Follow Up on Metrics Webinar: Filtered Denominator for Retention Metrics Hey @helpful-application-7107, thanks for the Q&A today! 🙏 For our business stakeholders it is quite confusing if they see different retention metrics from our dashboards (where we filter out users that did not yet have enough time to return), so would be great to replicate that in Growthbook. We do have a table that stores the user activity on a daily basis with one row per user and date.

Right, I think the best way to do this is to create a funnel metric where the denominator is a 1 if 7 days have passed or not, but I don't think we have a great way to do that right now. I think it's a pretty good solution. There's also a way to create an Experiment Dimension which does the same thing, but that would require more work to show to your stakeholders.

Right, I think the best way to do this is to create a funnel metric where the denominator is a 1 if 7 days have passed or not, but I don’t think we have a great way to do that right now. I think it’s a pretty good solution.

Okay I tried this with two metrics: • Users Returning after 7 days with a 168 Metric Delay • denominator: Users exposed to the experiment more than 7 days ago (no metric delay), defined as this:

select
  userid,
  timestamp
from assigned_events
where date
    between date('{{date startDateISO "yyyy-MM-dd"}}')
    and date('{{date endDateISO "yyyy-MM-dd"}}')
    
    and timestamp
    between timestamp '{{startDate}}'
    and timestamp '{{endDate}}'
    
    and date_diff('day', date(timestamp), current_date) > 7
    
    and experiment_id like '{{experimentId}}'

But it looks this doesn’t work with an activation function as the generated query filters out events that happened after the activation, but the exposure always happens before the activation 🤔

I think you should add 7 days to the timestamp in that denominator metric, but keep the filter as is.