maternal-app/docs/implementation-docs/product-analytics-dashboard.md

Product Analytics Dashboard Guide

Metrics, KPIs, and Data-Driven Decision Making

---

Overview

This document defines the key metrics, analytics dashboards, and monitoring strategies for the Maternal App to enable data-driven product decisions and rapid iteration based on user behavior.

Success Criteria (from Implementation Plan)

MVP Launch (Month 1)

• 1,000 downloads
• 60% day-7 retention
• 4.0+ app store rating
• <2% crash rate
• 5+ activities logged per day per active user
• 70% of users trying AI assistant

3-Month Goals

• 10,000 active users
• 500 premium subscribers
• 50% month-over-month growth
• 4.5+ app store rating

6-Month Vision

• 50,000 active users
• 2,500 premium subscribers
• Break-even on operational costs

---

Key Performance Indicators (KPIs)

1. User Acquisition Metrics

Download & Registration Funnel

Metric                          Target      Formula
─────────────────────────────────────────────────────
App Store Impressions           100,000     Total views
Download Rate                   3%          Downloads / Impressions
Registration Rate               75%         Signups / Downloads
Onboarding Completion           90%         Completed / Started
Time to First Value             < 2 min     First activity logged

Dashboard Queries:

-- Daily registration funnel
SELECT
  DATE(created_at) as date,
  COUNT(*) FILTER (WHERE step = 'download') as downloads,
  COUNT(*) FILTER (WHERE step = 'registration_started') as started_registration,
  COUNT(*) FILTER (WHERE step = 'registration_completed') as completed_registration,
  COUNT(*) FILTER (WHERE step = 'onboarding_completed') as completed_onboarding,
  COUNT(*) FILTER (WHERE step = 'first_activity') as first_activity
FROM user_funnel_events
WHERE created_at >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY DATE(created_at)
ORDER BY date DESC;

-- Conversion rates by channel
SELECT
  acquisition_channel,
  COUNT(*) as total_users,
  AVG(CASE WHEN onboarding_completed THEN 1 ELSE 0 END) as onboarding_completion_rate,
  AVG(time_to_first_activity_minutes) as avg_time_to_value
FROM users
WHERE created_at >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY acquisition_channel;

2. Engagement Metrics

Daily Active Users (DAU) / Monthly Active Users (MAU)

// Analytics service tracking
export interface EngagementMetrics {
  dau: number;               // Users active in last 24h
  wau: number;               // Users active in last 7 days
  mau: number;               // Users active in last 30 days
  dauMauRatio: number;       // Stickiness: DAU/MAU (target: >20%)
  averageSessionDuration: number;  // Minutes (target: >5 min)
  sessionsPerUser: number;   // Per day (target: >2)
}

Dashboard Queries:

-- DAU/WAU/MAU trend
WITH daily_users AS (
  SELECT
    DATE(last_active_at) as date,
    user_id
  FROM user_sessions
  WHERE last_active_at >= CURRENT_DATE - INTERVAL '30 days'
)
SELECT
  date,
  COUNT(DISTINCT user_id) as dau,
  COUNT(DISTINCT user_id) FILTER (
    WHERE date >= CURRENT_DATE - INTERVAL '7 days'
  ) OVER () as wau,
  COUNT(DISTINCT user_id) OVER () as mau,
  ROUND(COUNT(DISTINCT user_id)::numeric /
    NULLIF(COUNT(DISTINCT user_id) OVER (), 0) * 100, 2) as dau_mau_ratio
FROM daily_users
GROUP BY date
ORDER BY date DESC;

-- Power users (top 20% by activity)
SELECT
  user_id,
  COUNT(*) as total_activities,
  COUNT(DISTINCT DATE(created_at)) as active_days,
  AVG(session_duration_seconds) / 60 as avg_session_minutes
FROM activities
WHERE created_at >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY user_id
HAVING COUNT(*) > (
  SELECT PERCENTILE_CONT(0.8) WITHIN GROUP (ORDER BY activity_count)
  FROM (SELECT COUNT(*) as activity_count FROM activities GROUP BY user_id) counts
)
ORDER BY total_activities DESC;

Feature Adoption

export interface FeatureAdoption {
  feature: string;
  totalUsers: number;
  adoptionRate: number;      // % of total users
  timeToAdoption: number;    // Days since signup
  retentionAfterAdoption: number;  // % still using after 7 days
}

