Analytics Comparison
Google Analytics 4 vs Universal Analytics
This is a comparison with a settled ending: Universal Analytics stopped processing new data on 1 July 2023, and GA4 is now the only version Google supports. Still, understanding how the two differ explains why GA4 reports look so unfamiliar and how to think in the new model. If you learned analytics in the UA era, the mental shift below is the important part. For how to actually read the new attribution, see our GA4 attribution models guide.
The short answer
Universal Analytics is retired; there is no choice to make about which to use. The real task is unlearning the session-and-pageview mindset. GA4 is built on an event-based data model that measures across web and app, is designed for a privacy-first, cookieless future, and reports very differently. Expect the numbers to not match UA, and do not try to force them to.
Side by side
| GA4 | Universal Analytics | |
|---|---|---|
| Status | Current and supported. | Retired. Stopped collecting data on 1 July 2023. |
| Data model | Event-based. Everything is an event with parameters. | Session and pageview based. |
| Scope | Web and app in one property. | Web-focused (Firebase handled apps separately). |
| Key metric shift | Engaged sessions, engagement rate, events. | Sessions, bounce rate, pageviews. |
| Bounce rate | Redefined as the inverse of engagement rate. | Percentage of single-page sessions. |
| Attribution | Data-driven attribution is the default. | Last-click was the historical default. |
| Privacy | Cookieless-ready, IP anonymized by default, modeling for gaps. | Cookie-dependent, built for a pre-consent era. |
| Reporting | Explorations, fewer prebuilt reports, more custom. | Many prebuilt reports out of the box. |
The one shift that matters: events, not sessions
Universal Analytics organized everything around sessions and pageviews. GA4 organizes everything around events. A pageview is an event, a scroll is an event, a click is an event, a purchase is an event, each carrying parameters. This is why direct comparisons break: a UA session and a GA4 engaged session are not the same unit. Once you accept the event model, GA4 becomes more flexible, because you can measure any interaction without the rigid hit types UA imposed.
Why the numbers do not match
Teams often panic when GA4 traffic looks different from UA. It should. Different session timeout logic, a redefined bounce rate, data-driven attribution instead of last-click, and consent-based modeling all move the numbers. The right response is to stop reconciling old and new, set fresh baselines in GA4, and define the conversion events that matter to your business. If last-click attribution is distorting your view of what paid search contributes, our piece on attribution models for paid search covers the fix.
What to do now
Configure GA4 properly rather than relying on defaults: mark your real conversions as key events, set up the enhanced measurement toggles you want, and build Explorations for the questions you used to answer with UA reports. Pair it with Search Console for the organic query data GA4 does not show, and treat GA4 as the source of truth for on-site behavior. For the strategy layer on top of the data, see our content strategy guide.
How to choose
GA4 is right when
- Always, for any new setup. It is the only supported version.
- You need combined web and app measurement.
- You want data-driven attribution across channels.
- You are planning for a cookieless, consent-first world.
Universal Analytics context
- It no longer collects data and cannot be used going forward.
- Historical UA exports may still be worth archiving for reference.
- Do not benchmark GA4 against old UA numbers one-to-one.
- Its session model is the mental habit most worth unlearning.