You're tracking events in your app, but you need to understand which user segments or properties drive the most engagement. Event segmentation in Amplitude breaks down your events by any property—user tier, region, feature—so you see exactly which dimensions matter.
Instrument Events with Properties
Event segmentation requires clean event data with properties. You need to send descriptive properties in your events so Amplitude has dimensions to segment by.
Initialize the Amplitude SDK
Set up the Amplitude JavaScript SDK with your API key. This enables event tracking and sets up the client for your app.
import * as amplitude from '@amplitude/analytics-browser';
amplitude.init('YOUR_API_KEY', {
userId: 'user_123',
defaultTracking: true
});Send Events with Property Values
Track events and include properties you want to segment by later. Pass properties as an object—subscription_tier, user_type, region, etc. These become your segmentation dimensions.
amplitude.track('Feature Used', {
feature_name: 'export_data',
subscription_tier: 'pro',
user_type: 'analyst',
region: 'us-east'
});
amplitude.track('Feature Used', {
feature_name: 'dashboard_created',
subscription_tier: 'starter',
user_type: 'manager',
region: 'eu-west'
});Set User Properties for Reusable Segments
If a property applies to all events from a user (like account tier or company size), use user properties instead. This avoids repeating the same property on every event and keeps your payloads lean.
amplitude.setUserProperties({
subscription_tier: 'pro',
account_owner: true,
company_size: 'enterprise'
});
// Now all events from this user include these properties automatically
amplitude.track('Feature Used', {
feature_name: 'query_builder'
});subscription_tier: 'Pro' one time and subscription_tier: 'pro' another, they'll segment separately.Create Segmentation in the Analytics UI
Once events flow into Amplitude, segment them in the UI without touching code.
Open the Events Analysis View
Go to the Analytics tab, then click Events in the sidebar. This shows all tracked events, their volume, and unique user counts over time.
Select an Event and Add a Segment Breakdown
Find your event (e.g., 'Feature Used'). Below the chart, click Segment by and choose a property—subscription_tier, region, user_type. Amplitude displays the event count broken down by each property value.
// The property must exist in your event for it to appear in the Segment dropdown.
// If it doesn't show up, verify it's in your event payload:
amplitude.track('Feature Used', {
feature_name: 'export',
subscription_tier: 'pro' // Now appears as a segment option
});Read the Segment Results
Amplitude shows total count, unique users, and percentage for each segment value. Compare segments to answer questions like 'Which user tier engages most?' or 'How does adoption differ by region?'
Filter and Compare Segments
Combine segmentation with filters to zoom in on specific cohorts.
Add Filters to Narrow Your View
Click Filter above the chart to restrict analysis to specific users. For example, filter to subscription_tier = 'pro' to see how only premium users engage with the event.
// Filters apply server-side in Amplitude Analytics.
// Ensure your property values are consistent for accurate filtering:
amplitude.track('Feature Used', {
feature_name: 'automation',
subscription_tier: 'pro', // Consistent case and format
trial_expired: false
});Build Formulas for Derived Metrics
Use formulas to compare normalized metrics across segments. For instance, calculate 'Feature Used / Login' to see feature adoption per session, accounting for size differences between segments.
Common Pitfalls
- Not sending event properties at all. If you track events without properties, you can't segment them. Plan your property schema before you start instrumenting.
- Inconsistent property values (case, spacing, format). Amplitude treats 'Pro', 'pro', and 'PRO' as separate segments. Standardize values before shipping events.
- Segmenting by high-cardinality properties like email, user ID, or full timestamps. This creates hundreds of segments and makes patterns invisible. Use categorical properties instead.
- Skipping filters alongside segmentation. Always apply context—filter to the cohort you care about before interpreting segment results.
Wrapping Up
Event segmentation in Amplitude reveals which user properties and dimensions drive engagement with your features. Start by instrumenting events with descriptive properties, then use the UI to segment and filter. You'll quickly answer questions about adoption patterns across user tiers, regions, and feature adoption. If you want to set up event segmentation automatically across all your tools, Product Analyst can help.