to  the  controversial 1.5 sigma  shift.
    This is such a broad statement that it can hold for any methodology be it six sigma, lean, theory of constraints etc. … anything is possible.
    *SS has  been  broadly  critized   for  clinging to  the  concepts of  “attribute” and   “variable” data,rather  than  the  much  more  widely  accepted “nominal”,”ordinal”,”interval,and  “ratio’  model.
    This statement must come from someone who has obviously not the slightest clue about the theory of measurement as it evolved from physics to psychology. The four-level theory of measurement came about as a means to justify measurement in psychology by S.S. Stevens in 1947. The whole four-level model is an oxymoron as there is nothing “to measure” at a “nominal” scale. There are good justifications for using the four-level theory in psychometrics and survey research applications. For the hardcore engineers (which I assume most of the audience of Six Sigma is) there is no need to use a measurement theory that was develeped in psychology.
    *** SS can  be  considered just a  collection  of  tools  and  mehods,rather  than   a  methodology,itself..
    Even if … what matters is “does it work”?
    **** SS has  been  described   as  a  collection of  superficial  changes  that  ignore  many  of  the  major  factors affecting  quality  and  productivity.
    The Y = f(x) model should ensure that the factors “affecting quality” are not ignored. If it doesn’t then the author of the criticism should bring forward an alternative model.
    In essence, the “criticisims” completey miss the critical factors that inhibit a successful implementation of Six Sigma. They are superficial at best … sorry to be so blunt!

    0
    #140141

    Yes, I have a problem with this model.
    y = f(x) describes a physical law – it does not describe a real process.
    By way of contrast, Taguchi’s model:
    y (m, s) = f(x,s) x does represent a real process and clearly show how to apply variance reduction and scaling to minimise variance to target.
    However, there is much in SS to be recommneded – mostly those parts not contributed by Dr. Harry – specifically those parts Japanese company’s refer to as process excellence.
    Andy

    0
    #140144

    Hans
    Participant

    Andy,
    If Y = f(x) was to describe a physical law, then we would not need statistics. Historically, the model was developed because laws of physics do not explian phenomena in biology which are mere associated regularities that can be expressed in numbers. Y = f(x) implies variation (unlike it’s algebraic counterpart which is used to describe physical laws, and which are assumed to be constant at least in the classical Newtonian sense). All Taguchi does is expand on the equation by specifying process conditions. Thus, logically Taguchi’s formula can be seen as a special case of the more general formula used in current day statistics.

    0
    #140146

    Marlon Brando
    Participant

    I’m  really  glad  to  read  this debate.Adding  valuable  contributions  and  comments to  this  important  subject.

    0
    #140153

    Well, we’ll just have to disagree …

    0
    #140155

    Hans
    Participant

    It’s a pleasure to have a disagreement with a true gentleman. Thanks for the opportunity.

    0
    #140157

    Marlon Brando
    Participant

    The  use  of  “Black  Belts” as  itinerant change  agents is  controversial as  well as  it  has   created  a  cottage  industry of  training and  certification which  arguably relieves management  of  accountability for  change;pre-Six Sigma implementations,exemplified by  the  TPS and Japan’s industrial  ascension,simply used the  technical  talent at  hand–Design,Manufacturing and  Quality Engineers,Toolmakers,Maintenance and  Production workers–to  optimize  the  process.Meanwhile  ,for  companies not  solely devoted  to  manafacturing,the  spillover  effects  of  SS have  been troubling,especially as  executives trained  in  SS methodology for  change and  growth,clash withh  the  creative minds behind less  industrial   business  functions. 

    0
    #140171

    Hemanth
    Participant

    Well Yes
    I do agree that six sigma or by that means any process reduces creatvitiy and creates repeatable actions. This leads to demotivation, this is especially true in software industry where people take pride in creation (by their own way). But yes, not having enough processes does not make business sense.
    Six Sigma is a framework that is used for process improvement and this framework is so descriptive that it takes out the creativity in doing process improvement work. But is this bad?? Not sure…

    0
    #140173

    Samir Palsule
    Member

    I see Six-Sigma is a problem solving methodology, which works on D-MAGICS…
     
    D – Define the problem.
    M – Measure and generate the data points.
    A – Analyze the data points.
    G – Grasp the situation or appreciate the situation in which you are.
    I – Improvise the situation through root cause analysis. Initiate the corrective / preventive action loops.
    C – Control the resources provided for performing the processes and check the effectiveness of the action.
    S – Sustain the improvement and Standardize the same.
     
