You work on the onboarding flow for a workforce software product. Your team has built a shorter admin setup experience and believes it will increase the share of newly invited company admins who complete payroll setup in their first session. You plan to run a 50/50 A/B test, but after launch the treatment arm is receiving noticeably fewer eligible admins than expected. Leadership still wants a recommendation on whether the experiment can be trusted and how to proceed.
How would you design and analyze this experiment so that you can detect a meaningful lift while explicitly checking for sample ratio mismatch, and how would SRM affect your interpretation and ship decision if the primary metric looks positive?