// Target adoption rates:
const targetAdoption = {
  activityTracking: 0.95,     // 95% core feature
  aiAssistant: 0.70,           // 70% AI engagement
  voiceInput: 0.40,            // 40% voice adoption
  familySharing: 0.60,         // 60% multi-user
  analytics: 0.80,             // 80% view insights
  exportReports: 0.25,         // 25% premium feature
};

Dashboard Queries:

-- Feature adoption over time
SELECT
  feature_name,
  COUNT(DISTINCT user_id) as users,
  COUNT(DISTINCT user_id)::float /
    (SELECT COUNT(*) FROM users WHERE created_at <= CURRENT_DATE) as adoption_rate,
  AVG(EXTRACT(DAY FROM first_use_at - u.created_at)) as avg_days_to_adoption
FROM feature_usage fu
JOIN users u ON fu.user_id = u.id
WHERE fu.first_use_at >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY feature_name
ORDER BY adoption_rate DESC;

3. Retention Metrics

Cohort Retention Analysis

export interface CohortRetention {
  cohort: string;            // e.g., "2025-01-W1"
  day0: number;              // 100% (signup)
  day1: number;              // Target: >40%
  day7: number;              // Target: >60%
  day30: number;             // Target: >40%
  day90: number;             // Target: >30%
}

Dashboard Queries:

-- Weekly cohort retention
WITH cohorts AS (
  SELECT
    user_id,
    DATE_TRUNC('week', created_at) as cohort_week
  FROM users
),
retention AS (
  SELECT
    c.cohort_week,
    COUNT(DISTINCT c.user_id) as cohort_size,
    COUNT(DISTINCT CASE
      WHEN DATE(s.last_active_at) = DATE(c.cohort_week)
      THEN s.user_id
    END) as day0,
    COUNT(DISTINCT CASE
      WHEN DATE(s.last_active_at) = DATE(c.cohort_week) + INTERVAL '1 day'
      THEN s.user_id
    END) as day1,
    COUNT(DISTINCT CASE
      WHEN DATE(s.last_active_at) BETWEEN
        DATE(c.cohort_week) AND DATE(c.cohort_week) + INTERVAL '7 days'
      THEN s.user_id
    END) as day7,
    COUNT(DISTINCT CASE
      WHEN DATE(s.last_active_at) BETWEEN
        DATE(c.cohort_week) AND DATE(c.cohort_week) + INTERVAL '30 days'
      THEN s.user_id
    END) as day30
  FROM cohorts c
  LEFT JOIN user_sessions s ON c.user_id = s.user_id
  GROUP BY c.cohort_week
)
SELECT
  cohort_week,
  cohort_size,
  ROUND(day0::numeric / cohort_size * 100, 2) as day0_retention,
  ROUND(day1::numeric / cohort_size * 100, 2) as day1_retention,
  ROUND(day7::numeric / cohort_size * 100, 2) as day7_retention,
  ROUND(day30::numeric / cohort_size * 100, 2) as day30_retention
FROM retention
ORDER BY cohort_week DESC;

4. Monetization Metrics

Conversion & Revenue

export interface MonetizationMetrics {
  // Trial & Subscription
  trialStarts: number;
  trialToPayingConversion: number;  // Target: >30%
  churnRate: number;                // Target: <5% monthly

  // Revenue
  mrr: number;                      // Monthly Recurring Revenue
  arpu: number;                     // Average Revenue Per User
  ltv: number;                      // Lifetime Value
  cac: number;                      // Customer Acquisition Cost
  ltvCacRatio: number;              // Target: >3

  // Pricing tiers
  premiumSubscribers: number;
  premiumAdoptionRate: number;      // % of active users
}

Dashboard Queries:

-- MRR trend and growth
SELECT
  DATE_TRUNC('month', subscription_start_date) as month,
  COUNT(*) as new_subscriptions,
  COUNT(*) FILTER (WHERE previous_subscription_id IS NOT NULL) as upgrades,
  COUNT(*) FILTER (WHERE subscription_end_date IS NOT NULL) as churned,
  SUM(price) as mrr,
  LAG(SUM(price)) OVER (ORDER BY DATE_TRUNC('month', subscription_start_date)) as previous_mrr,
  ROUND((SUM(price) - LAG(SUM(price)) OVER (ORDER BY DATE_TRUNC('month', subscription_start_date))) /
    NULLIF(LAG(SUM(price)) OVER (ORDER BY DATE_TRUNC('month', subscription_start_date)), 0) * 100, 2
  ) as mrr_growth_rate
FROM subscriptions
GROUP BY month
ORDER BY month DESC;

