Setting Up Analytics That Actually Help

The Analytics Problem

Most developers fall into one of two traps. They either track nothing and make decisions based on gut feeling, or they track everything and drown in data they never look at. Both approaches waste your time.

Good analytics means tracking the right things, reviewing them regularly, and making decisions based on what you find. This guide covers exactly what to track, how to set it up, and how to turn data into action.

Choosing an Analytics Tool

For mobile apps, the main options are:

PostHog

Open source, self-hostable, or cloud-hosted
Event tracking, session replays, feature flags, A/B testing
Generous free tier (1 million events/month)
Good React Native SDK
Best for: indie developers who want one tool that does everything

Mixpanel

Event-based analytics with powerful segmentation
Funnel analysis and retention reports
Free up to 20 million events/month
Best for: apps that need detailed funnel analysis

Amplitude

Similar to Mixpanel with strong behavioral analytics
Free tier up to 10 million events/month
Best for: product teams focused on user behavior

Firebase Analytics (Google Analytics for Firebase)

Free, no event limits
Integrated with other Firebase services
Limited customization compared to Mixpanel/Amplitude
Best for: apps already using Firebase

Our Recommendation

For most indie developers, PostHog is the best starting point. One tool for events, funnels, retention, and feature flags. The free tier covers most indie apps indefinitely.

Setting Up PostHog in React Native

npm install posthog-react-native

// lib/posthog.ts
import PostHog from 'posthog-react-native';

const apiKey = process.env.EXPO_PUBLIC_POSTHOG_API_KEY;

export const posthog = apiKey
  ? new PostHog(apiKey, {
      host: 'https://us.i.posthog.com',
    })
  : null;

export function trackEvent(
  name: string,
  properties?: Record<string, any>
) {
  if (!posthog) return;
  posthog.capture(name, properties);
}

export function identifyUser(
  userId: string,
  traits?: Record<string, any>
) {
  if (!posthog) return;
  posthog.identify(userId, traits);
}

Call identifyUser when the user logs in. Call trackEvent when they do something meaningful.

What to Track: The Essential Events

You do not need to track every tap. Track events that answer these five questions:

1. Are Users Signing Up?

// Track when signup completes
trackEvent('signup_completed', {
  method: 'email', // or 'google', 'apple'
});

// Track when onboarding completes
trackEvent('onboarding_completed', {
  steps_completed: 4,
  time_spent_seconds: 45,
});

2. Are Users Reaching the Core Value?

Every app has a moment where the user first gets value. For a workout app, it is completing a workout. For a budget app, it is seeing their spending breakdown. Track this moment explicitly.

// The "aha moment" for a workout app
trackEvent('first_workout_completed', {
  exercise_count: 5,
  duration_minutes: 22,
});

// For a budget app
trackEvent('first_budget_created', {
  categories: 8,
  monthly_amount: 3500,
});

3. Are Users Coming Back?

Track session starts with context:

trackEvent('app_opened', {
  days_since_last_open: 2,
  session_count: 15,
});

Your analytics tool calculates retention automatically, but tracking session context helps you understand patterns.

4. Are Users Paying?

trackEvent('paywall_viewed', {
  source: 'feature_limit', // or 'settings', 'onboarding'
});

trackEvent('subscription_started', {
  plan: 'annual',
  price: 59.99,
  trial: true,
});

trackEvent('subscription_cancelled', {
  plan: 'monthly',
  reason: 'too_expensive', // if you ask
  tenure_months: 3,
});

5. Are Users Having Problems?

trackEvent('error_occurred', {
  screen: 'dashboard',
  error_type: 'network',
  message: 'Failed to load data',
});

trackEvent('feature_failed', {
  feature: 'export_pdf',
  reason: 'file_too_large',
});

What NOT to Track

Every Screen View

Screen views create noise. If you need to know which screens are popular, track navigation events selectively for the screens that matter, not every single transition.

Every Button Tap

Tracking "user tapped button X" for every button in your app generates millions of events and tells you very little. Track actions that represent meaningful user intent, not UI interactions.

Personally Identifiable Information

Never track names, email addresses, phone numbers, or other PII in analytics events. Besides being a privacy concern, it often violates GDPR and app store policies.

Vanity Metrics

Total downloads, total signups, and total page views look impressive but tell you nothing about whether your app is healthy. Focus on active users, retention, and revenue.

The Metrics That Matter

Retention Rate

The single most important metric. If users do not come back, nothing else matters.

