Idan D.

So the results you’re looking for is the one on the right, the one on the left shows different results because: 1. it seems like the cohorts are different 2. I wish i knew why it’s like that but it’s just the way cohorts work, if you want to see how many users are in the cohort at any given time you plot it overtime.. when adding filter it seems as though just the latest period makes sense

I hope i understood the question correctly, Instead of “$Track” you should use “$Set” - Is that what you needed to know?

Give it a try, let me know what you get.

Alex K. your cohort is dynamic so if you create and insights report with that cohort (set to 12 months ago) you will get what you need: the unique active users at any given time.

Do you have group analytics set up? The use case you’ve mentioned is a good fit for the feature. You can also create a cohort per agency and then create a set of reports and filter by cohort. Reports at the user level would only be readable if there aren’t many users per agency, otherwise you can also flip it and show them how many users per “Usage Bucket” (you can use distribution on an event to do this and pre-define ranges of usage buckets)

You can locate the users that were associated to the same mock user id by checking the identify history under Distinct ID, create a cohort of the affected users and then fix manually for that set of users

If i’m not mistaken, it would only be possible if you add them as 2 different cohorts, new on a specific week, and then new a week before and retained the week after

I usually export to google sheets when I need to do complex things - what are you trying to achieve maybe there is another way I could suggest (in mixpanel)

I agree, many places in the product would benefit from a tooltip of what stands behind the logic