    By following D-MAGICS, it may apparently be seen that we are killing the creativity of an individual through standardized manner. But let us understand that business requires R&R (Repeatability & Reproducibility) in all the processes that they are conducting. Hence every business demands standardization.
     
    There are certain negative points of Six-Sigma. But looking at the positive points, I hope it is good try this methodology to reap the positive results. We can have the creativity in data generation, collection, analysis & intiating the actions. The most important ASPECT Of Six-Sigma is “it works on actual facts / data points, which eliminates subjectivity”.
     
    – Samir.

    0
    #140174

    Hemanth
    Participant

    Hi Samir
    D-MAGICS is interesting… but not sure how appreciating the situation would help… can you explain this in more details.
    Thanks
    Hemanth

    0
    #140176

    Samir Palsule
    Member

    Hemanth;
     
    What I mean from “APPRECIATING THE SITUATION” is…
     
    You understand very well that whenever you know the exact status of the problem, you can act & respond in better manner.
     
    The step GRASP helps us to understand the exact status of the activity and brings the team working on the issue on common platform. If data shows some negativity, then our immediate reaction is, I was doing so many things to control the situation but the data shows something different. Then we start blaming data gathering system or data is totally wrong. HENCE BEFORE WE MOVE TO ANAYSIS PHASE, IT IS ESSENTIAL TO UNDERSTAND EXACT SITUATION AND UNDERSTAND THAT WHATEVER DATA SHOWS IS CORRECT AND WHATEVER CONTROL POINTS I HAD WERE NOT EFFECTIVE TO TAKE CARE. In absence of this step, I am afraid, data analysis may go wrong.
     
    Knowing the exact situation in which you are, we can chart the action plan more effectively and measurables for the process can be defined in better manner. YOU CAN IMPROVE ONLY THOSE THINGS WHICH YOU CAN  MEASURE.
     
    Hope I have clarified ydoubt that you had.
     
    – Samir.

    0
    #140178

    Hemanth
    Participant

    Hi Samir
    I agree with you it is very essential to ensure the current process baseline and agreement on data collection mechanism before moving to analyse phase. Though the Grasp and Appreciate… step comes after the Analyse phase in the roadmap you shared.. and hence the question in my previous post…
    Regards
    Hemanth

    0
    #140179

    Samir Palsule
    Member

    Hemanth;
     
    It was a typographic mistake in my previous post. In the road map of D-MAGICS, Grasp comes after Analyze phase. Grasp provides us the confidence and validates assumptions behind measurement system before moving to improvize phase.
     
    At the end of Grasp stage, the team has to agree on the analysis part and every member has to confirm the correctness of analysis (also completing the major milestone).
     
    – Samir.

    0
    #140193

    Heebeegeebee BB
    Participant

    So, how would you quantify, “Grasping the situation”??
    I say, use the KISS principle and stick with DMAIC/MAIC
    (or DMADV for DFSS)

    0
    #140199

    Samir Palsule
    Member

    Grasping thje situation is an improtant mile stone in the journey and it can be concluded with sign off from the team that they agree with the ANALYSIS phase.
    You can work on D-MAIC, as far as all the team members confirm the correctness of ANALYSIS phase. D-MAGICS only provides ‘ highlights G & S (additional) as these are inportant milestione.
    – Samir.

    0
    #140201

    I’d suggest grasping the situation before defining the problem. I would go so far as to say grasping is part of this first step – otherwise one is grasping at straws.

