The Risk of Losses in A.B. Tests

When you have a big sample size, that also means big risk. So per unit time that you're running the test, you have bigger risk of losses if these variants are worse than the control. Or if they are truly better, you're running a higher risk of missing out on revenue. I would actually recommend obviously to just run the test on average revenue per user when you can,. But yeah, that would be a side topic to consider.