Is it possible to exclude users from the experiment (not just metrics). We want to include only new users but we don't know if they are new until they log in. Can we at that point somehow exclude the ones that are not new?

you want the experiment to target anon users?

Kind of. We want to exclude users who have already posted projects on our platform. But we only know that after they log in. And we're running an experiment on a page where you don't need to necessarily log in yet. So we're wondering if we can place them in the experiment but exclude them later somehow. We can exclude them from metrics (based on some custom events/data we could save in our data warehouse). But that doesn't change the number of experiment participants thus skewing the metrics. Not sure if I'm explaining this well enough. 😕

you can use activation metrics for this

Ok, and that can be set afterwards with analysis setup?

yes

Ok, perfect.

its used as a precondition for experiment analysis inclusion. So for example, if you had to bucket users in an experiment before you were sure if they saw it or not, like if there is a modal window- then you can add a metric for the 'modal opened' event, and add that as the activation metric for that experiment

Ah ok, I see there's a special metric you assign during analysis. I thought all binomial metrics act as activation metrics. That didn't make sense to me. Now it's clear. 🙂