Whenever you're ready, walk me through how you'd approach this.
How would you design an A/B test to evaluate a new onboarding flow for small businesses versus enterprise organizations on Slack? Explain how you would define the hypothesis, choose the randomization unit, and compare results for the two organization types. Specify the primary metric, guardrails, traffic and duration assumptions, an explicit MDE and power calculation, the analysis plan, and a ship or do-not-ship rule. Address peeking, sample ratio mismatch, novelty effects, network effects, SUTVA, and differences between the unit of randomization and unit of analysis.