What Is Google Analytics 4 and How Its Data Model Works

Short answer

Google Analytics 4 (GA4) is Google's analytics service for websites and mobile apps. It answers three basic questions: who comes to you, where from, and what they do — view pages, click, submit forms, buy.

Its core idea: every interaction is an event. A page view, a scroll, a click on a phone number and a purchase are all recorded the same way: an event name plus parameters with details.

GA4 replaced Universal Analytics, which stopped processing data in 2023. The main shift is from a model built on sessions and pageviews to a model built on events and users.

The data model: four building blocks

Events

An event is anything a user does. GA4 has four kinds:

• Automatically collected: for example first_visit, session_start.
• Enhanced measurement: turned on with a switch, no code: page views, scrolls, outbound clicks, site search, file downloads, video and form interactions.
• Recommended: names Google suggests for common actions, such as generate_lead, sign_up, purchase.
• Custom: your own events when nothing above fits.

Each event carries parameters: page address, button text, order value. To use your own parameter in reports, register it as a custom dimension or metric.

Sending a lead event from a site where GA4 is installed with gtag.js:

gtag('event', 'generate_lead', {
  form_name: 'contact',
  value: 1
});

Users

GA4 counts users by a device and browser identifier, and optionally by your own User-ID for logged-in people. Reports show active users by default, along with new users.

Sessions

A session starts with the session_start event and ends after a period of inactivity, 30 minutes by default. GA4 adds the idea of an engaged session: one that lasted longer than 10 seconds, included a key event or had at least two page views. Engagement rate is the share of such sessions; bounce rate in GA4 is simply its inverse.

Key events

Key events are the events important to your business: a request, a call, a sign-up, a purchase. You mark any event as key in the admin panel, and GA4 then shows them across reports and attributes them to traffic sources. Earlier versions of GA4 called them conversions; Google Ads still uses that word.

How data gets in

1. Create a property and a data stream for your site or app.
2. Install the tag: paste the gtag.js snippet with your measurement ID (it starts with G-), use Google Tag Manager, or a CMS integration.
3. Check enhanced measurement settings.
4. Send the events you need and mark the important ones as key events.
5. Link Google Ads and, if needed, BigQuery for raw data export.

Standard reports

• Realtime — what is happening now; handy for checking that tracking works.
• Acquisition — where users and sessions come from: channels, sources, campaigns.
• Engagement — events, pages and screens, key events.
• Monetization — purchases and revenue, if e-commerce is set up.
• Retention — how often users come back.
• User attributes and Tech — geography, language, devices, browsers.

For questions the standard reports do not answer, use Explorations: free-form tables, funnels and path analysis.

How GA4 differs from Universal Analytics

Common beginner mistakes

• Not marking any key events, so reports show traffic but not results.
• Counting the same form twice: enhanced measurement and a custom event.
• Using custom parameters without registering them as dimensions.
• Keeping the default short data retention, which limits explorations to recent months.
• Not filtering internal traffic from the team.