experiment-metrics

Choose trustworthy A/B testing metrics by evaluating candidates against the STEDII framework, ensuring measures are sensitive, timely, efficient, debuggable, interpretable, and isolated. This process aligns selected metrics with your PRD success criteria and segmentation plans.

The STEDII framework is a methodology for experiment metric selection that verifies metrics meet six criteria: Sensitivity, Timeliness, Efficiency, Debuggability, Interpretability, and Isolation. It ensures decisions are based on reliable, action-oriented measures.

Define guardrail metrics and estimate sample size by applying the STEDII framework to your planning data. This helps select a primary metric alongside 3 to 5 guardrails, and calculates required sample sizes for desired statistical power and detectable effects.

Yes, you can align experiment metrics with PRD success criteria and a statistical plan. The framework ensures compatibility between selected metrics and your segmentation plans, guardrails, and decision criteria for robust pre-experiment planning.

Without a structured framework, experiment metrics risk being insensitive, delayed, or uninterpretable. Applying the STEDII criteria prevents these issues by enforcing isolation and debuggability, ensuring your decisions are based on valid, action-oriented measures.