GA4 vs Universal Analytics: The Complete Comparison (2026)

Universal Analytics is gone. Google shut it down in July 2024, and if you're still comparing the two, it's likely because you're migrating a client, trying to reconcile historical data, or explaining to a stakeholder why the numbers look different. This guide covers everything you need to know.

The Core Philosophical Difference

Universal Analytics was built on a session-and-pageview model. A user visits → a session starts → pageviews are counted → the session ends. All metrics derived from this: time on site, bounce rate, pages per session.

GA4 is built on an event-based model. Everything is an event. A pageview is an event. A click is an event. A purchase is an event. Sessions are derived from events, not the other way around. This is architecturally more flexible and aligns with how modern apps actually work (SPAs, mobile apps, offline interactions).

Key Differences Side by Side

Data Model

Bounce Rate vs Engagement Rate

UA bounce rate measured single-page sessions. GA4 replaced this with engagement rate — the percentage of sessions where the user was engaged (10+ seconds, 2+ pages, or triggered a conversion). Bounce rate in GA4 is simply 1 - engagement rate.

This is a more honest metric. A user who reads your article for 12 minutes and bounces is not a bad user experience — UA was just wrong about that.

Attribution

UA used last non-direct click attribution by default. Organic search and email often got credit that really belonged to paid social that introduced the user weeks earlier.

GA4 uses data-driven attribution by default (for properties with enough data). It distributes credit across all touchpoints in the customer journey using machine learning. For properties with insufficient data, it falls back to last click.

This means your channel-level conversion numbers will differ between UA and GA4 even for the same historical period — not because of a bug, but because the attribution model is genuinely different.

Reporting Interface

UA had a flat, hierarchical left-nav with dozens of pre-built reports. Most analysts knew exactly where to find things.

GA4 has: - Reports — pre-built reports (Acquisition, Engagement, Monetization, Retention) - Explore — ad hoc analysis with drag-and-drop (replaces Custom Reports + Segments) - Advertising — attribution and conversion paths - Configure — events, conversions, audiences, custom dimensions

The learning curve is real. Explore (formerly Analysis Hub) is significantly more powerful than UA's Custom Reports, but requires more setup.

Funnels and Path Analysis

UA's Funnel Visualization was limited to pre-defined, rigid funnels using pageviews only.

GA4's Funnel Exploration supports: - Open funnels (users can enter at any step) - Closed funnels (must enter at step 1) - Event-based steps (not just pageviews) - Segment breakdowns within the funnel - Elapsed time between steps

GA4's Path Exploration lets you see what events happen before or after any given event. There is no UA equivalent.

Predictive Metrics

GA4 includes built-in predictive audiences powered by machine learning: - Purchase probability (7-day) - Churn probability (7-day) - Predicted revenue (28-day)

These are available as audience segments for Google Ads remarketing. UA had no equivalent.

Data Retention

UA retained data indefinitely (or until the property was deleted).

GA4 retains event-level data for 2 months by default, or up to 14 months if you change it. For longer retention, you need to export to BigQuery.

This is a major operational difference. If you want to query data older than 14 months in GA4, you need the BigQuery export running from day one.

Why GA4 Is Better for Most Use Cases

1. Cross-Platform Measurement

If you have a website and a mobile app, GA4 can unify that data into one property using the same user identity. UA required separate properties for web and app.

2. More Accurate Session Counting

UA had a known issue: a new session started at midnight regardless of user activity, and a new session started every time a UTM parameter changed mid-session. GA4 fixed both of these.

3. BigQuery Native Integration

Every GA4 property (even on the standard tier) can export raw, event-level data to BigQuery daily. This means you can run SQL queries against your raw analytics data without limits. UA only offered this on the premium (360) tier.

-- BigQuery: purchase conversion rate by traffic source
SELECT
  traffic_source.source,
  COUNT(DISTINCT user_pseudo_id) AS users,
  COUNTIF(event_name = 'purchase') AS purchases,
  ROUND(COUNTIF(event_name = 'purchase') / COUNT(DISTINCT user_pseudo_id) * 100, 2) AS cvr
FROM `your_project.analytics_XXXXXXXXX.events_*`
WHERE _TABLE_SUFFIX BETWEEN '20260101' AND '20260331'
GROUP BY 1
ORDER BY 3 DESC
sql

4. Privacy-Ready Architecture

GA4 was designed with consent mode, cookieless modeling, and user deletion APIs built in. UA was retrofitted for privacy, which created gaps.

Common Metric Discrepancies Between UA and GA4

If you're comparing historical UA data to GA4, expect differences. Key reasons:

• Session counting: GA4 counts sessions differently (no midnight resets, no UTM-triggered resets)
• User counting: GA4 uses a different device fingerprinting methodology
• Attribution: Different default attribution models change conversion counts by channel
• Bounce rate: Completely different definition — not comparable
• Pageviews: If your SPA doesn't fire virtual pageviews, GA4 will show fewer

A 10-20% variance on most metrics is normal. Larger variances usually indicate a tracking implementation issue on either side.

Migration Checklist

If you're still in the process of fully migrating:

• [ ] GA4 property created and tag deployed (web + app)
• [ ] Historical UA data exported to BigQuery or Google Sheets before UA deletion
• [ ] All key events (conversions) recreated in GA4
• [ ] Consent Mode V2 implemented
• [ ] Cross-domain tracking configured (if applicable)
• [ ] Internal traffic filters set up
• [ ] Data retention set to 14 months
• [ ] BigQuery export enabled
• [ ] Custom dimensions/metrics recreated in GA4
• [ ] Audiences recreated for Google Ads remarketing
• [ ] Google Ads linked to GA4 property

Verify your GA4 implementation is complete with a full property audit — start your audit →