How do I identify returning users vs. new users in GA4?
Applies to: Google Analytics 4 (GA4)
Last updated: May 2025
Problem
You want to distinguish between new users (first-time visitors) and returning users (who have visited before) to understand audience loyalty, engagement, and retention trends.
Solution
GA4 provides a built-in New/Returning Users dimension, which is automatically collected based on user ID or device ID. You can use this in standard reports, Explorations, and audience segments.
Step-by-Step Guide
Step 1: Use Standard Reports
- Go to https://analytics.google.com
- Navigate to Reports > User > User Attributes > New vs Returning
- This report shows:
- Total users
- Engagement metrics (like sessions, time, conversions) broken down by:
- New
- Returning
Step 2: Use in Explorations for Custom Analysis
- Go to Explore > Free Form
- Add the following:
- Dimension:
New/Returning - Metrics:
Users,Engagement Rate,Conversions, etc.
- Dimension:
- Drag dimensions and metrics into rows/columns
- Segment performance by new vs. returning across:
- Traffic sources
- Devices
- Pages
Step 3: Build Audience Segments
To isolate behavior by user type:
- Go to Explore > Segment
- Click User segment > Add condition
- Choose
New/ReturningequalsReturning(orNew) - Compare performance side-by-side in any Exploration
This is useful for identifying which content or campaigns resonate better with each group.
Step 4: Export for Retargeting (Optional)
If your GA4 is linked to Google Ads, you can:
- Create audiences for returning users
- Retarget them with tailored ads or promotions
- Use exclusions (e.g., exclude returning users from acquisition campaigns)
How GA4 Determines User Type
- New user: First visit to your site/app (based on device or user ID)
- Returning user: Has visited at least once before, and is recognized (same browser, user ID, or app instance)
Note: Users who clear cookies or switch devices may be counted as new again.
Notes
- GA4 does not use a “session-based” model like Universal Analytics; it’s event-based, but user type is tracked per user
- You can include
new_vs_returningas a custom dimension in exports or BigQuery for advanced analysis





