Mobile Analytics Implementation Plan Template

Start with 15-25 core events that answer your key product questions. You can always add more, but removing events is harder because historical data gaps cannot be backfilled. Focus on lifecycle events (install, onboard, activate, retain), core feature interactions, and [conversion](/glossary/conversion-rate) milestones. Avoid tracking every tap. If you find yourself with 100+ events, your taxonomy probably needs restructuring.

Implement analytics before your first beta release, not after. If you wait until GA to add tracking, you will miss the most valuable learning window: the first cohort of real users. Plan the taxonomy during PRD writing, implement during development, and QA during beta. Use the [launch checklist](/templates/app-release-checklist-template) to ensure analytics validation is part of your release process.

Firebase Analytics (Google Analytics for Firebase) is free and tightly integrated with other Firebase services, but its analysis capabilities are limited compared to Amplitude and Mixpanel. Amplitude and Mixpanel offer more advanced funnel, retention, and cohort analysis tools. Amplitude's free tier (10M events/month) is more generous. Choose Firebase if cost is the primary constraint and you are already in the Google ecosystem. Choose Amplitude or Mixpanel if product analytics is a core competency for your team.

ATT only affects cross-app tracking (IDFA). First-party analytics (tracking events within your own app) does not require ATT consent. You can still track all user behavior inside your app without prompting for ATT. ATT matters when you want to attribute installs to ad campaigns or share data with third-party ad networks. Design your analytics plan around first-party data and you will have full visibility regardless of ATT opt-in rates.

Use snake_case with a verb_noun pattern: `feature_action_object`. This convention is readable in dashboards, sortable alphabetically by feature, and avoids the ambiguity of camelCase (is it `checkoutStarted` or `CheckoutStarted`?). Prefix events with their category when you have 50+ events (e.g., `commerce_checkout_started`, `onboarding_step_completed`). Whatever convention you pick, enforce it with a linter or validation layer. Inconsistent naming is the number one analytics debt pattern. ---