Bayesian Predictive Probabilities for Online Experimentation

TL;DR

This paper introduces a Bayesian predictive probability system for online experiments that allows interim analyses without inflating error rates, demonstrated on Instagram data to improve decision-making efficiency.

Contribution

It presents a novel Bayesian predictive probability approach for interim analysis in online experiments, avoiding numerical integration and ensuring experiment fidelity.

Findings

Enables error-controlled interim analyses in online experiments.

Reduces reliance on numerical integration for predictive probability estimation.

Demonstrates practical benefits using Instagram experiment data.

Abstract

The widespread adoption of online randomized controlled experiments (A/B Tests) for decision-making has created ongoing capacity constraints which necessitate interim analyses. As a consequence, platform users are increasingly motivated to use ad-hoc means of optimizing limited resources via peeking. Such processes, however, are error prone and often misaligned with end-of-experiment outcomes (e.g., inflated type-I error). We introduce a system based on Bayesian Predictive Probabilities that enable us to perform interim analyses without compromising fidelity of the experiment; This idea has been widely utilized in applications outside of the technology domain to more efficiently make decisions in experiments. Motivated by at-scale deployment within an experimentation platform, we demonstrate how predictive probabilities can be estimated without numerical integration techniques and recommend systems to study its properties at scale as an ongoing health check, along with system design recommendations - all on experiment data from Instagram - to demonstrate practical benefits that it enables.