Skip to content
prod e051e98
Browse

Growth & scale

Objective — Measure who uses what, fix onboarding drop-off, experiment safely, and scale infra before traffic spikes hurt you.

LayerQuestion it answersTypical tool
Error monitoringIs the app broken?Sentry
Web analyticsWho visits and from where?GA4 / Rybbit
Product analyticsWhich features drive retention?PostHog, Mixpanel, or GA4 events
ExperimentsDid the change help?A/B framework or feature flags
graph LR
Traffic["Traffic"] --> Activate["Activation<br/>(onboarding)"]
Activate --> Retain["Retention<br/>(feature use)"]
Retain --> Revenue["Revenue"]
Revenue --> Scale["Infra scale"]
  • Define north-star and 3–5 key events (signup, trial start, core action, upgrade).
  • Instrument server- and client-side events consistently (user_id in staging only when safe).
  • Build one dashboard: funnel from landing → activation → paid.

Enrich detail: see Monitoring & analytics and setup-new phases for Sentry/GA wiring.

  • Map first-session steps; measure drop-off per step.
  • Remove optional fields; defer advanced setup to later sessions.
  • Add in-app checklist or progress bar tied to activation event.
  • Hypothesis → single change → pre-defined success metric → minimum sample window.
  • Never run overlapping tests on the same page without segmentation.
  • Ship winner; document in changelog.
  • Watch CPU, queue depth, and DB connections after marketing pushes.
  • Enable opcode cache, horizon/queue workers, and CDN for static assets before vertical scaling.
  • Load-test staging with 2× expected peak.
  • Core events firing in analytics dashboard
  • Onboarding funnel baseline captured
  • One experiment doc with metric + stop rule
  • Staging load test recorded