As a general rule, it is the other way around - Pp and Ppk are used for the initial process study. This is when you still have a limited population to sample from.
Once you go into serial production, you should be using Cp and Cpk to monitor your ongoing process capability.
With that being said, I have had customers insist that I do it the other way around - if that is what they want, then that is what I give them.
I realize that this is your experience. However, it is not universal as you point out that different Customers ask for different things.
the whole designation of Ppk and Cpk is what is confusing. What really matters is:
- the intent of the study (long term or short term)
- The time period and availability of the data (pre launch or post launch)
- the formula used for the standard deviation (within subgroup or total standard deviation)
the terms Cpk and Ppk are indiscriminately interchanged by just about everyone and so it essential to confirm the above.
I know one customer who simply uses Cpk and its the original formula with total standard deviation. I know some call this Ppk. I know one that uses Cpk for short term and it has the within subgroup standard deviation formula. I also know some Customers use Ppk designation with total standard deviation as the approach for short term capability during pre-launch when there is limited data and no subgrouping is done...in the beginning there was only Cpk (no Cp) and it was only long term post launch using the total standard deviation. then the madness started
it is just plain crazy - this complication and confusion are not necessary but everybody just has to be 'different'. (and all for a completely useless index)