https://www.growthbook.io/ logo
#ask-questions
Title
# ask-questions
c

calm-lamp-46034

02/22/2024, 4:36 PM
Hey team, we are using Attributes for the first time to filter out users from our analysis. We are passing a boolean flag, which gets updated after our experiment flag is already triggered. We later update the boolean flag using .setAttributes(). It seems that the attribute is set correctly, which we confirmed using .getAttributes(). However, users are not getting filtered at all in our Experiments (we tried using Target by Attributes and Target by Saved Groups). We're stuck at this point and not sure how to proceed, so would appreciate any help.
f

fresh-football-47124

02/22/2024, 4:51 PM
Changing the attribute may cause the user to get included or excluded from an experiment - but if the initial state is included, the trackingCallback will fire, and you'll see that user in the analysis
sounds like what you want is activation metrics
c

calm-lamp-46034

02/22/2024, 4:55 PM
I see, so updating attributes won't retroactively update analysis? I'll check out activation metrics
I'm actually thinking of using Segments rather than activation metrics here. Seems like activation metrics are event based, but our filter is on a user property. I guess we could pass setting the property as an event also, but that seems redundant. The identifier type for the segment is 'distinct_id' while the feature targeting attribute is 'id'. Would the difference here be an issue?
f

fresh-football-47124

02/26/2024, 11:15 AM
should be okay
you'll be able to see the query GrowthBook generates
10 Views