This message was deleted.

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