Skip to main content

Hey everyone, quick question!

Does anyone know know how to measure the average number of times a user has triggered an event, viewed cumulatively over time rather than period by period?

  • For example - I have an app where users can upload their recipes and I want to understand whether users are continuing to upload more recipes over time
  • Ideally, I want to set up a line graph that shows me whether or not the average number of recipes per user is increasing over time
  • i.e. for the week ending 20 August, the value on the graph will be the average number of recipes per user since app launch until 21 August
  • for the week ending 27 august, the value on the graph will be the average number of recipes per user since app launch until 27 August
  • The closest I’ve gotten so far is creating a line graph that calculates the average number of recipe uploads in a given period, but not over time

Hope that makes sense 

Thanks @willkent for posting this here! 

This can be achieved with a custom formula in a segmentation chart using CUMSUM and AVG.

Here’s the logic:


AVG is the average number of times a user has triggered event (A), in this case Connect With Friend. AVG(A) = TOTALS(A)/UNIQUES(A)


CUMSUM will display a running total of events in a single data point and deduplicate uniques for you. In my example, with the custom formula CUMSUM(AVG,A), it shows me that between July 1st and August 15th, the average number of times a user has Connected With Friend is 2.94 in this time period. Between July 1st and July 31st, it is 2.55.

Reply


Terms & ConditionsCookie settings