Designing with Data provides an extensive background to A/B testing.
As with all other research methods, we need to start with a research question. A/B testing concerns itself with changes in user behavior, meaning that our questions need to be centered on measurable goals. In many cases, these will be conversions already defined as goals in our web or app analytics. In the video William Hudson discusses good and bad examples of research questions as well as introducing a few advanced topics:
References & Where to Learn More
Rochelle King, Elizabeth Churchill and Caitlin Tan (2016), Designing with Data: Improving the User Experience with A/B Testing
© Amazon, Fair Use