Design experiment for fake accounts impact

Your team will remove detected fake accounts and wants to estimate causal impact on real users' experience. Design an end-to-end experiment plan that addresses network interference and product trade-offs. Be specific: 1) Experiment unit and randomization: Choose and justify between user-level, ego-network cluster, or geography-level randomization. Describe how you would construct clusters to minimize cross-treatment contamination while maintaining power. 2) Primary and guardrail metrics: Specify exact definitions (e.g., comments_per_view, 7-day retention of real accounts, abuse reports per 1K views). Define metric windows and whether they are exposure- or calendar-based. 3) Power and duration: Provide a concrete back-of-envelope sample-size calculation assuming a 0.5% relative change in comments_per_view, baseline 0.12, overdispersion, and an intra-cluster correlation of 0.02. 4) Interference diagnostics: Propose two separate tests to quantify spillovers (e.g., ghost exposure analysis for users connected to treated removals; edge-cut A/A). Define the expected null. 5) Noncompliance and misclassification: Detection is imperfect. Outline an IV or CUPED/DID approach to recover LATE using removal intensity as an instrument; list assumptions and falsification checks. 6) Ramp and ethics: Define a staged rollout with kill-switch criteria using guardrails (e.g., creator reach drop >1% with p<0.05). Include how you will prevent label leakage in feeds and notifications. 7) Beyond experimentation: If randomization is infeasible in some markets, provide a quasi-experimental backup (synthetic control or staggered DID) and specify exact covariates required from logs. Conclude with a product recommendation you might make if engagement dips short-term but abuse reports drop materially.