Structured framework for designing statistically rigorous A/B tests with mandatory validation gates.

• Enforces hypothesis lock, metric freezing, and sample size calculation before any implementation begins
• Includes hard gates at three critical points: hypothesis validation, assumptions review, and execution readiness
• Defines primary, secondary, and guardrail metrics to prevent harmful wins and ensure valid interpretation
• Covers test type selection, duration estimation, and strict rules against mid-test changes or early stopping
• Provides refusal conditions and documentation requirements to build institutional learning and avoid repeated test failures