Baseline conversion is 5%. You want 90% power at α=0.05 (two-sided) to detect a 6% relative lift. Compute per-variant sample size for a Bernoulli outcome using a normal approximation, then adjust for 10% bot traffic and a 7-day window assuming independence by day. After running: A has 50,000 users and 2,650 conversions; B has 49,500 users and 2,820 conversions, with one interim look at day 3. Calculate the p-value and 95% CI for the difference in proportions, correct for the interim look (e.g., O’Brien–Fleming or Bonferroni), check for sample-ratio mismatch, and conclude whether to ship while discussing risks of Type S/M errors and baseline mis-specification.