Meta Product Analyst Interview Loop
Most Meta roles called Product Analyst are really Product Growth Analyst roles, and the Meta product analyst interview is built for that job. There's a recruiter call, then a 45-minute screen that opens with a product case and then asks for the SQL behind it. After that come four to five interviews: two product-improvement rounds, an analytical round, and execution and behavioral. The candidates who get caught out are the ones who prepared for a SQL-heavy screen. The case comes first.
What decides it
Product judgment, expressed as metrics. SQL gets you in the door. Round 3 is the interview.
Practice by round
SQL basics on product tables, done fast so the conversation can get back to the product. Practice the query as the second half of an answer. Name the metric first, then compute it.
Engagement, retention and funnel SQL at medium to hard, with cohort logic and window functions. Then come the follow-ups Meta always asks. How would you measure this? What could confound it? Which of two experiment results would you trust?
| Question | Difficulty |
|---|---|
| Blocked Users | Medium |
| Number of Conversations | Medium |
| Premium Accounts | Medium |
| Customers Who Purchased the Same Product | Medium |
| Daily Interactions By Users Count | Medium |
| Customer Revenue In March | Medium |
| User Feature Completion | Medium |
| User Streaks | Hard |
| Common Interests Amongst Users | Hard |
| Time Between Two Events | Hard |
| Average Time Between Steps | Hard |
These are the two product-improvement interviews, and they're the heart of the loop. You go from a product goal to a success metric and the experiment you'd run first. The questions include defining a product success metric, launch or no-launch, feed health metrics, notification success metrics, a feed-ranking impact experiment and People You May Know. Interviewers push on whether your metrics follow from your reasoning. They may even tell you a plan you built from first principles came out of nowhere. Keep your structure and walk them through each step.
Experiments at the analyst's level. That means setting up an A/B test, what sample size depends on, reading a conversion rate, and what network effects do to a test. Here you choose and read experiments. You don't compute them.
| Question | Difficulty |
|---|---|
| Conversion Rate | Easy |
| Negative R Squared | Easy |
| Expectation of Variance | Medium |
| A/B Testing | Medium |
| Propensity Matching for an Opt-In Feature | Medium |