    0
    #140202

    Hans,
    An associate of mine asked me to explain why I disagree with you; so I decided to post it.
    Your first counter argument is not relevant to my quote  by ‘Jeffries’ – a famous theoretical physicist. Why you would think a physical law y= f(x) is restricted to not using statistics is beyond me.
    You are correct though when you identify the very point of my comment – the statement of a physical law does not include any parameterisation of the process. At least we can agree on this point.
    I’m also concerned that so few people in Six Sigma seem to understant the importance of ‘scaling.’ Perhaps this is one of the reasons why so many deployments fail. (Please not I never claim any credit for pointing these issues out – I leave that to the generocity of the reader.)
    Of course I could test someone’s undestanding of these finer points – but it is not my goal to challenge anyone’s knowledge; only to try to point our several fallacies in Six Sigma thinking which for the past three years I’d hoped would inspire quality professionals to correct.
    Unfortunately, this does not seem to be the case – Six SIgma apparently is not for change itself, and therefore as a result of Dr. Harry’s muddled thinking many people are confused, many people struggle with the irrationality of some concepts, there is still a large distance between Taguchi Methods and ‘classical experimental design and its defence of psudeo interactions,’ and now everyone is talking about Lean and have by and large forgotten about process excellence (use of a natural process capability.)
    Andy

    0
    #140215

    Marlon Brando
    Participant

    Wish to  read  the  opinions of Darth,Stan & Mike  Carnell

    0
    #140225

    Mikel
    Member

    I assume you want a response to your original post.
    All of the criticisms are true for the majority of Six Sigma, Lean, Lean Six Sigma,TOC, … (insert latest, most hip label) implementations.
    Just recognize these are all labels and labels make no difference in anything. Just like the guy who threw in DMAGIC – what a joke!
    Most of the consulting companies have cash flow issues if they don’t keep the pipeline full – they are too big to have gaps in revenue. What this means is they take all customers and sell them training. They could care less about the long term sustainability. They keep on coming up with new stuff – new labels, new tools – when what the majority of their customers really need is basics.
    The basics are – have a rational strategy based on who you are and the market you are in, have measures and rewards aligned with that strategy; value stream map your key processes and identify gaps in execution (yes, Goldratt is right about this); put discipline in place to address those gaps that can be addressed by consistency; put projects in place for the other gaps using a pareto of pain, benefit, and resource availability; support the change agents and teams with the most visible thing possible – time.Oh yea, a good meaurement system aligned with strategic needs would also be good.
    The need for growth and efficiency is as true today as it was 50 years ago and as it will be 50 years from now. Six Sigma, Lean, Triz, TQM, TOC,… are all passing labels. What every company needs is not new tools but people who know how to execute strategy. When the student is ready, the teacher and the tools will come. Look for help from those who know how to execte, not from those with the latest twist or label.
    Did I answer your question?

    0
    #140228

    Hear, hear ..
    Andy

    0
    #140232

    Savage
    Participant

    Bravo.  Excellent post Stan.

    0
    #140266

    Heebeegeebee BB
    Participant

    Stan,
    Clear, succinct and well thought out.
    Good reply.
     

    0
    #140271

    Hans
    Participant

    Andy,
    Thanks for taking the time to clarify our argument and to point out the agreement that we have.
    I think we even agree more than you are pointing out. My argument was based on the fact that statistical regularities are not identical with physical “laws”. (Simply put, the fact that the ratio of males to females in regards to births has been consistently shown to be 51% to 49%, does not mean that this is a physcial law. In the 18th century this regularity was interpreted as “god’s will”, and in the 19th century as a “social law”. We now take it as a statitical regularity that is still looking for an explanation)
    As you correctly state, physics has moved into the direction of statistical reasoning, and statistical reasoning has moved away from viewing its “law-like regularities” as physical laws. I think that we can agree on that.
    With that in mind, I think we also agree that in Six Sigma we are using models to validate our findings. Your point is that Six Sigma needs to be more careful in justifying its model assumptions. Thus, you point out that Taguchi’s assumptions have an advantage over classical DOE. I agree with you.
    My other point is that Six Sigma itself makes assumptions about its applicability to business performance improvement. Just as you point out that when statistics is used to enhance performance of a process the assumptions need to be clearly understood, so I point out that the application of Six Sigma itself rests on assumptions that are not often very clear. Where we would probably disagree is the impact of one error (not using the right statistical methods) versus the other error (using Six Sigma and deploying Six Sigma in the first place). Again, I think we agree that there are some serious issues with regards to the testing of the assumptions of Six Sigma in applying its tools both at a macro-level (my point of view) and at a micro-level (your point of view).
    So, in essence we both agree that there is “muddled thinking” in Six Sigma and that we shouldn’t leave it to the “Dr” Harry’s to “muddle” our thinking in the first place. I truly appreciate your insights.

