iSixSigma

Question on Improving Poor Capability

Six Sigma – iSixSigma Forums Old Forums General Question on Improving Poor Capability

Viewing 24 posts - 1 through 24 (of 24 total)
  • Author
    Posts
  • #32882

    Bullet Boy
    Participant

    Hi Guys,
    I am currently involved in a project with an OEM who has requested capability data. Before I sent it, I analysed it and found I had a Cpk of 0.7. Therefore, I had no alternative but to “massage” the data and achieved a Cpk 1.43.
    Will the OEM be able to statistically prove that I have removed the high readings, or do you think I’ll get away with it?
    If this stops him from proving liability, then I will have made considerable savings for my company- if you get my drift.

    0
    #88342

    Gabriel
    Participant

    Most probably, they will never notice that not because it can’t be noted, but because it is not like they are looking for fake data.
    Now, are you sure that was the best for your company? If you are delivering parts at a Cpk of 0.7, then you are delivering a lot of non conforming parts. They probably will not discover that they are non conforming immediatly, as most OEMs are not making incoming inspection these days. But then, assuming that the specification is in line with the product requirements, they will have a problem in their own process, or worse, customer complaints, and will finally find the non conforming parts. By then, there will be a lot of parts in stock (yours or theirs), there may be a product recall, warranty charges, production line stopage, and the like. Who do you think that will pay all those costs?
    And then they will issue a non-confority complaint, where you will have to explain how on earth a process with a Cpk of 1.43 delivered that ammout of garbage!
    And all that, without even starting to talk about ethics and the like.

    0
    #88343

    Picklyk
    Participant

    yep.  i’ve seen this everywhere i’ve worked.  lying to OEM’s.  supplying false data.  never a good thing but nonetheless, i’m sure the direction you’ve recieved from management was to massage the data.  so do you do the right thing or keep your job???
    one instance you can get caught is if you simply cut off, or trumcate, the portion out of specification.  if you have a nice normal looking distribution that all of the sudden is trumcated, that would throw up a flag.  make sure your distribution looks good in the event they ask for raw data…unfortunately, i’m very good at this.
    a good OE would ask for parts himself and perform his own capability study…but he’s probably just passing along paperwork like most as well.

    0
    #88344

    JB
    Participant

    Is Right!! (JB)

    0
    #88346

    Gabriel
    Participant

    And by the way, you are also lying to the forum with your false subject “Question on improving poor capability”. You don’t mention “improve” in all your message. You are just asking a “Question on fudging poor capability”.

    0
    #88347

    Robert Butler
    Participant

    If you have been ordered to cook the data and your job depends on it then you can take the approach of following orders and doing as told. However, before playing the orders-is-orders game and getting worked up about the Cpk of the process as computed by some formula I would recommend that you sit down and really examine the data that you have. 
    First-check to see if your data is normal.  If it isn’t then you and your customer are wasting time and there are other issues that need to be addressed.  The quick fix here is to explain to all and sundry that the Cpk computation is the wrong one and sit down with everyone and discuss sampling and acceptance plans that focus on the behavior of the process and not on an inappropriately applied formula.  
      If it meets the criteria for normality then:
     Plot the data in time sequence. Now really look at it. What is the structure of the extreme points that are “ruining” you Cpk computation. Do they actually occur at random, here and there in the process, or is it the more likely case that they are clumped and thus represent the situation where your Cpk computation is “degraded” by the occurrence of one or two actual bad runs? 
     If it is the bad run situation can you explain why the bad runs occurred? Can you show that corrective action was taken? Can you show that this action worked?  If you don’t know why these bad runs occurred can you show what percentage of your runs might be affected? Can you compute the probability of their occurrence? 
     If you lay the results of the process out in this fashion (do this regardless of the structure of the bad data) and can explain and document the process using graphs and calculations I think you will find that your customer will be willing to work with you and perhaps the two of you can get on with the business of making quality products and an honest dollar. 
      While your customer may walk, it has been my experience that the worst that happens is that you are driven to the short term solution of lot selection and/or 100% inspection.  The fact that you are honest about the shortcomings of your process, can show what they are, and are actively looking for process improvement should carry a lot of weight.
      If you are going to take the cooked data approach I would be very careful.  If you over do it your customer may decide that you are really good and they will come back with a request for even tighter specs.  Since you are going to have to do lot selection in order to live up to the lie of your original fudged data you will have to redouble your efforts in this regard to meet any new requirements they may have…or you will have to ‘fess up to the fact that you were lying in the first place.  Either way – it’s more trouble than it is worth. 

