HI Matheus - your conversion numbers are still quite small, even if the denominator is large
fresh-football-47124
04/12/2024, 7:44 AM
its not uncommon for experimentation platforms to set a 200 conversion event limit before even showing results. You have 70 vs 93 and 44 vs 66 - which is not unreasonable that this is just natural variance or skewed by a few outliers (you can add percentile capping or winsorization to help see)
fresh-football-47124
04/12/2024, 7:44 AM
I believe by default we use 150
s
sparse-needle-41578
04/12/2024, 8:10 PM
That makes sense, thank you so much for looking into it! I changed the limit back to 150 and I'll keep the experiment running