Hi dears, I have a questions about experimentation splitting. Context: We are testing using test users growthbook splitting, on QA envs of ios and android, in order to understand that feature toggle is working as feature and to conduct A/A test on prod. Firstly, I created experiment with splitting 50%-50%(feature toggle), but all 7 users assigned to False (control group). So, in order to quickly understand that everything is working and to not to use so many test users that we have, I deleted experiment, and created new with the same name with splitting 0%-100%, so it means all users must assgined to variation_id=1. But, when we go through using previous used users they're still assinged to False. Questions: Does it mean that he some problems on side of our mobile? or growthbook somehow see that user is have been in False group in the experiment with the same name historically(with different splitting rule)? and that is why this user still assigns to False, while with new rule should be assigned to True, logically saying. recreated experiment with the same name(feature toggle was recreated too, with the same name), but with different rule:

Hi Amina, thanks for your question. If you are using the same tracking key in the new experiment, and the same start date, these exposure events will still be in your events table and be logged in your new experiment. You may want to start a new phase or modify the dates of your current phase. Best, Luke