User stickiness—how often your users come back—is one of the most important signals of product-market fit. If you're losing users after their first week, no amount of acquisition can fix it. Amplitude's retention analysis and cohort tools let you track this in real time and identify which user segments are actually sticky.
Track Core Engagement Events
Before you can measure stickiness, you need to track the actions that define engagement for your product.
Initialize Amplitude and log key events
Set up the Amplitude SDK with your API key and log events that represent active engagement. For a SaaS product, this might be logging in, viewing a report, or exporting data. Pick 1-3 core actions that reflect actual value delivery.
import * as amplitude from '@amplitude/analytics-browser';
// Initialize with your API key
amplitude.init('YOUR_API_KEY', {
defaultTracking: {
sessions: true,
pageViews: true
}
});
// Log engagement events
amplitude.track('Report Viewed', {
report_type: 'custom_dashboard',
time_spent_seconds: 120
});
amplitude.track('Data Exported', {
export_format: 'csv',
rows_exported: 5000
});Add user properties to segment by stickiness
Set user properties that help you understand *why* users are or aren't sticky. Track things like signup date, company size, or feature adoption—these become your segmentation axes.
// Set user properties for segmentation
amplitude.setUserId(user.id);
amplitude.setUserProperties({
'signup_date': new Date(),
'company_size': 'mid-market',
'has_custom_dashboard': true,
'onboarding_completed': true,
'feature_adoption_score': 7
});Analyze Retention with Cohorts
Once you're logging engagement events, use Amplitude's Retention feature to measure how many users come back over time.
Set up a retention analysis in the UI
In Amplitude, go to Analyze > Retention and define your activation event (usually signup or first login) and your returning event (the engagement action you want to track). Amplitude automatically calculates what percentage of your cohorts return on day 1, 7, 14, 30, etc.
// Amplitude Retention UI setup:
// 1. Go to Analyze > Retention
// 2. Activation event: "User Signed Up" (tracked automatically)
// 3. Returning event: "Report Viewed" (your custom event)
// 4. Date range: Last 90 days
// 5. View by: Cohort (groups users by signup date)
//
// Amplitude calculates retention automatically:
// Day 1: 85% of users returned
// Day 7: 42% of users returned
// Day 30: 18% of users returnedCompare stickiness across user segments
Use the Property filter in the Retention view to segment your analysis by user type. Are enterprise customers stickier than startups? Do users who completed onboarding stick around longer? These comparisons reveal where to focus.
// Apply property filters in Amplitude's Retention view:
//
// Segment A: company_size = "enterprise"
// Day 7 retention: 68%
// Day 30 retention: 45%
//
// Segment B: company_size = "startup"
// Day 7 retention: 28%
// Day 30 retention: 8%
//
// Insight: Enterprise users are 2.4x stickier
// Action: Invest in enterprise features, audit startup onboardingExport retention data for deeper analysis
Click Export on your retention table to download the raw cohort data as CSV. This lets you combine Amplitude retention with churn reasons from your customer data or support tickets.
// After exporting from Amplitude Retention UI, you get a CSV:
// cohort_date, day_0, day_1, day_7, day_14, day_30
// 2025-12-01, 1000, 850, 420, 210, 180
// 2025-12-08, 950, 808, 399, 199, 160
//
// Combine with churn survey data in your DB:
// SELECT user_id, churn_reason
// FROM churn_feedback
// WHERE cohort_date = '2025-12-01'
// GROUP BY churn_reason;
//
// Match day-7 droppers with feedback to find patternsMonitor Stickiness Over Time
Retention analysis is a point-in-time snapshot. To track stickiness as a live metric, set up a dashboard and alerts.
Build a stickiness dashboard
In Amplitude, create a new Dashboard and add cards for daily active users (DAU), your retention table, and engagement events over time. This becomes your command center for stickiness health.
// Amplitude Dashboard setup:
// Card 1: DAU (Daily Active Users)
// - Event: "Report Viewed" or "Data Exported"
// - Metric: Uniques
// - Breakdown: By day
// - Filter: Exclude trials or test accounts
//
// Card 2: Retention (rolling)
// - Activation: "User Signed Up"
// - Returning: "Report Viewed"
// - Cohort span: Last 90 days
// - Group by: company_size
//
// Card 3: Event trend
// - Events: "Report Viewed", "Data Exported"
// - By day, compare week-over-week
//
// Refresh: Automatic, updates every 24 hoursSet up alerts for retention drops
Use Amplitude's Alerts feature to notify you when day-7 retention drops below your baseline. If retention was 42% and suddenly hits 30%, you need to investigate what broke.
// Create an alert in Amplitude:
// Chart to monitor: Retention (day 7)
// Condition: Falls below 35% (or your baseline)
// Frequency: Daily
// Notification: Slack or email
//
// Example trigger:
// "⚠️ Day 7 retention dropped to 30% (threshold: 35%)"
// "Last week: 42%, This week: 30% (-12 percentage points)"
// "Investigate: Recent product changes, onboarding updates, or traffic quality"Common Pitfalls
- Confusing DAU with stickiness—a high DAU hides cohorts churning in the background. Always break retention down by cohort or user segment.
- Wrong activation event—if your activation event is 'signup' but real engagement happens a week later, your retention will look terrible. Match activation to when users actually start using the product.
- Too many events in retention analysis—tracking 50 events fragments your data. Focus on 1-3 signals that directly indicate value delivered.
- Ignoring onboarding's impact on retention—a retention cliff in week 1 usually means your onboarding is broken, not that the product isn't sticky. Separate post-onboarding retention from the activation flow.
Wrapping Up
Stickiness is the metric that matters most—it tells you whether you've built something people actually use repeatedly. By tracking engagement events, analyzing retention by cohort, and monitoring with a dashboard, you'll catch problems early and validate fixes. If you want to track this automatically across tools and set up smart alerts, Product Analyst can help.