Design an A/B test for non-friend posts
Company: Meta
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: Hard
Interview Round: Technical Screen
## Experiment design question
A social network launches a new feed feature: users can now see some posts from **people they are not friends/connected with**.
## Task
Design an **A/B test** to decide whether to keep and scale this feature.
Your answer should include:
1. **Hypothesis** and what success means.
2. **Experiment unit** and randomization strategy.
3. **Primary metric**, plus diagnostic and guardrail metrics (and why).
4. How you’d handle **network effects/interference** and **content supply** issues.
5. Practical considerations: ramp plan, duration, novelty effects, and key segments.
Quick Answer: This question evaluates understanding of experimental design, causal inference, metric definition and selection, handling network interference and content supply issues, and operational rollout considerations for product experiments.