    0
    #88348

    CL
    Participant

    BB is probably not concerned about misleading the forum if heshe has no qualms about massaging data for a customer

    0
    #88352

    billybob
    Participant

    Hello folks,
    Werid company that Bullet Boy works for.  They don’t know their process capability until the cusomer asks for it.
    Later,
    Billybob

    0
    #88355

    Gabriel
    Participant

    Bravo!

    0
    #88360

    Mike Carnell
    Participant

    Bullet Boy,
    With the things that have been in  the news papers these days this doesn’t surprise me. The thing that surprises me is that you are only concerned that you may get caught and that you have the audacity to ask this in a public forum. It does not appear to bother you in the least that you seem to completely lack any type of ethics or you are just a sociopath.
    Bottom line, if you sent them truncated short term data when in the long term it was 0.7 it will be pretty easy to detect. I really hope they do. At the very least you should be fired.

    0
    #88363

    Denzel
    Participant

    Bullet Boy,
    (Is that as in “Gonna take a ….?”)
    1) As you are working on a project, where liability is at stake, I presume there are already parts in the field, which are causing some customer dissatisfaction. Therefore if it is your parts/process at fault, you are only delaying the inevitable. As soon as return parts are measured, and found out of specification, you are going to struggle to explain that away.
    2) As your knowledge of Statistics / Six Sigma seems limited, are you sure you have good measured data to provide to the OEM? What are your MSA figures like? Do you have acceptable Gauge R&R results, especially for %Contribution to Tolerance – Did you even do a full Type 2 study? So many suppliers to us seem to believe that by measuring the same part several times without moving it, they have a good measurement system – In my experience, it may get you through a QS or TS audit, but will be of almost no use to accurately measure the process in the real world.
    In summary, I believe you are wasting your & their time by “massaging” the data, and if you are at risk of losing your job if you don’t, then maybe it is time to find new employment!
    As a colleague of mine says, “If you don’t play by the rules, it’s only a matter of time before someone threatens to smash your face in….”
    Good Luck, which ever way it goes!

    0
    #88364

    Rick Pastor
    Member

    Just because you have a small Cpk does not imply that you deliver nonconforming parts.  If you are doing 100% inspection prior to shipment, you should catch most of the nonconforming parts. 
    What a small Cpk means is that you are not making as much or any money on the parts that you sell.  The question comes down to this.  What is your Cp?  If it is pretty good, then you need to adjust the mean of the process.  If neither is pretty good, you have to worry about reducing the variation in the process, which typically is harder than shifting a mean. 
    Here is my suggestion:  Improve the process and everyone is happy.

    0
    #88370

    Eileen
    Participant

    Unbelievable!! This thread is quite a commentary about the state of OEM suppliers. No wonder the defect rate on domestic cars is so poor.
    If your product is a liabiity issue and you are being asked for capability data, chances are you are already toast. Fudge the data, manipulate the data, change the data, you are already in the crosshairs. When I have had the chance to deal with a supplier like you, the consequences are quite severe. A hoard of locusts (process evaluators) will swarm on your facility and you WILL be desourced.
    Trying to find a way to trick your valued customer, you don’t deserve to be in business. Why not spend your energy and time making a better product? What is your product?
    Eileen
    Quality Disciplines
     

    0
    #88372

    Rick Pastor
    Member

    I avoided comments about the integrity of a company that would fudge their data to make themselves look good when I read Dan’s response.  But Eileen your comment, “…you are already in the crosshairs” was great.  While Bullet Boy was a messenger, and it appears that he will be shot, I wonder if his management will suffer the same punishment. 
    Improve the process is the only way to be successful in my book.  I leave with one more comment.  The goal, objective, purpose of technology is to improve the condition of the human race.  Company profits are critical to achieving that goal.  If the end result is that the human race is marching backward through time for only a few people to reap the rewards, then we are doing something very wrong.  The question to me is the following was Bullet Boy trying to step up in the corporate world without concern for the consequences or was management pressuring him.  A person that robs a bank may steal a few thousand dollars and scare a hand full of people.  A president with support from other top managers that manipulates accounting numbers or facts reaps millions of dollars in their personal accounts.  These people should scare the world, which by the way they have. 

