Senior Product Analyst
SQL Interview Sandbox
A real SQLite database, pre-loaded with a small A/B test cohort. Write a query and run it. Everything runs in your browser, nothing is saved.
Schema
users
| user_id | integer |
| arm | 'control' or 'variant' |
| signup_date | date |
reservations
| reservation_id | integer |
| user_id | integer |
| created_at | date |
| status | confirmed / cancelled / refunded |
subscriptions
| subscription_id | integer |
| user_id | integer |
| started_at | date |
| canceled_at | date or null |
| plan | monthly / annual |
Questions
- Warm-up. Per arm, number of signups and the share with at least one confirmed reservation.
- Weekly cohort. Weekly signup cohorts per arm, and the share of each cohort that made a first confirmed reservation within 14 days of signup.
- Retention curve. Per arm, users with a confirmed reservation in week N after signup, for N from 0 to 4.
- 30-day conversion. Per arm, the share of signups that converted to an active subscription within 30 days, handling cancellations correctly.
- Sample ratio. Counts per arm. Then: are the arms split 50/50, how would you test it, and what would you do if not?
Cmd / Ctrl + Enter to run
Loading SQLite…