hi! we have launched our first test using GrowthBook. we used assignment by deviceId, but when we analysed raw data, we saw that the same deviceId was assigned to group A, and then to group B. maybe anyone could explain why this issue can happen? thanks a lot!

Hi Hanna - did you change the percent split at all?

hi! no, we didn’t change the split

we use deterministic hashing, so as long as the experiment name hasn’t changed, the same deviceId will always give the same variation