iSixSigma

Xgames

Forum Replies Created

Forum Replies Created

Viewing 10 posts - 1 through 10 (of 10 total)
  • Author
    Posts
  • #186349

    Xgames
    Member

    Please enlighten me Stan…..

    0
    #186340

    Xgames
    Member

    With lots of data (greater than 25 or 30), the central limit theorem takes cares of things (when checking for differences between the means).

    ANOVA is pretty robust to normality, though….the assumption in the model is for the residuals to be normal.  If they’re grossly skewed, then will most likely need to do a transformation on the data.

    0
    #178807

    Xgames
    Member

    DLW..thanks and just to clarify I am a “him”. I like your point on data driven analysis. There are a lot of measurements you can use to show improvement using lean without attaching a sigma level. To name a few..reduction in NVA activities, cycle time, distance traveled, and paper reduction (PC). Like you mentioned the name of the game is to continually improve.
    Krshna..I agree with your statement about cost benefit analysis vs. improvement. However with my IT group we use the DFSS tools with great success. You mentioned CTQ these are developed using QFD. This will also rank your requirements to show where you can make that cut off regarding bugs. The DFMEA is also critical to mitigate any risk. If this is all done upfront your risk of having bugs and rework in the first place is greatly diminished.
    my 2 cents
    Xgames

    0
    #178782

    Xgames
    Member

    I work as a transactional BB. We did one project this year in which our starting sigma level was .63. We improved the process to a sigma level of 1.05 (nothing earth shattering). There are several factors which prohibit this process from producing at a 6 sigma level but as Outlier stated it doesn’t neccessarily need to be. Our savings were close to 1M. I can tell you that many folks weren’t worried about the sigma level. The process is also very manual and spawned a DFSS project to improve it even more but we justified that the benefits were there. We’ll probably wind up around the 2 sigma level but that will be a huge success in everyone’s eyes.
    Another thing I’ve found is that the lean toolbox works great in transactional at eliminating waste and the best part is you don’t need sigma levels to show you’ve made improvement:)

    0
    #177841

    Xgames
    Member

    I’d start with the 4-5 general issues and work with the customers to drill down to a more specific issue (before and during service). At the same time you can query them as to what is important from their viewpoint. I might have nothing to do with the 4-5 general issues. From there you should be able to build your criteria for what is CTQ.

    0
    #177811

    Xgames
    Member

    What are you tracking now that says you need to make an improvement?

    0
    #175763

    Xgames
    Member

    Udayan,
    Use the sumproduct function (=SUMPRODUCT($D4:$D23,F4:F23) where the “D” column would represent your weights for your customer requirements and “F” column being what you scored your design requirements (how’s). That will give you the weighted importance of your “how”. Then divide the weight importance of that “how” by the sum of all the weighted importance of all “how’s” and that will give you the percent importance of the “how”. From there you can see which are the most important.

    0
    #175013

    Xgames
    Member

    Really…so if I tell you recruitment is all the rage you’ll do a project on that??
    You need to understand the VOC and what is causing the pain in your process and go do a project on that.

    0
    #174734

    Xgames
    Member

    Takt time is your hours available divided by customer demand.

    0
    #168789

    Xgames
    Member

    Rasu, you are correct there are only 2 types of data. What your BB is talking about are your inputs and outputs (x and y). If you had a discrete x and a discrete y you would use a chi-square test etc.

    0
Viewing 10 posts - 1 through 10 (of 10 total)