Let's picture a scene:
What should we do? Pause ad A or ad B?
First, Pareto Ads would run a conversion experiment. With this, we would see that the significance level is only 78% for conversion, which is inconclusive.
Then we would run a CTR experiment, leading our software to conclude with a 98% significance level that ad A is better than ad B.
So even if your campaign is targeting a better ROAS, you should go ahead and pause AD B, since it's better ROAS has no significance level.