Catalog
8 branches of analytics, product and AI interview questions
Explore eight skill tracks, called branches, filterable by subtopic, company, tool and difficulty. SQL questions run against sample tables, choice and numeric answers are checked instantly, and written cases are scored against evaluation criteria.
Filters
157 questions · page 5 of 7
Measure the health of a Facebook Group
You own analytics for Facebook Groups.
Written case · Metrics · 18 min · Medium
Product Analytics & Metrics · Reported at Meta
Improve Google Maps for daily commuters
ProHow would you improve Google Maps for daily commuters?
Written case · Product sense · 20 min · Medium
Product Management · Reported at Google
Measure surge pricing's effect on driver retention
ProUber is considering a surge-pricing policy that changes marketplace conditions for every driver and rider in a zone.
Written case · Design · 25 min · Hard
Experimentation & A/B Testing · Reported at Uber
Explain a p-value without misleading
An engineer asks: 'The experiment's p-value is 0.03.
Written case · Writeup · 12 min · Medium
Analytics Case & Communication · Reported at Amazon
Longest monthly activity streak
For each user, return the length of their longest streak of consecutive calendar months with at least one activity.
Live SQL · Windows · PostgreSQL · 18 min · Medium
SQL & Data Modeling · Original Uplift question
Design an AI travel-itinerary assistant
As the product manager at Booking.com, design a generative-AI assistant that creates personalized travel itineraries using destination, activity, and accommodation preferences.
Written case · Product metrics · 20 min · Medium
AI Agents & Automation · Reported at Booking.com
Is 550 heads in 1,000 flips suspicious?
A coin is flipped 1,000 times and lands heads 550 times.
Calculation · Intervals · 10 min · Medium
Statistics & Inference · Reported at D. E. Shaw
Choose one agent or a multi-agent system
ProYou are designing a research workflow that must search sources, extract evidence, draft a report, and verify citations.
Written case · Evaluation · 20 min · Hard
AI Agents & Automation · Reported interview question
Sessionize events with a 30-minute gap
ProCreate sessions separately for each user.
Live SQL · Funnels · PostgreSQL · 25 min · Hard
SQL & Data Modeling · Original Uplift question
Expected minimum of two uniforms
ProLet X and Y be independent Uniform(0,1) random variables.
Calculation · Probability puzzles · 15 min · Hard
Statistics & Inference · Reported at Google
Measure an email campaign's success
Airbnb sends a destination-inspiration email to eligible travelers.
Written case · Campaigns · 16 min · Medium
Digital & Marketing Analytics · Reported at Airbnb
Convince stakeholders to test an Eats promotion
ProUber Eats stakeholders want to judge a new discount by comparing customers who redeem it with customers who do not.
Written case · Structure · 22 min · Hard
Analytics Case & Communication · Reported at Uber
Evaluate engagement of new users
A consumer app acquired 100,000 new users last month.
Written case · Retention · 16 min · Medium
Product Analytics & Metrics · Reported interview question
Search relevance gains but pages slow down
ProA new Amazon search feature improves an offline relevance metric by 10% but adds two seconds to result-page load time.
Written case · Prioritization · 20 min · Hard
Product Management · Reported at Amazon
Triage a sample-ratio mismatch
ProA nominal 50/50 Meta messaging experiment reports 520,000 control users and 480,000 treatment users.
Written case · Readout · 20 min · Hard
Experimentation & A/B Testing · Reported at Meta
Allocate budget with conflicting attribution models
ProA customer sees a paid-social ad, later searches the brand, clicks a paid-search ad, and buys.
Written case · Attribution · 22 min · Hard
Digital & Marketing Analytics · Reported interview question
High bounce and short sessions on Meta mobile
ProMeta's mobile app suddenly shows higher bounce rate and lower session duration.
Written case · Diagnosis · 22 min · Hard
Product Analytics & Metrics · Reported at Meta
Inference with severely skewed revenue
Customer revenue is zero for many users and has a very long right tail.
Written case · Regression · 15 min · Medium
Statistics & Inference · Reported interview question
Week-one retention by signup cohort
Group users by Monday-starting signup week.
Live SQL · Funnels · PostgreSQL · 20 min · Medium
SQL & Data Modeling · Original Uplift question
Defend a tool-using agent from prompt injection
ProAn email assistant can read messages, search internal documents, draft replies, and create calendar events.
Written case · Evaluation · 22 min · Hard
AI Agents & Automation · Reported interview question
Write the decision from a failed experiment
ProA six-week onboarding experiment estimated +0.3 percentage points on activation with a 95% confidence interval of -0.4 to +1.0 points.
Written case · Writeup · 15 min · Medium
Analytics Case & Communication · Reported interview question
Launch Uber in a new city
ProYou are responsible for launching Uber rides in a new city.
Written case · Product sense · 25 min · Hard
Product Management · Reported at Uber
Test a new shoppable-ad ranking model
Meta has a new ranking model for shoppable-feed ads.
Written case · Design · 20 min · Medium
Experimentation & A/B Testing · Reported at Meta
Surface promotions that need attention
ProTarget runs hundreds of promotions across channels.
Written case · Campaigns · 18 min · Medium
Digital & Marketing Analytics · Reported at Target