A/B Testing Conversion Rates
Using the Beta-Binomial model, we introduce a common pattern to perform Bayesian A/B testing. There are three strong reasons to prefer Bayesian A/B testing over traditional hypothesis testing: uncertainty visualisation, interpretable statistics and more flexibility of computing additional statistics.
- What is A/B testing?
- How do we use Bayesian statistics in A/B testing?
- How do we determine if group A is better than group B?
- And by how much is group A better than B?
We Also Recommend