Facebook Groups Product Health and Feature Experiment Design

Context

You are evaluating the current health of Facebook Groups and deciding whether to launch a proposed "comment collapsing" feature. Assume the feature automatically collapses some comments within Group posts (e.g., low-ranked, off-topic, or long subthreads), with an affordance to expand. The goal is to reduce clutter and help people find valuable comments more efficiently without harming group engagement or safety.

Tasks

1. Define the core (north-star) and guardrail metrics you would use to judge the overall health of Facebook Groups today.
2. Design an experiment to evaluate whether the comment-collapsing feature should be launched. Include:
   • Hypotheses and success criteria
   • Experiment unit and randomization
   • Key metrics (primary, secondary, guardrails)
   • Duration, sample size/power approach
   • Analysis plan and heterogeneity cuts
   • Risks, trade-offs, and mitigations

Hints

• Be explicit about what the north-star captures and what guardrails protect.
• Outline how you would measure quality, satisfaction, and safety.
• Consider network effects, interference, and bias in your design.
• Propose practical thresholds for launch vs. iterate vs. do-not-launch.