What problem does it solve? It turns raw Google Analytics data into one concrete, testable website copy improvement per week, removing the guesswork from deciding which words on a site to change and how to measure whether the change worked. ## Core Features & Use Cases - Analytics-driven problem finding: Pulls 28 days of GA4 data, filters bot traffic, and identifies copy-shaped problems such as weak CTAs, low-engagement landing pages, or mismatched headlines. - Falsifiable hypotheses: Forces every change to be stated as a measurable hypothesis naming its metric before any edit is made, and supports no-op runs when the data does not justify a change. - Safe implementation workflow: Makes copy edits on a separate git branch in a temporary worktree, verifies the Next.js build, and never pushes, merges, or deploys without user approval. - Use Case: A designer runs the loop every Monday; it notices a high-traffic case study with weak dwell time, rewrites its title and description on a branch, and delivers a plain-English report for approval. ## Quick Start Ask the assistant to run the growth loop to analyze last month's GA data and propose one copy change on a review branch.
