Big Tech platforms Product Analyst Interview Loop
Product analyst interviews at the big platforms test product sense, and SQL is the gate. Meta's version is the one to know. There's a recruiter call, then a technical screen that opens with a product case and then asks for the SQL that would prove it. After that come four to five interviews: two product-improvement rounds, an analytical round, and execution and behavioral. Google's loop follows its data analyst pattern with a heavier case. Amazon hires for this work as Business Analyst, with Leadership Principles in every round. The candidates who get caught out expected a SQL-heavy screen.
Practice by round
SQL basics on product tables like users, sessions and purchases, written 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 every platform asks. How would you measure this? What could confound it? Which of two experiment results would you trust?
These are the product-improvement interviews, and they're the heart of the loop. You take a product and go from a goal to a success metric to the first experiment you'd run. The questions include defining a product success metric, launch or no-launch, a north star for e-commerce, notification and feed-health metrics, and a user-satisfaction metric. Interviewers push on whether your metrics follow from your reasoning. Keep your structure and walk them through each step.
| Case study | Difficulty |
|---|---|
| Graph of Sales and Revenue | Easy |
| New Notification System | Medium |
| Clicks on Search Result | Medium |
| Teenager's Parents on Facebook | Medium |
| North Star Metric for E-Commerce | Medium |
| Measuring Social Media Campaign Success | Medium |
Experiments at the analyst level: setting up an A/B test, what sample size depends on, reading a conversion rate, and what network effects do to a test. You choose and read experiments here. You don't compute them.
| Question | Difficulty |
|---|---|
| Z and t distributions | Easy |
| Outliers Detection | Easy |
| Assessing Multicollinearity | Easy |
| Right vs Left-Skewed Distributions | Medium |
| Linear Regression & T-test | Medium |