iSixSigma

setting goals outside control limits

Six Sigma – iSixSigma Forums Old Forums General setting goals outside control limits

Viewing 28 posts - 1 through 28 (of 28 total)
  • Author
    Posts
  • #48385

    Pipkin
    Participant

    I am relatively new to the six sigma process, but the government agency I work with has produced a number of control charts with upper and lower control limits.  However the goals that the agency HQ have set for the field to attain on a number of the control charts fall well outside the control limits.  This does not make sense to me.  Can or should you plan objectives or goals that fall outside your control limits?  Any guidance or insight would be appreciated.

    0
    #162886

    BTDT
    Participant

    Jack:When you say goals, do you mean Upper and Lower specification limits?Cheers, Alastair

    0
    #162888

    Dr. Scott
    Participant

    Jack,
    First my answer to your questions is yes and yes.
    Below are the recent tests for “speed” (it does not matter the speed of what). My target is a speed of 2000, my minimal acceptance is 1900, and my maximum acceptance is 2200 (due to operation failure or danger to the operator). What do you see below? What does the control chart say, and how does that compare to my “goal” or requirements?
    Speed1600.831332.311142.871116.071796.891561.211609.481546.121156.441383.201138.771323.441363.551322.621356.151216.601478.051197.461364.861449.021438.091092.861457.401603.811620.711745.981365.321595.161291.791691.441054.631360.741361.571695.561481.041039.281418.251636.911433.141522.721367.731602.421203.541578.54946.831639.511094.731531.261487.30921.651348.191129.811608.561452.02890.671610.421248.711684.42981.431566.801330.291258.951114.901047.591479.101447.851247.781349.731521.891527.941110.861263.541591.411521.61901.451365.061666.921488.241268.921476.051053.871638.571493.841582.331319.361352.651193.371184.351230.461455.331769.79913.451253.311372.681322.391590.611366.821574.381328.711422.211453.211233.611548.821833.591859.841292.771495.711379.051550.921640.051319.321225.261619.961779.811390.861182.241575.351168.141276.881500.571536.471490.731446.921573.251148.021729.801034.641449.221358.601193.071387.431651.741230.431504.671518.601540.241638.491436.351161.521512.341365.871382.511296.131026.441366.601353.611453.521712.061299.391087.581457.951516.271234.601278.651496.35976.701385.081307.011317.531365.221375.861527.571347.781436.591559.161563.991421.581743.631650.721323.911433.281325.49957.001276.091436.161417.721563.781398.661593.331534.661400.741053.14949.481215.991141.451221.961397.481559.211311.551425.811606.381408.601731.841213.481337.051500.861179.441196.801450.761375.54
     Get back to me with your answers?
     
    Dr. Scott

    0
    #162890

    Brandon
    Participant

    Jack, could it be these goals have to do with collection of tax dollars?
    If so, your superior’s objectives don’t surprise me.

    0
    #162895

    Mikel
    Member

    It’s done all the time, it’s called a breakthrough objective.

    0
    #162896

    Pipkin
    Participant

    Dr. Scott,
    I am new at this, but here goes.  All the values fall outside the upper and lower  tolerances that you provided (2200 and 1900) and none of the values meet the goal (2000).  The graph showed that the speed numbers were mostly within + or – 3 standard deviation and therefore within statistical control limits.  You could conclude that the process is within control limits, but does not mee the tolerances or goal specified.  To achieve the desired resutls the entire process as currently configured and employed could never produce the desired results and would have to be totally reworked to meet the desired goal and tolerances.
    Again hope my ignorance doesn’t show too badly.  And thanx for your response. I look forward to your critique of my answer.

    0
    #162900

    Pipkin
    Participant

    Alastair,
    No, I think what they are trying to do is set goals that are untainable.  For example if you track unemployment and the the trends shows mean unemployment over the past decade at say 5.0 % and you set control limits of + or – 3 SD so the control limits are upper say 8.0% and lower is 3% they want the goal to be 1%.  Using the current processess I don’t see how you can get there statistically.

    0
    #162903

    BTDT
    Participant

    Jack:Your control chart can be used to show management that the present process is not capable of achieving the goal – you are not merely having a few bad months. The present process is in control, but can not achieve the goal without having a project directed at changing the process.Cheers, Alastair

    0
    #162914

    Mike Carnell
    Participant

    Jack,
    I am with Stan on this. Juran documented this in his book “Managerial Breakthrough” in 1964. Take a look there is a picture of exactly what you are speaking about depicted graphically. You are missing the point of differentiating between Breakthrough and Control.
    I am not sure why you are having such an issue with this. The control chart is a reflection of where the process is operating currently. Why would you believe it cannot operate some place else?
    Generally the mean can be shifted with knob variables. Go twist one and see what hapens to your data. The process will generally shift someplace else and settle in at some new spot. That is why you try to eliminate people who tweek machines constantly – that are constantly shifting the process.
    Good luck.

    0
    #162918

    Dr. Scott
    Participant

    Jack,
    You pretty much hit the nail on the head, so to speak, with your anaylsis. And you answered your own question. Of course the goal can be outside of the control limits of a process. So the process has to be changed (As Mike said, “Go twist one and see what hapens to your data”). Except if I were you I would twist more than one “knob” in a designed way (i.e., use DOE to figure out how to improve the process).
    Control limits are the “voice” of the process, it is telling you what it is doing (like a young child saying this is what I am doing). Spec limits are the order to the process (like the young child’s parent saying here is what I want you to do). So if the process is not behaving as it is told to, then you have the ability to figure out how to make it do what you want. Jumping other steps, use a DOE.
    Hope this helps.
    Dr. Scott

    0
    #162931

    Pipkin
    Participant

    Dr Scott,
    Thank you for explanation. You are an excellent teacher.