    0
    #140280

    Hans,
    Thank you for your eloquent reply.
    Yes, on the whole I agree .. including your perspective of our view points – marco vs micro.
    But to clarify, my microscopic concern is for macroscopic principles, such as the important principle of using scaling to reduce error. My point was this principle does not appear to be encapsulated by y = f(x) and of course it is one of Taguchi’s strategies – the other is which set point to choose to transmit the least error.
    Thus, there has been little appreciation for the value of Taguchi’s signal-to-noise ratio –  used in many of Moto’s waferfabs and at a time when Dr. Harry should have discovered the real reason why Austin waferfabs were so successful.
    Regards,
    Andy
     

    0
    #140345

    Marlon Brando
    Participant

    Excellent Answer

    0
    #140359

    Mike Carnell
    Participant

    Marlon Brando,
    My apologies for the late response. I have been preoccupied with moving my home and traveling back to South Africa.
    I would be interested in reading the article you are refering to and seeing who wrote it. Because someone gets something published doesn’t necessarily mean it has any credibility i.e. the USA Today article in the mid 90’s that had a technician at Allied saying he didn’t think much of Six Sigma and then a few quotes from Bossidy. Let me see if I can figure out who has the more strategic view of the effect of Six Sigma on Allied Signal the technician or the CEO. 
    There are no lifeguards at the shallow end of the gene pool and occasionally the stars line up badly and a couple of those guys get together (a writer and a publisher) and get something into print. Probably a little judgemental without reading it but that is my initial feeling. A quote from Harry Anderson (Night Court) “It is important to keep an open mind, but not so open that your brains fall out.” so maybe I need to hold off until you give me a reference.
    Six Sigma has its roots in business improvement not quality tools. It is interesting how the quality community has laid claim to statistics as “theirs.” There are a few tools that were probably born in the quality community but overall the substance comes from statistics which do not belong to quality, the exception may be MSA. It isn’t much different than the comments about Six Sigma being a dirrivative of TQM. Once someone put the TQM tools together TQM owned the tools? Where is the logic in that?
    The 1.5 sigma shift is acdemic. We mention it in training because it is important that a Belt be aware of it as a part of their heritage. Does the success or failure of a project depend on it? No. It has no effect beyond a belt being aware that there is a good chance that a short term capability will probably be over stated. If they want to know by how much – get the data after all this is Six Sigma and we are supposed to use data.
    As far as clinging to the concept of attribute and variable data. If you consider the number of belts in the world the reference to “widely accepted” is probably incorrect. Our belts are aware of nominal, ordinal, interval and ratio data. The question becomes would spending time on the different types of data change the results significantly? No.
    When we began the Allied deployment in 1995 we could get all the trining material intoone 3 inch binder (we gave out about 8 different books). Now we have five 3 inch binders full and the projects have not changed significantly. It isn’t about tools and a better understanding of tools. It is about how do you take a problem apart and understand what is and is not important. Blooms Taxonomy may give you some insight into this. Any belt that isn’t at Application, minimum, is a disfunctional Belt.
    Depending on who schooled this person on Six Sigma from they could easily have been taught that it is a collection of tools. There are a lot of people who have been taught the technique of playing the guitar but there are not many people who play the guitar like Clapton. It isn’t about technique – lots of people know the technique for playing the guitar – it is about understanding at an intuitive level what you are trying to accomplish, what questions get you there and what tools answer the questions. We spend more time understanding what is going on in the problem solving situation than focusing on tools. The tools need to be understood on some level. They only answer questions. With the advent of programs such as Minitab we can spend more time on understanding the relationships between factors and determining which ones are relevant.
    The last statement is a glittering generality and I would be willing to place my mouth where the money is to see the data that substantiates that nonsense.
    Just my opinion.
    Good luck.

    0
    #140384

    Marlon Brando
    Participant

    Thank You  Mike  for  your  comprehensive  reply.It  was   5-pages  article in  the  Internet,signed  by nobody,but some 17  references  were recorded at  the  end,including  Motorola  University……Joseph A.De Foe & William W Barnard (Juran  Institute’s SS Brakthrough and Beyond,and  finally  as  the  last  reference:Mikel Harry & Richard Schroeder,best  regards

    0
Viewing 28 posts - 1 through 28 (of 28 total)

The forum ‘General’ is closed to new topics and replies.