← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Senior

Senior
May 2026

Summary

Meta DS interview focused entirely on a product metrics and experimentation case around push notifications. One big question, lots of moving parts, and I definitely underestimated how deep they wanted to go on the stats side.

Questions Asked (1)

Q1

A mobile app sends push notifications and wants to ensure only high-quality notifications are shipped. Propose and justify a metric for notification quality, explain how you'd use historical CTR data to establish whether current performance is good or bad, and design an A/B test for a new notification strategy including setup, success metrics, and in-depth result analysis.

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

This question ate up the whole session.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining a composite quality metric that balances user engagement and negative feedback, then use historical CTR distributions to set performance benchmarks. For the A/B test, outline a clear hypothesis, randomization unit, and success metrics, and plan for deep analysis including segmentation and novelty effects.

Pro tip: Always consider the trade-off between short-term CTR and long-term user retention; a notification that boosts clicks but increases opt-outs is not high-quality. Also, ensure your A/B test accounts for network effects and interference, especially in social apps like Meta's.

1. Define Notification Quality Metric

Propose a metric that combines positive signals (e.g., click-through rate, conversion rate) and negative signals (e.g., dismissal rate, opt-out rate, report rate) into a single score, such as a weighted sum or ratio. Justify why this captures 'quality' from both user and business perspectives.

2. Establish Historical Benchmarks

Analyze historical CTR data to understand its distribution (e.g., mean, median, percentiles) and trends over time. Use this to set thresholds for what constitutes 'good' or 'bad' performance, considering seasonality and user segments.

3. Design A/B Test for New Strategy

Define the hypothesis, control and treatment groups, randomization unit (e.g., user), and sample size calculation. Specify primary and guardrail metrics, and outline the test duration to capture full user behavior.

4. Analyze Results In-Depth

Beyond average treatment effect, perform segmentation analysis (e.g., by user demographics, past engagement), check for novelty effects, and evaluate long-term impact on retention. Use statistical tests to ensure significance and consider practical significance.

5. Synthesize and Recommend

Combine findings from the metric, historical analysis, and A/B test to make a recommendation on whether to ship the new strategy, iterate, or abandon. Discuss potential trade-offs and next steps.

Key Points to Mention

  • Composite metric design: balance CTR with negative feedback like opt-outs and reports.
  • Use of percentiles and time-series analysis for historical CTR to account for seasonality and trends.
  • A/B test setup: randomization unit, sample size, power analysis, and guardrail metrics.
  • Segmentation analysis to uncover heterogeneous treatment effects.
  • Novelty effect and long-term holdout to measure sustained impact.
  • Statistical significance vs. practical significance and business impact.

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