Hi all! I thought I'd post this here as well, since I know development teams often do extensive data analysis.
We are beginning several new data analysis initiatives here and we’re curious about best/recommended practices.
We’ve identified several goals for acquisition, re-acquisition and retention and now need to do some trend analysis and segmentation of certain groups. However, we’re unsure of the best way to pull this information. Do most of you use SQL queries, or Lists and/or Extractions, any other methods?
Second, what kind of verification process do you have when doing SQL queries? We have one person who can do SQL queries, but we’d like to develop checks and balances to verify the data is accurate to the best of our ability.
Any information you can provide about your process is appreciated!
Thank you!
Marley
I'm not sure what kind of analysis you're thinking about, but I like extractions whenever I'm trying to do anything involving segmentation. It's the simplest way I know to make sure your constituents don't end up in more than one segment through tiny (or not so tiny) mistakes in your lists/queries. Plus, sometimes it's easier to break complex questions down into several smaller segments and/or suppressions rather than trying to fit all the conditions you need into one list.
For larger projects I've also found it useful to put the full results of the extraction into Excel, and then I can use =vlookup to join in other data I've pulled via output sets or reports. Once you have all of that information in one place you can start to do strong analysis quickly with pivot tables or charts or however else.
As for the second part of your question, the only thing I can really offer is test, test, test! Wherever possible I like to try to check the numbers from my extractions against numbers from reports. The suppression report also comes in quite handy when one of your segments seems suspiciously small.
And of course, if you can answer any of your questions with T-Stats, you should give that a try first. Good luck!
This is great information! Thank you!