Product analytics
Answers from every event, on one model
Trends, funnels, retention, paths, stickiness and lifecycle over the same events, with SQL when you need to go further. Charts show you the query behind them.
The insights you expect
Every insight compiles to parameterised SQL that you can read, and runs next to your data in its region.
- Trends with total, unique people, sessions or groups, sums, averages and percentiles; formulas and period comparison
- Funnels of up to 20 steps: ordered, strict or unordered, with conversion windows and time to convert
- Retention, first-time or recurring; paths; stickiness; lifecycle
- Breakdowns by event, person or group property; cohorts as filters
- Dashboards, annotations, and a Home dashboard for the numbers you check every morning
SQL that cannot leak
Write ClickHouse SQL over virtual tables for events, people, sessions, groups, cohorts and revenue. Queries are compiled against an allowlist, forced into your project, and checked again by row policies in the database.
A tracking plan that knows what is firing
The Catalog records the question each event answers and flips planned events to implemented when they first arrive. atrix codegen turns the plan into typed wrappers for TypeScript, Swift, Kotlin, Python and Go.
Identity you can audit and undo
Anonymous traffic stays personless until sign-in. Merges follow deterministic rules, two identified users are never joined by accident, and any merge can be reverted from the person's identity graph.
Go deeper: Getting started · SQL · Identity