## Metadata - Author: [[Mårten Schultzberg]] - Full Title:: Why Spotify Is Not Using Bayesian a/B Testing - Category:: #🗞️Articles - URL:: https://engineering.atspotify.com/2026/9/why-spotify-is-not-using-bayesian-a-b-testing?utm_source=substack&utm_medium=email - Read date:: [[2026-10-03]] ## Highlights > Any experimentation program produces false positive and false negative results over time, regardless of whether its inference is based on Bayesian or frequentist philosophy ([View Highlight](https://read.readwise.io/read/01m40zvwaspdja3w1rajx73xqh)) > the posterior remains a valid belief update regardless of when or why data collection stopped. This is correct, but it is a narrow guarantee. Error-rate control, estimation precision, and decision-theoretic performance all depend on the prior and stopping rule. ([View Highlight](https://read.readwise.io/read/01m41034q44qfmb0qa7xaj0gs6))