Glossary · Business Analytics

Bayesian Inference

also: Bayesian statistics · Bayes' theorem

Definition

Bayesian inference updates prior beliefs about a parameter using observed data via Bayes' theorem to produce a posterior distribution. In A/B testing it directly answers 'what is the probability that B beats A?' — the question product teams actually ask — unlike the indirect counterfactual framing of frequentist p-values.

Bayes' theorem P(θ|D) = P(D|θ)·P(θ) / P(D) is the foundation. Practically, the posterior is often analytically tractable for conjugate models (Beta-Binomial for conversion testing) or sampled via MCMC for complex hierarchical models. Bayesian A/B testing allows continuous monitoring without inflating error rates, eliminates the peeking problem, and produces directly interpretable probabilities and expected-loss estimates.

Essays on this concept