[encoderer]

Alright, last comment from my side, just to clarify:

>> You are literally saying that if the result is close enough, you will prefer not to reject the null hypothesis, and therefore whichever variation you have arbitrarily chosen to be your null hypothesis will be the answer.

This is a misunderstanding. The null hypothesis is that your two variants have no statistical impact on conversion and any edge you see is just random. That is the hurdle you have to overcome to gain any useful direction from B testing.

GL!

[Silhouette]

Fair enough, my phrasing before was a little casual, but the underlying point is sound. A hypothesis test might tell you that there is no significant impact on conversion at your chosen level. However, you still have to make a choice between option A and option B. If you have no a priori reason to favour one as the default and no additional data to consider -- which, again, is a crucial detail in the situation closed was talking about -- you should logically still choose whichever option that was most successful during your experiment. This is simply because if your conclusion was correct and there is no impact on conversion then which you pick doesn't matter, but if your result was a false negative then it is more likely that the more successful option during the experiment is the better choice. Given that you're going to pick that one anyway, your hypothesis test hasn't actually provided any useful information to help inform your decision in this scenario.

In any case, we seem to be talking at cross-purposes here, so perhaps we'll have to agree to disagree on this one.