-- LTV calculation
WITH user_revenue AS (
  SELECT
    user_id,
    SUM(amount) as total_revenue,
    MIN(payment_date) as first_payment,
    MAX(payment_date) as last_payment,
    COUNT(*) as payment_count
  FROM payments
  WHERE status = 'completed'
  GROUP BY user_id
)
SELECT
  AVG(total_revenue) as avg_ltv,
  PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY total_revenue) as median_ltv,
  AVG(EXTRACT(DAY FROM last_payment - first_payment)) as avg_lifetime_days
FROM user_revenue;

-- Churn analysis
SELECT
  DATE_TRUNC('month', cancelled_at) as month,
  COUNT(*) as churned_users,
  AVG(EXTRACT(DAY FROM cancelled_at - subscription_start_date)) as avg_days_before_churn,
  cancellation_reason,
  COUNT(*) FILTER (WHERE cancellation_reason IS NOT NULL) as reason_count
FROM subscriptions
WHERE cancelled_at IS NOT NULL
GROUP BY month, cancellation_reason
ORDER BY month DESC, reason_count DESC;

5. Product Quality Metrics

Technical Health

export interface QualityMetrics {
  // Performance
  apiResponseTimeP95: number;       // Target: <2s
  apiResponseTimeP99: number;       // Target: <3s
  errorRate: number;                // Target: <1%

  // Reliability
  uptime: number;                   // Target: >99.9%
  crashFreeUsers: number;           // Target: >98%
  crashFreeS essions: number;        // Target: >99.5%

  // User satisfaction
  appStoreRating: number;           // Target: >4.0
  nps: number;                      // Net Promoter Score (target: >50)
  csat: number;                     // Customer Satisfaction (target: >80%)
}

Dashboard Queries:

-- API performance monitoring
SELECT
  DATE_TRUNC('hour', timestamp) as hour,
  endpoint,
  COUNT(*) as request_count,
  ROUND(AVG(response_time_ms), 2) as avg_response_time,
  PERCENTILE_CONT(0.95) WITHIN GROUP (ORDER BY response_time_ms) as p95_response_time,
  PERCENTILE_CONT(0.99) WITHIN GROUP (ORDER BY response_time_ms) as p99_response_time,
  COUNT(*) FILTER (WHERE status_code >= 500) as server_errors,
  COUNT(*) FILTER (WHERE status_code >= 400 AND status_code < 500) as client_errors
FROM api_logs
WHERE timestamp >= NOW() - INTERVAL '24 hours'
GROUP BY hour, endpoint
HAVING COUNT(*) > 100  -- Only endpoints with significant traffic
ORDER BY hour DESC, p99_response_time DESC;

-- Crash analytics
SELECT
  DATE(created_at) as date,
  platform,
  app_version,
  COUNT(DISTINCT user_id) as affected_users,
  COUNT(*) as crash_count,
  error_message,
  stack_trace
FROM error_logs
WHERE severity = 'fatal'
  AND created_at >= CURRENT_DATE - INTERVAL '7 days'
GROUP BY date, platform, app_version, error_message, stack_trace
ORDER BY affected_users DESC;

---

Analytics Dashboard Templates

1. Executive Dashboard (Daily Review)

Key Metrics Card Layout:

┌─────────────────────────────────────────────────────┐
│ Daily Active Users    │  MRR              │  Uptime │
│ 5,234 ↑ 12%          │  $12,450 ↑ 8%     │  99.98% │
├─────────────────────────────────────────────────────┤
│ New Signups          │  Churn Rate        │  NPS    │
│ 342 ↑ 5%            │  4.2% ↓ 0.3%       │  62 ↑ 3 │
└─────────────────────────────────────────────────────┘

📊 7-Day User Growth Trend
[Line chart: DAU over time]

📊 Feature Adoption (Last 7 Days)
[Bar chart: % of users by feature]

🚨 Alerts & Issues
• P95 response time elevated (2.3s, target: 2.0s)
• Crash rate on Android 1.2.0 (3.1%, target: <2%)

