Metrics &
KPIs

// analytics.js — structured event tracking for AARRR
// Every event has a name, stage, and properties object

const analytics = {
  track(event, stage, props = {}) {
    // In prod: segment.track() / mixpanel.track() / amplitude.track()
    console.log({ event, stage, props, timestamp: new Date().toISOString() });
  }
};

// ACQUISITION — how users arrive
analytics.track('page_viewed', 'acquisition', {
  source: document.referrer || 'direct',
  utm_campaign: new URLSearchParams(location.search).get('utm_campaign'),
});

// ACTIVATION — first value moment
async function handleSignup(user) {
  analytics.track('signup_completed', 'activation', { method: 'email' });
  await createProject(user);  // force the first value moment
  analytics.track('first_project_created', 'activation');
}

// RETENTION — returning users
const lastVisit = localStorage.getItem('last_visit');
const daysSince = lastVisit
  ? (Date.now() - new Date(lastVisit)) / 86400000
  : null;
if (daysSince !== null) {
  analytics.track('session_started', 'retention', { days_since_last: daysSince });
}
localStorage.setItem('last_visit', new Date().toISOString());

// REVENUE — paid conversion
analytics.track('subscription_started', 'revenue', {
  plan: 'pro', mrr: 29, trial_days: 14
});

// REFERRAL — sharing behaviour
analytics.track('invite_sent', 'referral', { channel: 'email', count: 3 });

-- Retention cohort analysis (PostgreSQL)
-- Shows % of users from each signup week who returned in weeks 1—8

WITH cohorts AS (
  SELECT
    user_id,
    DATE_TRUNC('week', created_at) AS cohort_week
  FROM users
  WHERE created_at >= '2026-01-01'
),
activity AS (
  SELECT
    user_id,
    DATE_TRUNC('week', event_at) AS activity_week
  FROM events
  WHERE event_name = 'session_started'
)
SELECT
  c.cohort_week,
  COUNT(DISTINCT c.user_id)                       AS cohort_size,
  COUNT(DISTINCT CASE WHEN week_number = 1  THEN a.user_id END) * 100.0
    / COUNT(DISTINCT c.user_id)                     AS week_1_retention,
  COUNT(DISTINCT CASE WHEN week_number = 4  THEN a.user_id END) * 100.0
    / COUNT(DISTINCT c.user_id)                     AS week_4_retention,
  COUNT(DISTINCT CASE WHEN week_number = 8  THEN a.user_id END) * 100.0
    / COUNT(DISTINCT c.user_id)                     AS week_8_retention
FROM cohorts c
JOIN activity a ON c.user_id = a.user_id
CROSS JOIN LATERAL (
  SELECT EXTRACT('week' FROM a.activity_week - c.cohort_week) AS week_number
) w
GROUP BY c.cohort_week
ORDER BY c.cohort_week;

-- Healthy SaaS benchmark: week-4 retention > 25%
-- "Smiling" retention curve = product has found product-market fit