    0
    #88373

    Ropp
    Participant

    Summary of my fellow posters’ excellent points:
    1) Check for Normality
    2) Is Cpk the right index?
    3) Look at time-phased data
    My comments:  Sadly, it is mostly a paperwork game.  When I registered to QS-9000 on an existing product line, I found that all the PPAPs that the Customer had approved 2 years prior did not come close to the PPAP Manual standard.  I had to resubmit them with indices less than 1.33, which the customer was “forced” to accept since they didn’t want to redesign the part.  The customer also did not understand why I wanted to submit Ppk instead of Cpk.  QUALITY PLANNING CANNOT BE DONE IN A VACUUM.
    ALL STATISTICAL INDICES, LIKE CPK, ARE A JOKE, BECAUSE TOLERANCING’S GOALPOST MENTALITY HAS NO PLACE IN A 21ST CENTURY PRODUCT DESIGN PROGRAM.  THE CORRECT APPROACH IS TO MANAGE PROCESSES SO THAT YIELDS, HOWEVER THEY ARE DEFINED, ARE ACCEPTABLE.  As I recently suggested to a colleague, why not assign alpha symbols to all key characteristics and allow their target values to vary as required over time?  Then you would never be enslaved by an MRB.

    0
    #88374

    Et Al
    Participant

    Isn’t the idea of most of these “statistical indices” to provide some method of early warning, and ranking to help focus and limit resources to particular ‘problem’ items within a group, BEFORE yields are impacted?
    …lest we become just fire fighters – blindly running from fire to fire or, worse, doing nothing but sitting and waiting for the next fire.

    0
    #88375

    Bullet Boy
    Participant

    Denzel,
    Thanks for your thoughts. I hadn’t really considered the OEM measuring parts from failed units. However, I think I’m covered as I know for a fact the OEM cannot measure the characteristic they requested data for themselves.
    As for MSA, we got the gauge to pass & PSW signed off by the OEM by performing a static MSA with 1 part measuring 30 times. However, I can get the value to vary depending on the way I place it in the fixture. Luckily the OEM doesn’t know this, so even if they come to our plant to measure the parts, I should be able to get a lower measurement by  biasing the part to one side of the fixture.
    Thanks for the tip-off!

    0
    #88377

    Rex
    Participant

    Interesting. I’ve been in the midst of a failure analysis where the supplier reported that their Cpk, etc data refuted our hypothesis of failure cause and pointed us in other ‘possible’ directions – which we’ve expended time, resources and cost investigating. However, we’re not inclined to completely rule-out the supplier simply because we aren’t setup to measure it. Since we’ve ruled out most everything else, and this is seemingly all that remains, we’re having parts made by another supplier to the same requirements. This is new technology for the other supplier, but they are eager to try it and thus far we’re seeing very promising results; while their data suggests lower process capability and more defects than the current supplier, the parts they are yielding to us perform much better and we as well as our customer are willing to pay a higher cost for this new initial investment with a supplier who has now contracted with us to show continuous improvement. They have even recommended changes to the requirements specifying datums and restraint conditions that might better control the measurement of the characteristic relative to its intended interface in our application; they’ve taught as a bit more about GD&T and functional test on this product than we ever knew, so there’s a qualitive ROI that we believe is more than we get with the current supplier. Despite some shortcomings in our design requirements, as the new supplier has pointed out, we have enlisted the help from an independent lab to inspect the current suppliers parts and he is convinced we can show that they have failed to measure the parts according to our requirements and thus prove liability – which we intend to add the cost of failure analysis, which includes the rabit trails the supplier sent us down, as well as drop them altogether based on our new, eager, source of supply.