2. Product Analytics Dashboard

User Journey Funnel:

-- Onboarding funnel conversion
SELECT
  'App Download' as step,
  1 as step_number,
  COUNT(*) as users,
  100.0 as conversion_rate
FROM downloads
WHERE created_at >= CURRENT_DATE - INTERVAL '7 days'

UNION ALL

SELECT
  'Registration Started' as step,
  2,
  COUNT(*),
  ROUND(COUNT(*)::numeric / (SELECT COUNT(*) FROM downloads WHERE created_at >= CURRENT_DATE - INTERVAL '7 days') * 100, 2)
FROM users
WHERE created_at >= CURRENT_DATE - INTERVAL '7 days'

UNION ALL

SELECT
  'Onboarding Completed' as step,
  3,
  COUNT(*),
  ROUND(COUNT(*)::numeric / (SELECT COUNT(*) FROM downloads WHERE created_at >= CURRENT_DATE - INTERVAL '7 days') * 100, 2)
FROM users
WHERE onboarding_completed_at IS NOT NULL
  AND created_at >= CURRENT_DATE - INTERVAL '7 days'

UNION ALL

SELECT
  'First Activity Logged' as step,
  4,
  COUNT(DISTINCT user_id),
  ROUND(COUNT(DISTINCT user_id)::numeric / (SELECT COUNT(*) FROM downloads WHERE created_at >= CURRENT_DATE - INTERVAL '7 days') * 100, 2)
FROM activities
WHERE created_at >= CURRENT_DATE - INTERVAL '7 days'
  AND created_at <= (SELECT created_at FROM users WHERE user_id = activities.user_id) + INTERVAL '24 hours'

ORDER BY step_number;

User Segmentation:

export enum UserSegment {
  NEW_USER = 'new_user',              // < 7 days
  ENGAGED = 'engaged',                // 3+ activities/day
  AT_RISK = 'at_risk',                // No activity in 7 days
  POWER_USER = 'power_user',          // Top 20% by activity
  PREMIUM = 'premium',                // Paid subscription
  CHURNED = 'churned',                // No activity in 30 days
}

// Segment users for targeted interventions
const segments = {
  new_user: {
    criteria: 'days_since_signup < 7',
    action: 'Send onboarding emails',
  },
  engaged: {
    criteria: 'activities_per_day >= 3',
    action: 'Upsell premium features',
  },
  at_risk: {
    criteria: 'days_since_last_activity >= 7 AND < 30',
    action: 'Re-engagement campaign',
  },
  churned: {
    criteria: 'days_since_last_activity >= 30',
    action: 'Win-back campaign',
  },
};

3. A/B Testing Dashboard

Experiment Tracking:

export interface ABTest {
  id: string;
  name: string;
  hypothesis: string;
  variants: {
    control: {
      users: number;
      conversionRate: number;
    };
    variant: {
      users: number;
      conversionRate: number;
    };
  };
  pValue: number;                     // Statistical significance
  winner?: 'control' | 'variant';
  status: 'running' | 'completed' | 'cancelled';
}

// Example: Test new onboarding flow
const onboardingTest: ABTest = {
  id: 'exp_001',
  name: 'New Onboarding Flow',
  hypothesis: 'Simplified 3-step onboarding will increase completion rate from 75% to 85%',
  variants: {
    control: {
      users: 1000,
      conversionRate: 0.75,
    },
    variant: {
      users: 1000,
      conversionRate: 0.82,
    },
  },
  pValue: 0.03,  // Statistically significant (< 0.05)
  winner: 'variant',
  status: 'completed',
};

Dashboard Queries:

-- A/B test results
WITH test_users AS (
  SELECT
    experiment_id,
    variant,
    user_id,
    CASE WHEN action_completed THEN 1 ELSE 0 END as converted
  FROM ab_test_assignments
  WHERE experiment_id = 'exp_001'
)
SELECT
  variant,
  COUNT(*) as total_users,
  SUM(converted) as conversions,
  ROUND(AVG(converted) * 100, 2) as conversion_rate,
  ROUND(STDDEV(converted), 4) as std_dev
FROM test_users
GROUP BY variant;

-- Calculate statistical significance (chi-square test)
-- Use external tool or statistics library

