Practice real Meta analytics interview cases
Real business problems from Meta's product analytics interviews. Real datasets. Write SQL, investigate, and present your findings — then get feedback on your analytical thinking.
The Friend Request Freeze
Scenario
“Hey — friend requests sent across the platform dropped about 10% week-over-week. This is one of our core engagement metr...”
Data Schema
| week | total |
|---|---|
| 2025-01-06 | 248,192 |
| 2025-01-13 | 251,847 |
| 2025-01-20 | 246,503 |
| 2025-01-27 | 228,614 |
| 2025-02-03 | 219,847 |
| 2025-02-10 | 211,203 |
Meta Analytics Cases
These are the problems Meta's analytics interviewers give candidates. Solve them here first.
The Friend Request Freeze
Hey — friend requests sent across the platform dropped about 10% week-over-week. This is one of our core engagement metrics and leadership is asking questions. Can you figure out what's going on?
sql · investigation · segmentation · root_cause_analysis
Start this caseThe Threading Tradeoff
So threading is getting mixed signals. Comments per user are up about 10% — that's great. But posts created is down about 2% — that's not great. Head of Product wants a clear answer before the all-han
sql · product_analytics · segmentation · tradeoff_analysis
Start this caseThe Like Counter Removal
Was this a good decision? Should we keep it, roll it back, or modify it?
sql · product_analytics · segmentation · communication
Start this caseThe Close Friends Experiment
feature that lets users share Stories visible only to a curated list of friends. The test splits users 50/50 — treatment gets the feature, control doesn't. The test ran for 6 weeks. The PM says resu
sql · experimentation · segmentation · communication
Start this caseThe Green Dot Dilemma
dot next to users' profile photos in the messaging interface. The feature was rolled out to ALL users simultaneously — no A/B test was run because the engineering team said the real-time presence inf
sql · causal_inference · quasi_experiment · communication
Start this caseReels vs. TikTok: Closing the Engagement Gap
Instagram launched Reels 18 months ago to compete with TikTok's dominance in short-form video. Reels have grown to represent 30% of total time spent on the app, but growth has plat
sql · growth_analysis · segmentation · platform_dynamics
Start this caseThe Ad Load Dilemma
A photo/video social platform currently shows ads at a rate of 1 ad per 6 content posts in the feed (16.7% ad density). The monetization team wants to increase ad load to improve r
sql · experiment_analysis · revenue_optimization · segmentation
Start this caseThe Algorithm That Broke Event Search
A social platform's Events feature shows that the click-through rate on event search results increased from 18% to 19.8% (a 10% relative increase) week-over-week. The Events PM is
sql · investigation · metric_interpretation · root_cause_analysis
Start this caseThe Reels Recession
We're seeing a significant drop in the core Reels engagement metric. Average Reels watched per user has declined 18% in the past two weeks (from 23.4 to 19.2 per day). This is alarming — Reels is our
Metric interpretation · Count vs. quality decomposition · Multi-cause investigation · Product changelog analysis
Start this caseThe Groups Paradox
feature designed to surface relevant Groups in the News Feed. Here's what the metrics show: - Daily Active Users of Groups (DAU): +2.0% (from 18.4M to 18.8M) - Average time spent on Groups per activ
A/B test evaluation with conflicting metrics · Mix-shift / composition analysis · Cohort decomposition · Top-of-funnel vs. engagement tradeoff
Start this caseThe Weekend Seller Illusion
This is a huge opportunity! We should run a promotion where we pay sellers a $5 bonus for every item they list on Friday-Sunday. If we can shift selling to weekends, we'll boost GMV and take a cut. Wh
Selection bias identification · Confounding variable decomposition · Causal reasoning · Cost-benefit analysis
Start this case