2 Replies Latest reply on Mar 2, 2017 2:20 PM by Patrick Van Der Hyde

    Cohort Analysis Using Blended Data on Tableau Desktop?

    G Lambert

      I'm trying to do a basic cohort analysis in Tableau. I've already read other questions and the linked Tableau KB articles on how to do cohort analysis. Those all work well if your data is all in the same database. Mine is not.


      I'm using session data from Google Analytics and my data for the date the user was first seen is actually in an Excel file. I didn't set this up, just working with what I have. The two data sets are linked by a common user ID number. I read that I can't join GA and Excel data, that I have to use data blending.


      I just want it to look similar to this:

      Screen Shot 2017-02-21 at 4.08.38 PM.png

      where it shows me the number of users that had a session that month, for each consecutive month. The Minimum month on the left (the cohort) should be based on the date the user first joined (which again is in the Excel file).


      I tried to do it a couple of ways. I tried COUNTD, but it didn't work, I think because of the way it was aggregated. I also tried to do the DATEDIFF function - to get the number of days between the user's first appearance and their current session, but it didn't work because the date first seen data was in the Excel file and aggregated as ATTR.


      Is what I'm trying to do just impossible in Tableau Desktop 10.1, since my data has to be blended? It's fairly easy to do in Excel, but the data files are getting too large for my laptop. I thought Tableau would be (somewhat) simple.


      Is there a way for me to use the Date Created info in Excel with the rest of my data in GA? Perhaps I'm putting things on the wrong shelf or I'm not thinking about the right way to connect the data.


      I haven't attached a workbook, because the question is about connecting GA and Excel.