Design an A/B test with guardrails

Your team will test a new checkout flow expected to improve conversion by 3% relative (baseline 5%). (a) Choose a primary metric and at least two guardrail metrics; justify. (b) Compute the required sample size per group for 90% power at α=0.05, accounting for a 10% bot-traffic exclusion and a 5% day-of-week effect; specify all assumptions. (c) Propose a randomization and bucketing scheme that prevents contamination across sessions and devices. (d) Define a pre-registration plan, including a fixed-horizon vs sequential design and how you will handle peeking (e.g., group sequential or alpha spending). (e) Describe how you will segment results (e.g., new vs returning) while controlling multiplicity.