sql-queries

# SQL Queries Turn a product question into trustworthy SQL using a small set of battle-tested product-analytics patterns. Most product questions reduce to one of four shapes — funnel, retention/cohort, active-user ratio, or event count — and getting the shape right matters more than clever syntax. **Grounded in:** *Lean Analytics* — Croll & Yoskovitz: answer the product question (funnels, retention windows, DAU/MAU) — correctly, not approximately. **Go deeper (The Product Channel):** [Product Metrics](https://sidsaladi.substack.com/p/week-5-week-in-product-series-product) ## When to use this - A PM or analyst asks "how many users went from signup to first key action, and where do they drop off?" - You need 1/7/30-day retention or a weekly cohort triangle and want it correct, not approximately correct. - You want DAU, WAU, MAU, or the DAU/MAU stickiness ratio over time. - You need to count events or unique users per feature, segment, or time bucket — and trust the number. - A dashboard figure looks off and you want a query plus a sanity-check protocol to confirm or debunk it. ## Before you start (gather these) - **The question, stated as a metric** — e.g. "7-day retention of users who signed up in May," not "how's retention." - **The schema** — table names, key columns, and grain. At minimum: the events/activity table, its user-id column, its timestamp column, and how an "event type" is identified. - **The SQL dialect** — Postgres, BigQuery, Snowflake, Redshift, MySQL, Duc...