You are planning an A/B test for a change to an ecommerce checkout flow and need to decide whether the experiment is worth running. The team wants to understand how many users you need, what effect size is realistic to detect, and how those choices affect the timeline and confidence in the result.
How do you think about sample size and minimum detectable effect when planning this experiment? Walk through how you would choose the MDE, translate it into a required sample size, and balance that against practical constraints and guardrails.