---
title: "GA4 Explained: What Changed From Universal Analytics"
description: How GA4's event-based model differs from Universal Analytics' sessions, what it means for your reports, and why it matters for Google Ads bidding.
url: "https://steviemorris.com/analytics/ga4-explained-what-changed/"
canonical: "https://steviemorris.com/analytics/ga4-explained-what-changed/"
type: article
silo: na
category: analytics
location: 
primary_keyword: GA4 vs Universal Analytics
keywords:
  - GA4 explained
  - GA4 vs Universal Analytics
  - event-based analytics
  - GA4 Google Ads conversions
blueprint: 0
audience: UK business owners
date_published: 2026-07-14
date_modified: 2026-07-14
author: Stevie Morris
schema_types:
  - Article
  - FAQPage
  - BreadcrumbList
record_id: article__ga4-explained-what-changed
word_count: 1734
---

# GA4 Explained: What Changed From Universal Analytics and Why It Matters

GA4 replaced Universal Analytics' session-based model with event-based measurement — every interaction (page view, click, scroll, purchase) is logged as a discrete event rather than grouped into sessions. That changes how metrics like bounce rate and users are calculated, and it affects the conversion signals feeding Google Ads' automated bidding.

## The core shift: from sessions to events

Universal Analytics (UA) was built around the session. It watched a visitor arrive, bundled everything they did — pages viewed, time spent, whether they bounced — into one session, and reported on those sessions. GA4 throws that structure out. Everything a user does is logged as an individual event: page_view, scroll, click, video_start, purchase. Sessions still exist in GA4's reporting, but they're now calculated *from* the events (essentially, a session_start event plus everything that follows it), not the other way round.

This isn't a cosmetic change. It's a different measurement philosophy. UA was designed for a world of websites with fairly linear visits. GA4 was designed for a world where the same user might interact with you across a website, an app, and multiple devices, and Google wanted one model flexible enough to stitch all of that together. Event-based measurement is what makes that possible — you can send the same type of event from a web page or a mobile app and GA4 treats it consistently.

## What actually changed under the hood

- **Sessions are derived, not primary** — a session in GA4 is a group of events, not a fixed container events get sorted into, which changes how session counts can differ from UA for the same traffic
- **Bounce rate became engagement rate (inverted)** — UA counted a bounce as a single-page visit with no interaction; GA4 instead measures 'engaged sessions' (lasting 10+ seconds, with a conversion, or with 2+ page views), so the two numbers are not directly comparable
- **Users are counted differently** — GA4 leans more on 'active users' and can blend device-based and (if configured) identity-based signals, rather than UA's cookie-only user counting
- **Data-driven attribution is the default** — UA defaulted most accounts to last-click attribution; GA4 defaults to a data-driven model that distributes conversion credit across multiple touchpoints
- **Enhanced measurement events are automatic** — scrolls, outbound clicks, site search and video engagement are tracked out of the box in GA4 without extra tagging, which UA never did natively

## What this means for the reports you're used to reading

If you're a business owner who got used to checking sessions, bounce rate and average session duration in UA, GA4's standard reports can feel like they're speaking a different language — because in measurement terms, they are. A page that had a '45% bounce rate' in UA doesn't have a directly equivalent number in GA4; you'll see an engagement rate instead, and because engagement is measured differently, the figure can look better or worse even though nothing about the site changed.

The same applies to traffic and conversion totals. Two accounts I've looked at side by side — one still holding legacy UA data, one on GA4 — will rarely show identical numbers for the same period, even accounting for the UA/GA4 cut-over. Different session definitions, different attribution defaults and different bot-filtering logic all pull the numbers apart slightly. None of this means either tool is 'wrong'. It means year-on-year comparisons that span the UA-to-GA4 transition need to be read as directional, not exact — and it's part of why I always sense-check the story a client's GA4 dashboard is telling before I trust it as the basis for a bidding decision.

### The honest bit

> Universal Analytics stopped processing new data in 2023, and standard UA properties were later removed from the interface, so if your GA4 property was only switched on around the cut-over, you may have a gap or a mismatch in your historical trend line. If nobody actively migrated your goals, audiences and conversion events across from UA to GA4 at the time, there's a reasonable chance your GA4 setup is still running on Google's default configuration rather than one built around what your business actually needs to measure. That's worth checking rather than assuming.

