Beast mode to calculate month over month user count changes

sgoldina

One of the metrics we calculate as a business is our user change month over month. Our main data set we work off of is an activity log dataset which contains the date of the activity, the user id, and details about the activity. For active users we calculate the distinct number of users for that time period such as the month.

 

However we would also like to look at the changes in users month over month:

resurrected users - users that had no activity last month but are active this month.

returning users - users that were active last month and this month.

new users - users that have had their first activity this month.

 

Would it be possible to compare these user arrays using beast mode or would we be able to create a table using Magic ETL somehow?

jaeW_at_Onyx

It would not be possible to define your user states (Is Resurected , is Active etc. in Beast Modes with month over month comparisons in Analyzer.  You would have to be presenting the data at the UserID granularity in order to calculate the desired metric (because you'd have to have User in the GROUP BY CLAUSE ...)

 

INSTEAD what you can do is create a UNIVERSE of Months and User IDs where the granularity is one Row per User per Month since they were added to Domo and then calculate your isResurected, and isActive metrics as binary / boolean columns.  

 

Then you'd be able to cleanly perform your math without having to rely on Count Distinct.

It'll sound like a stretch, but this tutorial:  https://www.youtube.com/watch?v=Xb4QgKYgaqg&list=PLUy_qbtzH0S4CkHBUvpOVpLNJluk6upOn&index=26&t=418s will apply to your use case.