Hey @rahulsekhar 

From what I understood from your use case, I would have suggested using the Active % metric in the event segmentation chart using your Attendance event. But Active % graphs only the percentage of users who fired a specific event, compared to the total number of active users at each data point. So the denominator would be the active users within that datapoint and not all your subscribers in your cohort.

So if there were 100 subscribers in total for your product in April 2022, 80 were active during April 2022 and 40 of them fired the attendance event, then the % you are looking for in April will be 40% ( for the 100 ) and not 50% ( the active % ). Is this the right understanding for your analysis?

I haven’t quite found the best way yet to apply a static cohort count to a custom formula, but manually applying the count in an interval bound bar chart views seems like the short term approach.

In this demo chart, I have equated  “Money Transfer Completed” as my event for Premium Users in April 2022 assuming that my overall ( active + inactive ) premium user count for April is 100.

Another approach would be adding in an user property like events_attended and update the counter once the attendance event get fired.

Let me know if I have interpreted this correctly or otherwise.

Hi @rahulsekhar! Besides the methods that Saish have suggested, you can look into using the ‘average’ metric as well: 

The Average metric graphs the average number of times a specific event was fired. Here, the "average" for any data point is equal to its event totals divided by unique users.

Hi Saish and Ning. Thanks for your responses! Your understanding is absolutely correct, Saish - I did try Active % but the gap you described is why it didn’t fit this analysis. I believe Average would behave the same way, Ning?

I’d ideally like the cohort count to also change over time, so that the chart of attendance/member is independent of member growth over time. One idea I’ve had is to log an event each day for a fictional internal user, specifically to track the count of active members and add that event to the chart - wonder if that might work?

Saish can you say a bit more about how you’d use a user property like “events_attended” for this analysis? We actually do have a property like that.

Yea, the Average metric would work out the same way.

My initial thoughts on something like events_attended were that you could use the Average per property value, but then this is applicable to only the active users during the given time period. Not sure if this would satisfy your use case of looking at your entire user base.

The fictional internal user event logging might help, but I‘m always skeptical that it might mess up my data governance.

So tracking the daily average might be tricky.But if every subscribed member has the event like “Registered/Subscribed” , then calculating the aggregate average over a time interval becomes a bit easier using Total Sessions that have happened / Total count of “Registered/Subscribed” events.

There could be some other obvious and simpler approaches to this one which I might be failing to formulate here.