Skip to main content

Hi all,

I’m trying to formulate a question into an Amplitude query, but I’m not sure exactly how to. 

I have a view_page event (for any page view on our site) and a start_session event (for any session on our site). The simplest version of the question is: if someone performs more page views in a session, does that make them more likely to come back? Does a higher view_page divided by start_session mean more start_sessions in future?

I am ideally looking for an indicator of causation, but I’ll settle for correlation at this point.

I can create a segmentation chart that does a formula on totals(view_page)/totals(start_session) to see averages, no problem. I can even segment this view by people who have done 1 session last month v.s. 2 sessions v.s. 3-5 sessions etc etc. But I kind of want the inverse of this - to count the sessions, but segment by the totals(view_page)/totals(start_session) value.

The retention chart seems like it would be the right place, but gives me no way to do the totals(view_page)/totals(start_session) part. 

I’m stumped and would appreciate any advice on how to do this! Thanks

I have found a way to answer this question when I manually specify criteria - i.e. I have to define a set of tiers of people who did X page views in Y sessions over Z period, and then compare their average number of sessions during period A (which is later than Z). But how can I do this generically?

 


Reply