Analytics & Metrics

Sample Size

Quick answer

How many observations a test needs before results mean anything.

Definition

Sample requirements scale inversely with effect size: detecting a 20% improvement needs far fewer users than detecting 2%. Low-traffic sites testing headline colors chase differences their volumes cannot statistically perceive.

Why it matters

Calculating required sample before launching a test kills the most common failure mode - ending experiments at 80% completion and shipping noise.


All glossary terms

Your competitors booked
theirs.

Onboarding slots for this quarter are limited. We audit your funnel, show you where it leaks and what your next 90 days should look like - zero obligation.