Module 1: GA4 Foundations
How GA4 Differs from Universal Analytics
If you have ever opened an old Google Analytics report and then opened Google Analytics 4 (GA4), you will know the feeling: the familiar menus are gone, the numbers do not match, and words such as "sessions" seem to behave differently. This lesson explains why. Once you understand the change in thinking behind GA4, the rest of the tool becomes far easier to learn, and you will stop trying to force it to work like the old one.
Why Google replaced Universal Analytics
Universal Analytics (UA) was designed in an era when most people visited websites on a desktop computer. Its data model was built around sessions and pageviews: a visit started, a few pages were viewed, and the visit ended. Apps were tracked in a separate product. As customers began moving between phones, laptops and apps, and as privacy expectations changed, that model became harder to stretch. Google stopped processing new data in standard Universal Analytics properties in July 2023, so GA4 is now the version you will use for any new measurement work.
GA4 was designed to measure websites and apps in one property, to rely less on cookies over time, and to give you more flexibility in what you track. The price of that flexibility is that you must make more decisions during setup.
The key differences you will notice
| Topic | Universal Analytics | GA4 |
|---|---|---|
| Data model | Sessions and hit types (pageview, event, transaction) | Everything is an event with parameters |
| Goals | Up to 20 goals per view | Key events, which you mark from any event |
| Engagement | Bounce rate as the headline measure | Engaged sessions and engagement rate; bounce rate is the opposite of engagement rate |
| Views | Multiple filtered views per property | No views; you use filters, comparisons and data streams instead |
| Ad hoc analysis | Mostly fixed standard reports | Explorations for free-form analysis |
| Web and app | Separate products | One property with several data streams |
The most important row is the first. In GA4 a page being viewed is simply an event called page_view. A purchase is an event called purchase. A form submission is whatever event you decide to send. This sounds technical, but it means you can measure almost any meaningful action in the same way.
Why the numbers do not match
Many teams worry when GA4 shows different user or session counts from their old reports. Some difference is normal. GA4 counts sessions differently (for example, a new campaign source in the middle of a visit does not start a new session as it did in UA), it reports "active users" by default, and it may apply thresholds or modelling where consent is limited. Do not try to reconcile the two systems to the last digit. Instead, establish a new baseline in GA4 and compare GA4 with GA4 going forward.
A worked example
Imagine Wanjiku runs a small online shop in Nairobi selling handmade leather bags. In UA she had one goal: visiting the "thank you" page after checkout. Her bounce rate looked high at around 70 percent, and she assumed her homepage was failing.
When she moved to GA4, she set up a purchase event with the order value, plus events for add_to_cart and clicks on her WhatsApp button. She discovered that many "bouncing" visitors were actually reading a product page for over a minute and then tapping WhatsApp to ask a question. GA4 counted these as engaged sessions. Her homepage was not the problem; her checkout was, because people preferred to order by chat. She added a clearer "Order on WhatsApp" option and started measuring it as a key event.
The lesson is that GA4 encourages you to measure actions, not just page loads, and that often changes the story.
Common mistakes to avoid
- Expecting UA reports in the same place. Learn the new structure rather than hunting for old menus.
- Comparing GA4 totals directly with historic UA totals. Treat GA4 as a new baseline.
- Leaving GA4 on default settings only. Defaults collect page views and some automatic events, but your business actions need deliberate setup.
- Assuming bounce rate means the same thing. In GA4 it is the share of sessions that were not engaged.
In the next lesson you will look closely at events, the building block of everything in GA4.
Exercise
Write down the three most important actions a visitor can take on a website you manage or know well (for example, buying, submitting an enquiry, or tapping a WhatsApp button). Next to each, note how Universal Analytics would have tracked it and how you would describe it as a GA4 event with a short name such as generate_lead. Keep this list; you will use it throughout the course.
Key points
- GA4 is built on events, not sessions and pageviews.
- Universal Analytics stopped processing standard data in July 2023, so GA4 is the tool for new work.
- Expect GA4 numbers to differ from old UA numbers and set a fresh baseline.
- Engagement rate replaces bounce rate as the main quality signal.