Anirvan Sen

If the problem appears within 4 months of implementation, then as somebody mentioned there could be seasonal impact or other factors that influence your outcome.
One more point it highlights is a component of control mechanism which is “mistake-proofing(poka yoke)”. In other words, your mistake proofing should have been able to capture a…[Read more]

Depending on the density of calls, you are better off with a week (fiscal or whole week is up to you). A month maybe too large a sub-group especially, with the helpdesk cases and may have a lot of noise.
To answer, you second part if your Y is cases in helpdesk then your y can be cases per group.  Use your grouping/sub-grouping judiciously (tak…[Read more]

When u use Discrete data samples, it makes histograms. As the number of samples increase, the shape of the distribution starts looking more like an ND. You will have to plot the datapoints to figure out whether the distribution is normal or not.
I’m not sure if I understood your second question. For normal distribution, the curve follows a…[Read more]

Since it is for the org. I work for, therefore, may not be able to share all the details. However, if you want very specific input, I’m more than happy to share.
You can send me an email at [email protected] or use this forum.
cheers,
Anirvan

Kaizen is continous improvement and Kaikaku is one-time improvement. Depends on a lot of factors, which one should you follow. If there are cost implications, then you may want to do a Kaikaku.

It was Beethoven who became deaf and blind by the time he composed his last composition.

Surprisingly, this time I’ve to agree with Andejrad. Indeed you are looking at big “Y” at the end of the process and a small y at your process level. Moreover, look at your control limits and then see how well does your process perform (and then plot a SPC at look at how the data points w.r.t. 1-sigma; 2-sigma; 3-sigma…and you could go upto…[Read more]

Andejrad, +/- 6 SIGMA makes sense to be set as a target where one would like to see the process to be in. Chances are that the process in question, probably is far away from that level. By keeping the levels at +/- 3 sigma, you can get a fair understanding of your process behaves at this point in time.
There is no right or wrong approach.There…[Read more]

Sorry, I accidentally hit the post button without the complete answer.
One of the assumptions, you could make is to have the tolerance limit and the control limit as the same. This method is not full-proof but gives you a level of understanding of how good or bad your current process is.
A.

Tommy, Your tolerance limits are indeed your specification limits and usually need to be set by your customers. In real life, however, you often get a subjective problem e.g. length is not right…it is too long. 
Your tolerance limits do not affect your control limits (however, when you compare your process vis-a-vis the two limits, you draw c…[Read more]

Cp or Cpk uses spec limits (voice of customer) and Pp or Ppk uses the natural limits of your process (voice of your process).

Make sure nobody hits below the belts ;-)
A.

I Agree with FSC!! Moreover, discrete focusses on whether compared to continous (which focuses on by how much). Moreover, in general, you will need more number of samples for discrete compared to continous to arrive at the same level of statistical evidence.
A.

You could always use Capability Analysis using any std. statistical software.
A.

HornJM, I agree with you on getting MBBs and 6-sigma champions (Change Leaders). I had assumed that since they had already started working on 6-sigma therefore, they would have some MBBs and/or 6-sigma champions (Change Leaders).
Thanks for the correction.
Cheers,
A.

Sum of Squares refer to the sm of squared deviations. Therefore, it will be represented as :
summation of square of deviation of individual points and the mean of the subgroup.
 SSTotal: assesses the deviation of each individual point from the overall mean

SSBetween:assesses the deviation of each subgroup mean from the overall mean…[Read more]

CG, it refers to the variation within and between subgroups to get a better understanding of the special causes and compare short term and long term effects.
The formula is SStotal = SSwithin + SSbetween
where SS : Sum of square of variation
For your second problem, I need more details. At the same time, you may want to look at the way, you…[Read more]

Ben, when it comes to training even though Green Belt training is easier compared to Black Belt however, saying that there is no mandatory rule that to get a black belt training, you need to get Green Belt training earlier.
However, since your org is just getting started you probably are better off sending your own people for GB Training and get…[Read more]

Steve, the more I read your question the more intriguing it gets!
“My question is: Is it valid to remove special cause, 3 sigma zone violation, data from a capability study (even though I do not know the reason for the special cause event)?”
When you say valid, what is it that you are trying to validate?
Remove special cause: you are referring…[Read more]

Steve C., I’m assuming the reason why you would like to remove 1 outlier and check the capability, is to see whether the rest of the process is capable or not. You may want to substitute the outlier with a value that is the mean of the rest. This is a very crude measurement but atleast gives you an idea about the overall process.
If you indeed, f…[Read more]