A/B testing has become a popular method to investigate user behaviour on web sites , but the learning we get is of little value I think. Narrow the focus on "A or B ?" allow us to collect "encrypted" (more or less BIG) data about users behaviour, without no decryption key other than our cognitive biases. My latest post : "A/B Testing Is Fooling You"
Join over 70,000 growth pros from companies like Uber, Pinterest & Twitter