Not natively, but you could do it through feature flags. Basically, get the value of the feature and then do the redirect based on the value.
We don't recommend doing redirects as part of an A/B test if you can avoid it. It tends to affect the results since the new variation takes a significantly longer time to load. Doing the A/B test at the edge or in your backend is typically better if it's an option with your setup.
10/31/2022, 1:42 PM
The main problem I am trying to solve is that we need to split the traffic before the user loads the page. The problem with that is that in the BE we can only do for users that already registered so far because then we use our ID to identify then but for users that did not signed up we use mixpanel which we cant access in the BE.
The idea is for example users try to access url A but before loading it we decide which group he is, if he is control he stays in url A but if he is variation we send him to url B.
10/31/2022, 1:59 PM
Doing the redirect on the front-end is definitely possible, but it will introduce bias to your results. The variation will start with a deficit since it's adding an extra page load to the user experience. So it has to be a lot better to overcome that and win.
One thing you could do is to add your own unique visitor id cookie on the back-end. That way, you have a way to identify someone before the app loads and can split traffic on the back-end, even for anonymous users.
11/01/2022, 6:53 AM
I think that for that I will had to find a way to add this id and alias it in mixpanel since we use it for tracking, I will give it a try