A practical guide to designing reliable A/B tests, from hypotheses and metrics to MDE, statistical power, sample size, ...