Large-Scale Online Experimentation with Quantile Metrics

TL;DR

This paper introduces a fast, scalable, and statistically valid method for conducting A/B tests with quantile metrics, addressing a key challenge in large-scale online experimentation.

Contribution

It presents a novel variance estimation technique for quantile metrics that is both accurate and computationally efficient, suitable for large datasets.

Findings

Achieves over 500x speedup compared to bootstrap methods.

Maintains only 2% difference from bootstrap estimates.

Provides a scalable solution for quantile-based A/B testing.

Abstract

Online experimentation (or A/B testing) has been widely adopted in industry as the gold standard for measuring product impacts. Despite the wide adoption, few literatures discuss A/B testing with quantile metrics. Quantile metrics, such as 90th percentile page load time, are crucial to A/B testing as many key performance metrics including site speed and service latency are defined as quantiles. However, with LinkedIn's data size, quantile metric A/B testing is extremely challenging because there is no statistically valid and scalable variance estimator for the quantile of dependent samples: the bootstrap estimator is statistically valid, but takes days to compute; the standard asymptotic variance estimate is scalable but results in order-of-magnitude underestimation. In this paper, we present a statistically valid and scalable methodology for A/B testing with quantiles that is fully generalizable to other A/B testing platforms. It achieves over 500 times speed up compared to bootstrap and has only $2\%$ chance to differ from bootstrap estimates. Beyond methodology, we also share the implementation of a data pipeline using this methodology and insights on pipeline optimization.