Sometimes you need far fewer customers than you think for an A/B test

I might be wrong, but sometimes you need far fewer customers or less data than you think to make a decision through an A/B test or experiment.

When the difference is striking, the winner is obvious!