Universal Analytics is gone. Google pulled the plug in July 2023, and historical data access ended in July 2024. Yet I still hear from analysts who think about GA4 through the lens of Universal Analytics — trying to find their old reports, recreating UA-style dashboards, and wondering why the numbers don’t match.

Here’s the truth: GA4 isn’t an upgrade to Universal Analytics. It’s a fundamentally different platform. Having migrated over 40 properties from UA to GA4, I can tell you that the sooner you stop comparing and start understanding GA4 on its own terms, the faster you’ll unlock its real power.

This guide breaks down exactly what changed — the data model, the metrics, the reports, the implementation — so you can finally leave UA thinking behind.

Quick Comparison Table

Feature	Universal Analytics	GA4
Data Model	Session-based (hits)	Event-based (all interactions are events)
Session Definition	30-min inactivity, resets at midnight/new campaign	30-min inactivity only, no midnight reset
Bounce Rate	Single-page sessions	Non-engaged sessions (under 10s, no key event)
Goals/Conversions	Goals (max 20 per view)	Key events (unlimited)
Views/Filters	Multiple views with filters	No views — use comparisons and BigQuery
Cross-Platform	Web only	Web + app in single property
Attribution	Last non-direct click	Data-driven attribution (default)
Data Retention	Unlimited	2 or 14 months (user-level)
BigQuery Export	GA360 only (paid)	Free for all properties
Privacy/Consent	Basic cookie opt-out	Consent Mode v2, cookieless modeling
Reporting	100+ standard reports	~20 standard reports + Explorations

The Fundamental Shift — Events Replace Sessions

This is the single most important thing to understand about GA4. Universal Analytics was built on a session-based model. Everything — pageviews, events, transactions — was a “hit” inside a session. The session was the fundamental unit.

GA4 uses an event-based model. Every interaction is an event. A pageview is the page_view event. A session starting is session_start. A purchase is purchase. Even scrolling triggers the scroll event automatically.

In practical terms:

• No more hit types. UA had pageview, event, social, transaction, and timing hits. In GA4, they’re all events with different parameters.
• Events carry parameters. Instead of UA’s rigid Category/Action/Label, GA4 events have flexible key-value parameters.
• Sessions still exist, but they’re derived. GA4 calculates sessions from session_start events rather than treating them as primary containers.

If you’re setting up GA4 event tracking for the first time, understanding this shift is essential.

What GA4 Does Better

Event Flexibility Without Code Changes

GA4 automatically collects page_view, scroll, click (outbound), file_download, and video engagement — no extra configuration needed. In UA, you needed custom event tags for all of these. GA4 also lets you create new events from existing ones directly in the interface, without GTM.

True Cross-Platform Measurement

UA needed separate properties for web and app. GA4 uses data streams — one property can include your website, iOS app, and Android app. A user who browses on mobile web and converts on the app can be tracked as a single journey.

Free BigQuery Export

This was a GA360 feature that cost six figures annually. Now every GA4 property can export raw event data to BigQuery for free. You get row-level event data with no sampling. If you’re serious about analytics, this alone justifies GA4.

Predictive Analytics

GA4’s machine learning models can predict purchase probability, churn probability, and predicted revenue — then let you build audiences from those predictions. UA never could do this.

Built-In Consent Mode

Consent Mode v2 lets GA4 model conversions for users who decline cookies. With GDPR and privacy regulations tightening, this isn’t a nice-to-have — it’s essential.

What GA4 Still Struggles With

Steep learning curve. The interface is completely different. Settings are scattered. Even experienced analysts need weeks to feel comfortable.

Limited standard reports. UA had over 100 pre-built reports. GA4 has roughly 20. Google wants you to use Explorations, but that puts the burden on you.

Data thresholds. When Google Signals is enabled, GA4 hides rows to protect privacy. Your reports can show different totals depending on date range and dimensions.

Missing UA features. Annotations are still missing. Custom channel groupings took years to arrive. Content grouping requires manual event parameters. These gaps are real.

14-month retention. UA kept user-level data indefinitely. GA4 maxes out at 14 months. For long-term analysis, you need BigQuery exports. Build that pipeline early.

Head-to-Head: Key Metrics

This is where the confusion hits hardest. Here’s what actually changed.