---

Monitoring & Alerting

Alert Rules

Critical Alerts (PagerDuty/Slack)

alerts:
  - name: "High Error Rate"
    condition: "error_rate > 5%"
    window: "5 minutes"
    severity: "critical"
    notification: "pagerduty"

  - name: "API Response Time Degradation"
    condition: "p95_response_time > 3s"
    window: "10 minutes"
    severity: "high"
    notification: "slack"

  - name: "Database Connection Pool Exhausted"
    condition: "active_connections >= 95% of pool_size"
    window: "1 minute"
    severity: "critical"
    notification: "pagerduty"

  - name: "Crash Rate Spike"
    condition: "crash_rate > 2%"
    window: "1 hour"
    severity: "high"
    notification: "slack"

Business Alerts (Email/Slack)

alerts:
  - name: "Daily Active Users Drop"
    condition: "today_dau < yesterday_dau * 0.8"
    window: "daily"
    severity: "medium"
    notification: "email"

  - name: "Churn Rate Increase"
    condition: "monthly_churn > 7%"
    window: "weekly"
    severity: "medium"
    notification: "slack"

  - name: "Low Onboarding Completion"
    condition: "onboarding_completion_rate < 80%"
    window: "daily"
    severity: "low"
    notification: "email"

---

Rapid Iteration Framework

Week 1-2 Post-Launch: Monitoring & Triage

**Focus**: Identify and fix critical issues

Daily Tasks:
- [ ] Review crash reports (target: <2%)
- [ ] Check error logs and API failures
- [ ] Monitor onboarding completion rate (target: >90%)
- [ ] Track day-1 retention (target: >40%)

Weekly Review:
- Analyze user feedback from in-app surveys
- Identify top 3 pain points
- Prioritize bug fixes vs. feature requests
- Plan hotfix releases if needed

Week 3-4: Optimization

**Focus**: Improve core metrics

Experiments to Run:
1. A/B test onboarding flow variations
2. Test different push notification timings
3. Optimize AI response quality
4. Improve activity tracking UX

Success Metrics:
- Increase day-7 retention to 60%
- Increase AI assistant usage to 70%
- Reduce time-to-first-value to <2 minutes

Month 2: Feature Iteration

**Focus**: Expand value proposition

Based on Data:
- Identify most-used features (double down)
- Identify least-used features (improve or remove)
- Analyze user segments (power users vs. casual)
- Test premium feature adoption

New Features to Test:
- Sleep coaching (if sleep tracking is popular)
- Meal planning (if feeding tracking is high-engagement)
- Community forums (if users request social features)

---

Tools & Integration

Recommended Analytics Stack

Core Analytics: PostHog (open-source, self-hosted)

// Backend integration
import { PostHog } from 'posthog-node';

const posthog = new PostHog(
  process.env.POSTHOG_API_KEY,
  { host: 'https://app.posthog.com' }
);

// Track events
posthog.capture({
  distinctId: userId,
  event: 'activity_logged',
  properties: {
    type: 'feeding',
    method: 'voice',
  },
});

// User properties
posthog.identify({
  distinctId: userId,
  properties: {
    email: user.email,
    isPremium: subscription.isPremium,
    familySize: family.members.length,
  },
});

Error Tracking: Sentry

import * as Sentry from '@sentry/node';

Sentry.init({
  dsn: process.env.SENTRY_DSN,
  environment: process.env.NODE_ENV,
  tracesSampleRate: 1.0,
});

// Automatic error capture
// Manual events
Sentry.captureMessage('User encountered issue', 'warning');

Uptime Monitoring: UptimeRobot / Pingdom

checks:
  - name: "API Health"
    url: "https://api.maternalapp.com/health"
    interval: "1 minute"

  - name: "Web App"
    url: "https://app.maternalapp.com"
    interval: "1 minute"

Performance: Grafana + Prometheus

# prometheus.yml
scrape_configs:
  - job_name: 'maternal-app-backend'
    static_configs:
      - targets: ['localhost:3000/metrics']

---

Conclusion

This analytics framework enables:

1. Data-Driven Decisions: Track what matters
2. Rapid Iteration: Identify and fix issues quickly
3. User Understanding: Segment and personalize
4. Business Growth: Monitor revenue and churn
5. Product Quality: Maintain high standards

Review dashboards daily, iterate weekly, and adjust strategy monthly based on real-world usage data.