03/09/2023, 5:12 PM
Hey 👋🏿 ! We’ve been using GrowthBook for a while now, and for the first time we’re faced with a SRM on one of our A/B test, we’ve checked the implementation, and don’t figure out what’s wrong.
The help message attached to it says we should check: • Bucketing • Execution • Analysis • Interference
but I believe we don’t really have a hand on bucketing since GB does it by itself
03/09/2023, 5:37 PM
Hi Gania - it can happen if you changed the split percentage
how many users are in your experiment currently?
03/10/2023, 8:36 AM
We've changed the split at the beginning of the week
Can we set the beginning of the experiment in the board?
03/10/2023, 9:13 AM
03/10/2023, 9:15 AM
we actually found the answer
we were looking at the SQL queries, and weren’t sure where to choose the start date
thank you for hinting me the split % change
03/10/2023, 9:22 AM
glad I could help
if you change the overall exposure, you don’t have to change the split to ramp up the experiment, and you wont get multiple exposure problems or SRM problems