Howdy! I'm running into a small issue with running an experiment from when a user goes from unauthed to authed. What I want to happen is: 1. User lands on page unauthed and an id is assigned to them anonymously, experiment starts 2. User auths and I call

growthBook.setAttributes({
        ...growthBook.getAttributes(),
        id: user.userId,
      })

3. The experiments do not start again and the anonymous ID is merged with the user's real ID from my system Current state the experiments are called again and it messes up my event analytics

We had the same issue. We opted to cookie every visitor a unique id which we then set as the

deviceId

in attributes and used that attribute in our a/b test. But it means the same user can be in different variants if they use another device.

That's a good workaround. Idk if I feel like coding that though 😅 . We're using mixpanel for analysis and they have a

First time ever

filter which works for us, but it's a foot gun to explain the nuances of when you need it when doing cohort analysis