    0
    #162932

    Taylor
    Participant

    Jack
    “Most important, leaders can conceive and articulate goals that lift people out of their petty preoccupations and unite them in pursuit of objectives worthy of their best efforts.”
                                                  John Gardner
    This is why you have goals outside your control limits. Any company, organization, government, etc that doesn’t set goals outside its control limits will soon no longer be such.
    Chad Vader

    0
    #162935

    Dr. Scott
    Participant

    Jack,
    Thank you for saying so. That is nice to hear now and again. Let met know if I can offer help in the future.
    Good Luck,
    Dr. Scott

    0
    #163321

    Simple Simon
    Member

    Dr Scott,
    I love your child-parent analogy; this is probably the best, and simplest, explanation of VoC / VoP that I have ever seen.
    SSimon

    0
    #163325

    fake accrington alert
    Participant

    Yes  but  it  should  be  within the  specification  limits

    0
    #163326

    fake accrington alert
    Participant

    Dr.Scott
    All  I  can  say  that  you  really deserve  your  Phd Degree.
    best  regards

    0
    #163338

    JC
    Participant

    What’s that saying about the cart and the horse?  Seems to me that is what is happening here.  W/O specific actions taken to meet any goal your efforts will always be suspect.  I would encourage you to ask your HQ a series of questions that follow the DMAIC guideline.  In my business, success for us is when a measurement falls outside the upper control limit (sales).  Also, keep in mind that, depending on your reference material, 3-4 consecutive points above or below, the pre-DMAIC mean is considered a change in the “process”.  Success is not always defined by points outside the base control limits.  Good luck.

    0
    #163345

    Dr. Scott
    Participant

    fake accrington alert,
    Are you being sarcastic or sincere?
    Either way, Thanks
    Dr. Scott

    0
    #163348

    fake accrington alert
    Participant

    Sincere

    0
    #163353

    fake accrington alert
    Participant

    Just  explain (simplify)  the  role of  DOE in  this  context.I  find  it  difficult  to  understand.
    thanks  and  regards

    0
    #163360

    Dr. Scott
    Participant

    Fake,
    You have a process (let’s assume it is in or close to in control), and you have a goal which is well outside the control limits and the process is not capable.
    So, we have to change the results of the process. The only way to change the results of a process is to change the process itself (doing the same thing and expecting a different result is the definition of insanity). So we have to change the process.
    Well then, now we have to ask (those in the process) what might we change? I am sure you know we won’t be short of ideas on how to do that (if you actually ask those who run and operate the process). But which one or two of the 20 or so ideas do you choose and implement change on? How do you choose the changes that are best?
    My answer to your question, DOE!
    Regards,
     
    Dr. Scott
     

    0
    #163363

    Taylor
    Participant

    Dr. Scott,
    Although I totally agree with your explanation and use of DOE, it is entirely possible to achieve results that are already known without running trials and experiments. One must always examine the choices of change and implement immediately those of which are “Known” to make positive impact. It is only when the result of the change is Unknown that a DOE should be applied.
    Just the Engineer in me that wants to get it done
    Chad Vader

    0
    #163364

    fake accrington alert
    Participant

    Thank You
    You  have the  talent  of  a  teacher

    0
    #163370

    mand
    Member

    I think the question you are asking can be handled a bit more easily by looking at the concept of control limits vs. specification limits.
    Control Limits are statistically derived, ie they are based on historical process performance average as well as a stastical calculation to determine the height of them, this is sometimes 3 standard deviations, or it is Rbar over D2.
    Specification limits are not statistically derived and they are determined by a business either to ensure profitability, compliance, or just to be something to reach for.
    I would be less concerned about the control limits per se, and more concerned about how capable your process is in meeting the specification limits.
    S-

    0
    #163371

    Dr. Scott
    Participant

    Chad Vader,
    Ok, sounds good to me. Implement what is already “known” to work, right? But why wasn’t the “known” solution implemented BEFORE you or I get the call to come help.
    My experience is that everyone involved in a process has the “known” solution, but very, very, very few actually have the data (Y=f(X)) data to prove it as “known”.
    As a very well renowned process improvement expert said in front of me once, “show me you data or take a break from the room to wash you hands and face, then come back to the room”.
    When you have the data, then you know. If you don’t have the data then you don’t know, you just believe (which is a good thing in that we can learn from that belief using data in the future, i.e., a DOE). And finally, do you know how your “known” solution will affect all other parts of the process? You probably do not, even if you have data from another place (e.g., book, expert or other device). But a DOE will answer this for you.
    So the Engineer in you should be strongly in favor of DOE to know what “known” ideas work, which ones don’t do a thing, and which ones actually hurt, in concurrence with each other.
    Respectfully,
    Dr. Scott
     
     

    0
    #163372

    Dr. Scott
    Participant

    fake accrington alert,
    I will take that as a compliment. But what I teach comes from what I have experienced. That is, I don’t recommend it unless I have done it and it worked. That is where good teachers get their ability to teach.
    Best Regards,
    Dr. Scott

    0
    #163379

    fake accrington alert
    Participant

    Congratulation
     

    0
    #163817

    tcm
    Member

    Jack,
    Wif ref. to the post above:
    I analysed the data and found out that there is a out-of-control point… The process is not in control??
    Could it be i am using the wrong chart??? My eyesight got problem???
    Mayb my experience is not enuf??? -.- i got oni 3wks of reading charts experience, so please correct me if i am wrong…..
    (i will be very happy if i can get some comments…..)  ^.^
    tcm

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

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