Back to problems

Write SQL for post-trial conversion cohorts

SQL · OpenAI · Hard

You are given four PostgreSQL tables captured from an onboarding experiment run during June 2025. Build the three analyses described below and return one SQL query for each part. Schema All timestamps are UTC. A user's cohort entry point is their earliest row in exposures (call its timestamp first_exposed_at and its label first_variant); every later exposure row for that user is ignored, so a person is attributed to exactly one arm. If two rows tie for the earliest…

Checking your access…