It’s been a few months since Google announced the release of Google Analytics 4 (GA4), and we’re still seeing plenty of confusion around the key differences between GA4 and the older Universal Analytics.
Before you read on, if you aren’t sure what Google Analytics 4 is, the features it brings to the analytical table, and how to get yourself set up on the platform (without losing your historic data), you might want to read our in-depth Google Analytics 4 guide here…
Below, we’ll take a deeper look at three of the fundamental differences between the two Google Analytics platforms, and why it’s important that you’re aware of them.
Different Measuring Models
The biggest difference between Universal Analytics and GA4 is the measurement model that they use. Universal Analytics uses a measurement model based on sessions and pageviews.
A session is a group of user interactions (hits) with a website which take place over a given timeframe. A session can contain multiple pageviews, events and eCommerce transactions.
Contrasting that, Google Analytics 4 uses a measurement model based on events and parameters. The principle here is that any interaction can be captured as an event. As a result, all Universal Analytics hit types translate to events in GA4.
What can be confusing is that in Universal Analytics (and all previous version of GA) an event has a category, action and label and is its own hit type. In GA4 there is no category, action, or label. Every hit is an event and events can (but do not have to) contain parameters.
For example, in GA4 there is an event called page_view and this event contains the parameters page_location (page URL), page_referrer (previous page URL) and page_title.
Events in GA4 are grouped into 4 categories:
Automatically collected events – These are the events automatically tracked when you install the GA4 base code. They include events such as page_view, first_visit and session_start
Enhanced Measurement events – Also automatically-collected alongside the base code, but you can be enabled or disabled depending on your website functionality. These events include scrolls, outbound clicks, site search and video engagement.
Recommended Events – These are events that Google recommends you set up and have been split into industry type. Other than the eCommerce recommended events, the recommended events from Google just seem to be broad recommendations and the naming conventions are not essential
Custom Events – These are events and parameters that you can create and implement yourself based on website requirements. Current limits (and this could change in the future) are 500 distinctly named events
Removal of Monthly Hit Limits
Another significant difference between Universal Analytics and GA4 is the removal of monthly hit limits. The free version of Universal Analytics had a monthly limit of 10m hits. That’s gone.
Some of our clients were finding it an issue collecting all the data that they needed while staying within this limit.
Instead, GA4 has a limit on the number of different events that can be captured (500). There is, at the time of writing, no limit as to the volume of hits that can be collected. This has resulted in a number of clients already opting for a GA4-first approach to their analytics.
Free Connection to BigQuery
The final difference I’m going to cover is GA4’s free connection to BigQuery. Previously this feature was only available to GA360 customers (and was one of the big differentiators between the free and paid versions of GA).
For those not familiar with BigQuery, it enables very large and complex data sets to be queried very quickly. If you have ever tried to create complex segments in GA you will be aware of the issue that sampling can have on your ability to analyse data.
Big Query takes the data out of GA and gives you the ability to interrogate it without the issue of sampling.
Clearly, there are some important differences between Google Analytics 4 (GA4) and Universal Analytics. And it’s essential that you understand these fully before you go about switching over (besides, we’d always recommend creating a new property to run alongside Universal Analytics).