Day 1 retention: What percentage of users open the app the day after installing? Benchmark: 25-40% is average, 40%+ is good.
Day 7 retention: What percentage are still active after a week? Benchmark: 10-20% is average, 20%+ is good.
Day 30 retention: What percentage are still active after a month? Benchmark: 5-10% is average, 10%+ is good.

If your day 1 retention is below 20%, your onboarding or first experience needs work.

Activation Rate

What percentage of new users reach the "aha moment"? If only 30% of signups complete their first core action, you need to remove friction from the path between signup and value.

Conversion Rate

What percentage of free users become paying users? Track this as a funnel:

Download → Signup → Activation → Paywall View → Trial Start → Paid Conversion

Each step reveals where users drop off.

Revenue Metrics

MRR (Monthly Recurring Revenue): Total subscription revenue per month
ARPU (Average Revenue Per User): MRR divided by active users
LTV (Lifetime Value): Average total revenue per user over their lifetime
Churn rate: Percentage of subscribers who cancel per month

The One Metric That Matters (OMTM)

At any given time, focus on one primary metric. Early stage: focus on retention. Growth stage: focus on activation rate. Monetization stage: focus on conversion rate. Trying to optimize everything at once means optimizing nothing.

Building Dashboards

Create a simple dashboard you review weekly:

Weekly Review Dashboard

This Week:
- New signups: ___
- Day 1 retention: ___%
- Weekly active users: ___
- Paywall views: ___
- New subscribers: ___
- MRR: $___
- Top error: ___

Trends (vs. last week):
- Signups: ↑/↓ __%
- Retention: ↑/↓ __%
- Revenue: ↑/↓ __%

Setting Up Alerts

Configure alerts for critical events:

Crash rate exceeds 1%
Day 1 retention drops below 25%
Error rate on key features spikes
Revenue drops more than 20% week-over-week

You should not need to check your dashboard to know something is wrong. Alerts bring problems to you.

Funnels: Finding Where Users Drop Off

A funnel shows the percentage of users who complete each step in a sequence. Build funnels for your most critical flows:

Onboarding Funnel

App Install → Open App → Start Onboarding → Complete Step 1 →
Complete Step 2 → Complete Step 3 → Create Account → Enter App

If 80% of users complete step 1 but only 30% complete step 2, step 2 is your bottleneck. Investigate: Is it confusing? Too long? Asking for too much information?

Conversion Funnel

Free User → View Paywall → Start Trial → Use Premium Feature →
Trial Expires → Convert to Paid

Each drop-off point tells you something:

Low paywall views: Users do not know premium exists, or the free tier is too generous
Low trial starts: The paywall is not compelling, or the price is too high
Low conversion after trial: The premium features are not valuable enough

Cohort Analysis

Cohort analysis groups users by when they signed up and tracks their behavior over time. This reveals whether your product is improving.

If users who signed up in March have better retention than users who signed up in January, your product changes are working. If retention is declining across cohorts, something is getting worse.

Most analytics tools provide cohort analysis out of the box. Review monthly.

A/B Testing

Once you have baseline metrics, test changes:

Onboarding flow: Test a 3-step vs. 5-step onboarding
Paywall design: Test different layouts, copy, and pricing presentations
Feature placement: Test moving a key feature to the home screen

A/B Testing Rules

Test one change at a time
Run tests for at least 1-2 weeks
Need at least 100-200 users per variant for meaningful results
Define your success metric before starting the test
Do not peek at results early (let the test complete)

PostHog and most analytics tools include built-in A/B testing with feature flags.

Privacy and Compliance

GDPR

If you have users in Europe:

Disclose what you track in your privacy policy
Allow users to opt out of analytics
Do not track PII in events
Respond to data deletion requests

App Tracking Transparency (iOS)

Since iOS 14.5, you must ask permission before tracking users across apps and websites. Analytics within your own app do not require ATT permission, but ad tracking does.

Best Practices

Use anonymous IDs by default
Only identify users after they create an account
Provide a "Do not track" option in settings
Document your data practices in your privacy policy

The Analytics Workflow

Week 1: Set up analytics, define core events, build your first dashboard
Weekly: Review dashboard, check retention, note anomalies
Monthly: Run cohort analysis, review funnels, identify optimization opportunities
Quarterly: Review overall trends, set new metric targets, adjust tracking

Analytics is not a one-time setup. It is a habit. The teams that review their data consistently build better products than the teams that install analytics and forget about it.

Getting Started Today

Start with these five events:

signup_completed
core_action_completed (your app's key value moment)
paywall_viewed
subscription_started
error_occurred

These five events answer the fundamental questions: Are people signing up? Getting value? Considering paying? Actually paying? Having problems? Everything else can wait until you have mastered these basics.