This message was deleted.

One sample as 110, the other has 66. If you need to see results not, set the min sample size to <66

@fresh-football-47124, regarding this rule, I am confused why Growthbook considers that the number of conversions is what matters for statistical significance, not sample size, in particular for very rate events. Surely, if control has 5 events / 1M users, and treatment has 20 events / 1M users, it should be statistically significant, no?

yes