## Why GA4 accuracy matters for your Google Ads bidding

1. **Smart Bidding learns from your conversion signal** — Automated bid strategies like Maximise Conversions and [Target CPA/ROAS](/data/how-to-set-google-ads-targets-cpa-roas-2/) optimise entirely towards whatever conversion action you've told Google Ads to chase. If that signal comes from GA4 via a linked conversion, and GA4 is counting the wrong event — or double-counting, or missing conversions that happen after a delay — the algorithm is confidently optimising towards the wrong outcome.
2. **Performance Max leans on GA4 audience signals** — [PMax](/ppc-paid-ads/performance-max-explained-how-it-works/) accepts audience signals as a starting hint for its own targeting, and first-party data pulled through GA4 (site visitors, engaged users) is genuinely useful input here. A GA4 property with thin or misconfigured events gives PMax weaker material to work with than an account measuring the right actions properly.
3. **Data-driven attribution changes what gets credit** — Because GA4 defaults to data-driven attribution across multiple touchpoints, campaigns that assist earlier in the journey can start showing more value than they did under UA's last-click model — which matters when you're deciding which campaigns to scale and which to cut.
4. **Consent Mode affects how much signal you get at all** — With cookie consent requirements now standard, GA4's Consent Mode fills measurement gaps for users who decline cookies using modelled data. Without it configured correctly, you can lose a meaningful slice of your conversion signal before it ever reaches Google Ads' bidding algorithms.

## Getting GA4 right before you scale ad spend

None of this is a reason to distrust GA4 — it's a reason to check it's actually set up to measure what your business cares about before you let Google Ads' automated bidding lean on it. In practice that means confirming your key conversion events are marked correctly (and only the right ones are marked as conversions), that GA4 is properly linked to your Google Ads account, and that the numbers in GA4 broadly make sense against what you know about your own business — enquiries, orders, calls.

Conversion tracking and analytics setup is one of the things I do as part of [managing an account](/ppc-management/), precisely because Smart Bidding and PMax are only as good as the signal you feed them. If you're not sure whether your GA4 is actually tracking what you think it is, that's worth checking before you scale spend on the back of it.

### Related services

- [Get a PPC audit](/ppc-audits/)
- [Talk to a PPC consultant](/ppc-consultant/)
- [PPC management](/ppc-agency/)

### Related reading

- [What Is Attribution in Google Ads (and Why Last-Click Is Lying to You)](/analytics/attribution-in-google-ads-explained-2/)
- [What Is Conversion Tracking and Why Most Accounts Get It Wrong](/analytics/conversion-tracking-explained-common-mistakes-2/)
- [How to Read a Google Ads Report Without Being Fooled by Vanity Metrics](/data/how-to-read-a-google-ads-report-2/)
- [A Plain-English Glossary of Google Ads Terms Every Business Owner Should Know](/ppc-paid-ads/google-ads-glossary-plain-english-guide/)

## Frequently asked questions

### Do I need to do anything now that Universal Analytics has been switched off?

If you haven't already, yes — you need a GA4 property actively collecting data, with your key conversion events configured and linked to Google Ads. UA stopped processing new data in 2023, so any UA-only setup has been reporting nothing for some time.

### Why don't my GA4 numbers match my old Universal Analytics numbers?

Because the two tools measure differently, not because one is broken. GA4 calculates sessions from events rather than as a fixed container, uses engagement rate instead of bounce rate, and defaults to data-driven attribution instead of last-click. Expect the same traffic to produce different numbers in each tool.

### What's the difference between a GA4 conversion and a Google Ads conversion?

A GA4 conversion is an event you've marked as a conversion inside GA4 (for example a purchase or a form submit). A Google Ads conversion is what your Ads account uses for bidding — it can be imported directly from GA4 via the linked-property conversion import, or tracked separately with the Google Ads tag. If both exist for the same action, it's worth checking you're not double-counting.

### Does GA4 use cookies, and does that affect tracking accuracy?

GA4 still relies on identifiers similar to cookies for measurement. With consent requirements meaning some visitors decline tracking, GA4's Consent Mode uses modelling to estimate the conversions it can't directly observe — but that modelled data is only as good as the Consent Mode configuration behind it.

### Should I link GA4 to Google Ads even if I already track conversions with the Google Ads tag?

Generally yes — the two aren't mutually exclusive. GA4 linking gives you audience sharing into Performance Max and Smart Bidding, plus a consistent view of the customer journey across channels, which the Google Ads tag alone doesn't provide.

### What is Consent Mode and do I actually need it?

Consent Mode is Google's framework for adjusting how GA4 and Google Ads tags behave based on a visitor's cookie consent choice, filling measurement gaps with modelled data rather than losing that visitor's activity entirely. If you've got any EEA traffic, Google requires a verified consent signal — it's not optional in any practical sense. UK-only advertisers aren't held to that same requirement, but if you're running a cookie consent banner anyway, skipping Consent Mode just throws away conversion signal you don't need to lose.

## Not sure what your GA4 is actually measuring?

I check conversion tracking and analytics setup as part of managing an account, because Smart Bidding and PMax are only ever as good as the signal you're feeding them. Start a conversation and I'll give you an honest read on what your GA4 is really telling you.

**Call 07410 907 104**