GrowthBook

Rise standardized experiment statistics by running tests in GrowthBook on its warehouse. Rise uses Segment for tracking, Amplitude for product analytics and GrowthBook, which sits on the data warehouse, for test statistics. Leon Sasson called GrowthBook the biggest move toward standardizing experimentation at Rise.

Checks like sample ratio mismatch come built in. It is cheap to adopt if the data is already organized. Leon's advice: "never write your own statistics engine".

Subversive▶ 28:28

Define metrics once and keep product managers out of spreadsheet winner calls. Judging winners in spreadsheets or with a homegrown stats engine invites pitfalls such as sample ratio mismatch. A shared statistics engine with standard metric definitions removes that risk.

Metrics like trial start rate need clean definitions, because real user journeys are non-linear, with reinstalls and returns.

Subversive▶ 28:28