A/B testing is the most common way of establishing causality and proving that a feature or a product change results in an improvement of your metrics. This talk will give an overview of what A/B experiments are and provide a deep dive into most common challenges and pitfalls, and how to avoid them.