← Meta Interview Insights

Meta·Product Manager·Onsite - Product Sense / Strategy·Senior

Senior
May 2026

Summary

Meta PM interview, one case question about Facebook Reactions where you're handed a mixed-signal metrics scenario and have to figure out what to do. Pretty classic product analytics setup but the specifics made it trickier than expected.

Questions Asked (1)

Q1

You're the PM for Facebook Reactions. Reactions are up 20% but comments are down 10%. What do you do?

Product Analytics & MetricsRoot Cause AnalysisProduct Sense & Ideation
Author's notes

My first instinct was to say reactions are cannibalizing comments and just run with that.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the metric definitions and the time period, then hypothesize why reactions might be up while comments are down, considering both substitution and causal effects. Prioritize hypotheses by potential impact and ease of testing, and propose a plan to validate the root cause before jumping to solutions.

Pro tip: Don't assume the metrics are directly related; consider external factors like algorithm changes or UI updates that could independently affect both. Also, segment the data by user type, content type, and geography to uncover nuanced insights.

1. Clarify and Validate Metrics

Define what 'reactions up 20%' and 'comments down 10%' mean precisely: absolute vs relative, per user vs total, time frame, and whether these are statistically significant. Confirm the data source and any recent changes.

2. Segment and Explore Data

Break down the metrics by user demographics, content categories, device types, and geography to see if the trend is uniform or driven by a specific segment. Look for correlations with other metrics like time spent, shares, or reports.

3. Generate Hypotheses

Brainstorm possible reasons: reactions may be substituting for comments (e.g., easier to react), a UI change may have made commenting harder, algorithm changes may favor reaction-worthy content, or external events may drive more reactions but fewer comments.

4. Prioritize and Test Hypotheses

Rank hypotheses by impact and likelihood, then design quick tests or analyses (e.g., A/B tests, cohort analysis, user surveys) to validate or invalidate them. Start with the most plausible and easily testable.

5. Decide on Action and Monitor

Based on findings, decide whether to intervene (e.g., adjust UI, algorithm, or incentives) or accept the change if it aligns with goals. Define success metrics and monitor long-term effects on user engagement and well-being.

Key Points to Mention

  • Distinguish between correlation and causation; reactions and comments may not be directly linked.
  • Consider the trade-off between lightweight engagement (reactions) and deeper engagement (comments) and its impact on community health.
  • Check for recent product changes (e.g., UI redesign, algorithm updates) that could explain the shift.
  • Segment by user cohorts (e.g., power users vs. casual users) to see if the trend is driven by a specific group.
  • Evaluate whether the change is net positive or negative for the platform's goals (e.g., meaningful interactions).
  • Propose a data-driven approach: use A/B testing or holdout groups to isolate the cause.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.