Bounce Rate vs. Engagement Rate. In UA, bounce rate was single-page sessions. GA4 leads with engagement rate — sessions longer than 10 seconds, with a key event, or 2+ page views. GA4’s bounce rate is simply 100% minus engagement rate. A user reading your article for 15 seconds is no longer a bounce.

Sessions. GA4 sessions are often 5-15% lower than UA because they no longer reset at midnight or when campaign parameters change.

Users. GA4 defaults to “Active Users” instead of UA’s “Total Users.” This typically shows lower numbers.

Goals vs. Key Events. UA goals had a limit of 20 per view. GA4 key events have no limit — just toggle any event on. The simplicity is a genuine improvement.

Metric	UA Definition	GA4 Definition	Impact
Bounce Rate	Single-page sessions	Non-engaged sessions	GA4 bounce rate 10-30% lower
Sessions	Resets at midnight + new campaign	30-min timeout only	GA4 sessions 5-15% lower
Users	Total Users (default)	Active Users (default)	GA4 users usually lower
Conversions	1 per goal per session	Configurable per occurrence or session	GA4 can count higher

Head-to-Head: Reporting

Standard Reports. GA4’s reports cover basics: acquisition, engagement, retention, monetization, demographics. They’re customizable but minimal compared to UA.

Explorations. GA4’s real analytical power. Free-form reports, funnel analyses, path explorations, segment overlaps — all drag-and-drop. Nothing in UA came close.

Looker Studio. Combined with free BigQuery export, you can build dashboards from raw event data — no sampling. This GA4 + BigQuery + Looker Studio pipeline is far more powerful than anything UA offered. Make it part of your tracking plan.

Head-to-Head: Implementation

The Tag. UA used analytics.js or gtag.js with UA- IDs. GA4 uses gtag.js with G- measurement IDs. In GTM, the UA tag type is gone — use the Google tag.

GTM Migration. Every custom event needs recreating as a GA4 Event tag. But GA4 automatically tracks many events that needed manual tags in UA. Review your UA tags, remove anything GA4 handles, and rebuild the rest.

Data Layer. E-commerce structure changed completely. UA used an ecommerce object with actions. GA4 expects an items array with item_id, item_name, price. If you’re setting up e-commerce, see my e-commerce tracking guide for approaches that minimize code changes.

Which Matters Now?

Only GA4 matters. UA is dead. Every hour spent comparing is an hour you could spend mastering GA4. My advice:

• Stop recreating UA reports in GA4. Build reports that answer your business questions instead.
• Learn Explorations. This is where GA4’s depth lives.
• Set up BigQuery export now. With 14-month retention, this is non-negotiable.
• Accept the metrics won’t match. Both were “correct” within their own model.
• Audit your event tracking. GA4’s automatic events are great but don’t cover everything.

FAQ

Can I still access my Universal Analytics data?

No. Google shut down UA data access in July 2024. If you didn’t export before that deadline, it’s gone. Set up BigQuery export for GA4 immediately to avoid repeating this mistake.

Why are my GA4 numbers different from UA?

Different data models produce different numbers. Sessions don’t reset at midnight, GA4 defaults to Active Users instead of Total Users, and bounce rate is calculated completely differently. These aren’t errors — they’re different measurement approaches.

Is GA4 better than Universal Analytics?

In some ways yes — cross-platform tracking, free BigQuery, predictive audiences, consent mode. In other ways GA4 is still catching up — fewer reports, data thresholds, steeper learning curve. Overall, more powerful but less accessible.

Do I need to change my GTM setup for GA4?

Yes. UA tags don’t work with GA4. You need a Google tag plus GA4 Event tags. However, GA4 auto-tracks many interactions that needed manual UA tags, so your container may end up simpler.

How long does GA4 retain data?

User-level data: 2 or 14 months (configurable). Set it to 14 months immediately. Standard report data remains indefinitely. For long-term user-level analysis, export to BigQuery.

Moving Forward

The UA-to-GA4 transition was painful. But we’re past that now. GA4 is genuinely capable once you learn to work with it. The event-based model is more flexible. BigQuery integration opens up analysis that was impossible in UA. Start with the fundamentals: understand the event model, set up your key events, enable BigQuery export, and invest time in Explorations.