User Analytics and Growth Metrics
Track the metrics that matter. Set up analytics, build an internal dashboard, and make data-driven decisions about your product.
The metrics that matter for a new SaaS
Ignore vanity metrics (page views, total signups). Focus on: activation rate (what percentage of signups complete the core action within 24 hours), retention rate (what percentage return in week 2, week 4), conversion rate (what percentage upgrade to paid), churn rate (what percentage of paid users cancel per month), and monthly recurring revenue (MRR). Ask Claude Code: "Set up PostHog analytics. Create custom events for: signup, activation (define this as creating their first [core object]), feature_used (track which features are used and how often), upgrade, and churn. Add the PostHog client to the app and instrument these events at the appropriate points in the code."
Building an internal metrics dashboard
Ask Claude Code: "Create an admin dashboard at /admin/metrics showing: signups per day/week/month (line chart), activation funnel (signup to first action to second session), retention cohorts (weekly cohort table), MRR chart, and a real-time feed of user activity. Use the database directly for these queries — do not rely on PostHog's API for the dashboard. Cache the queries for performance."
Setting up alerts and automated reports
Ask Claude Code: "Create a weekly metrics email that sends every Monday at 9am to the admin. Include: signups this week vs last week, activation rate, MRR change, notable events (first paying customer, churn events), and the top 3 users by activity. Use Resend to send it. Also create a Slack webhook alert when: a new user upgrades to paid, MRR exceeds a milestone, or churn rate exceeds 5 percent."
Making decisions from data
Data without decisions is useless. Here is the framework: check activation rate first — if people sign up but never use the product, fix onboarding before anything else. Check retention second — if people use it once but never return, the product is not sticky enough. Check conversion third — if people use it regularly but do not pay, the pricing or upgrade prompt needs work. Ask Claude Code to add an experiment framework: "Add a simple A/B testing system that randomly assigns users to variants, tracks which variant they saw, and measures conversion for each variant. Support testing on: onboarding flow, pricing page, and upgrade prompts." Start with one experiment at a time. Run it for at least 2 weeks before declaring a winner.
Data-Driven Decision Making
This guide is hands-on and practical. The full curriculum covers the conceptual foundations in depth with structured lessons and quizzes.
Go to lesson