The A/B Testing skill transforms Claude into a specialized experimentation strategist for e-commerce brands. It moves beyond simple split-testing by enforcing a strict hypothesis-first approach and providing deep statistical rigor for sample size and Minimum Detectable Effect (MDE) calculations. Whether you are designing a new experiment or analyzing completed results, this skill ensures test isolation, prevents peeking-driven false positives, and helps operators decide when a test is worth running—or when it's better to pivot to higher-leverage changes based on traffic limitations.
主要功能
01Automated MDE and required sample size calculations for significance
02Sequential testing guidance to manage daily monitoring and stop-rules
033 GitHub stars
04Feasibility assessment based on monthly traffic and conversion baselines
05Rigorous X-Y-Z-mechanism hypothesis formulation and validation
06Test isolation frameworks to prevent cross-contamination on complex pages
使用场景
01Designing an isolated experiment to test if new H1 copy improves conversion rates
02Evaluating a completed test to determine if the results justify the operational cost of shipping
03Calculating if a proposed PDP change can reach statistical significance within 30 days