    0
    #88378

    Rick Pastor
    Member

    ETAL:
    Statistical indices: Different indices mean different thinks.  Etal said, “…BEFORE yields are impacted? ”  
    Cpk:   Cpk does not provide information before yields are impacted.  The purpose of  this statistical index is to provide information regarding the capability of your process.  For example, if you have a Cpk of 0.7 and you have been running at that value since the process was started, yield is poor and you have a process that is costing your company money.  The higher the value of Cpk the less scrap you are producing.
    Low Cpk Does Not Mean Poor Quality: Just because you have a small Cpk does not mean that your company produces poor quality products.  If you do 100% inspection, you might catch most of the poor quality before shipment to a customer. Unfortunately, you are now spending money on inspection and scrap when that money could be used for new product development. 
    A process with a large Cpk produces less scrap and with a large enough value you might be able to convert from 100% inspection.
    Dave said, “ALL STATISTICAL INDICES, LIKE CPK, ARE A JOKE…” Dave I wish you all the success in the world.  For the rest of us, I suggest that we take statistical indices seriously.  Of course, we must try to use the tools properly.  Statistical indices look into the process and tell us how we are doing.  Since there are statistical fluctuations for many sources in a process, indices change as a function of time, and we must track them to know how the process is doing.

    0
    #88379

    Et Al
    Participant

    Thanks Rick. I guess I was looking at it from the perspective that if I monitor my Cpk’s (I have several product and process characteristics being measured), and examine the data, control charts, etc for those with the lowest Cpk along with examination of the possible product or process causes, then I have a better chance of detecting processes that might be beginning to go out of spec so that I can adjust before they do go out of spec and effect yields. But I see your perspective as well. Thanks. 

    0
    #88382

    Mike Carnell
    Participant

    Eileen,
    Nice response. I am happy someone else sees that this isn’t a statistics question – it is an ethics question.
    Someone cited a quote earlier. My coauthor Barbara Wheat has one that seems applicable: “When you sleep with the devil, you’ll smell like brimstone.”
    I just wonder how many people would be offering up advice if the parts in question were killing people in car accidents. This thread really sucks.

    0
    #88386

    Psst
    Participant

    Bullet Boy –  I’m hoping you presented this REAL-WORLD issue to get a true measure of the response, and possible actions that could be taken to prevent them from happening to you; and I’m hoping that its not really your practice. It is about statistics, as well as ethics, and how they can be manipulated either intentionally or unintentionally, and how they can mislead the unwary. Over-looking and dismissing the problem (because it is ugly and does suck) denies a resolution, and helps nobody. From this perspective alone, I applaud you for raising it in this forum. I’m not sure how all the name calling, arm waving, philisophical responses and what-if scenarious help to address the problem, except maybe folks actually begin to think of it even if they do drop it to “hush-mode” just as quickly. I’ve heard consultants respond with everything from the same similar bullets, quickly moving on to the next subject as it throws a monkey wrench into all their neatly packaged toolkits, to the need for qualification and over-check on everything we do. I’d be curious to know what you come away with on this subject, or case study.

    0
    #88387

    Denzel
    Participant

    Mike Carnell & Eileen,
    Agree totally with the sentiment of both your responses, but sadly I believe that the appropriate action will never happen – As a BB for a OEM, to often we see suppliers providing misleading, inaccurate or even downright lies for data, and yet it is almost impossible to desource them due to single sourcing agreements, or even financial contracts that, despite what might be best, only focus on the initial (rather than lifetime) costs.
    And believe me, this includes safety critical systems.
    Henry Ford had a philosophy of bringing everthing he needed to manufacture his cars under one roof, so he had control of every aspect of the product. In a modern business world, this is no longer considered the best (read most profitable) way to work, and parts are sourced from “subject matter experts.” Now I’m not advocating that we should try and turn the clock back, but we need to find a way in which we are not competing on a pure cost basis with our suppliers/partners.
    If Bullet Boys OEM is like ours, he may well get away with it this time, but eventually he, or his company will get caught out. What happens then depends entirely on the people involved at the time.
    (And when I get back to the office, I will be double checking any supplier data submitted for my current projects.)

    0
    #88405

    Mick
    Participant

    I’m surprised that no one’s asked these questions yet…
    Is your process under control and is it capable (Cp=1.0 or better)?  No?  Then you shouldn’t even be discussing Cpk with an OEM.
    Are the points you “massaged” away legitimately removeable because of special causes in the process (and hence have been rejected and scrapped)?  If they were due to special causes, were corrective actions taken to bring the process back into control and eliminate the special cause? 

    0
Viewing 24 posts - 1 through 24 (of 24 total)

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