iSixSigma

Long term and short term variation

Six Sigma – iSixSigma Forums Old Forums General Long term and short term variation

Viewing 13 posts - 1 through 13 (of 13 total)
  • Author
    Posts
  • #50810

    chhabra
    Participant

    The goal of a six sigma project is to stabilize and improve capability of a process. How then should long term variation be different from short term variation after successful completion of a project? Why should we specify whether the new sigma level is short term or long term. If a process is already stabilized, why are we expecting the process to drift away and therefore become unstable ?

    0
    #175145

    ISS
    Participant

    Refer long term and short term which indicates the time factor into consideration, and as well bring into consideration special causes which shall fluctuate the process in the long term, inspite of tighter control limits and monitoring. This hs been found to be applicable in all the processes and hence been generalized as a concept.

    0
    #175146

    chhabra
    Participant

    But doesn’t that mean that we have failed to ensure that the process remains stable ? Also, how long is “long term” ?

    0
    #175148

    ISS
    Participant

    It depends on the process to define a long term or a short term time parameter.
    A manufacturing highly automated process may have a probability of a special cause making it fluctuate anywhere between 3-6 months, whereas a process may have a special cause occuring inΒ a much smaller time duration in a transactional environment where process has large human dependency.
    Its not completely right to say that we have not been able to control the process, its just that we were not able to built in all special causes in our solutions to take care off
    Β 

    0
    #175161

    Anonymous
    Participant

    These terms were invented by Mikel Harry.Β  They are completely meaningless.Β  In his original “proof”, “long term” was one day !

    0
    #175165

    GB
    Participant

    Almost…
    Pegging an actual shift was partially his idea.
    However, in over 15 years of analyzing processes, I’ve found that most of the processes I’ve had experience with, did in fact exhibit some “Process Slump” over time.Β Β  Whether it was due to tool wear, Operator complacency, or turnover, the causes varied.Β Β  But to peg all “shifts” at 1.5…now that is a stretch.

    0
    #175167

    John B
    Participant

    A,
    It is a bit confusing and debating the issue itself is debateable to some…but you are really looking at taking a measure of dispersion in a given moment or “snapshot” of time when first calculating the process capability.Β Β  Here you are taking consecutive measures where the variation is assumed to be common cause only (ie within subgroup), since you addressed this earlier in your “is it stable and app. normal” checks during MEASURE.Β  But this variation doesnt represent the process over time, simply the process at that point in time…so consecutive measures which then calculate the dispersion both within your samples and between them are combined and this becomes the “overall” measure of dispersion in your process….and how long is “long term”?Β  Depends on the nature of the process and what your MBB tells you…often there are organizational rules of thumb, like “3 months, quarters, etc)….That’s my understanding….good luck.
    Β 
    Β 
    Β 
    Β 

    0
    #175168

    mand
    Member

    Not surprising when most processes are so poorly managed … six sigma being partly responsible.

    0
    #175187

    GB
    Participant

    Sam,
    Wrong!Β Β  Six Sigma allowed the team to see the slump…It didn’t cause the slump.Β Β Β  When they applied the methodology and actually bothered measuring their process, the slump became visible.
    Nice try though…

    0
    #175271

    Bill Fowlkes
    Participant

    Very well put ISS.Β  As for some of the cynical remarks in this thread, you guys should give Mikel Harry a break, after all without him, you don’t have your training, your belt and probably don’t have a job.
    Bill Fowlkes, SSBB class of September 25, 1992, Schoenberg, IllinoisΒ Β (where it all started!)

    0
    #175273

    Eugene Lanning
    Participant

    Perhaps my perspective is wrong, but this is my perspective:
    Β 
    When I have a project to work on the issue is studied, and the available information is roughly Y=f(x1,x2, x3, x4, x5,x6)Β  — assuming that one has identified 6 variables at that time.
    Β 
    Some variables that could be manifested will be unknown to the team, variables x7, x8, Β….Β  The reason that they are unknown is that someone thought they were too trivial to be considered, that the probability of their occurrence was low, that they just did not remember them, or in some cases the team had to only consider a limited number of variables so that the problem was manageable.Β Β  The key here, in my opinion, is that there are more variables than were studied in the development of a solution.
    Β 
    As time goes on, occasionally variable x7, or x8, or an unrecognized x9 will become important and affect the output Y.Β  What is the effect of that?Β  The variation in the process is now larger by some amount.Β  I would not label the process unstable, because the variable x7 (or whatever) will likely return to its normal value and your Y will also return to its normal value because again the process is adequately described by the first 6 variables.Β  Using an SPC chart and tracking down any Special Causes will help identify the x7, x8, etc variables so they are better controlled in the future.
    Β 
    Just my perspective.
    Β 

    0
    #175276

    Jered Horn
    Participant

    This is a good question.Β  You throw in a little controversy when you insert the word “drift”, though.Β  You may not be, but it almost implies that you are simply using the 1.5 sigma shift to calculate your long term capability…and only measuring short term capability.
    Your company/organization, or even industry should have some guidelines for you to differentiate between a short term and a long term study.Β  Apply those guidelines along with rational subgrouping to a long term study at the end of a project.Β  Take your short term (within subgroups)Β and long term (between subgroups) results and there you have it.
    You’re right, the goal of the Control phase is to reduce that shift or “drift”.

    0
    #175293

    Rakesh Bishnoi
    Participant

    Hi Hornjm,
    Β Β Β Β Β Β Β Β Β Β Β Β Β Β Β Β  Sir m Rakesh Bishnoi working with a collection agency from last 2 yrs,just start working on a project based on Portfolio performance,n how to raise the benchmark of the portfolio,if possible i need some guidance from you,n even some better suggestion for improve the quality n accuracy of the project.
    Regards,
    Rakesh Bishnoi

    0
Viewing 13 posts - 1 through 13 (of 13 total)

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