The trend has been going on for years and is making its mark on personal data protection. In addition to legal structures (GDPR, CCPA), practical tools are piling up and, in addition to extensions, data transfer is being blocked at the level of browsers and even operating systems. Universal Analytics still relies only on the data it sees, and lately these have become less and less available.
Google responded to this situation by creating GA4, equipped with artificial intelligence and Machine Learning. GA4 collects all the data it can, passes it with algorithms and tries to model the real situation. That is, to show the picture of how it would be if it had the information about each user.
While Universal Analytics' measurement model was based on page views and sessions, GA4's measurement model is already based on events and parameters. So even in Google Analytics 4, a page view is considered an event. The reason for this change is that GA4, as the name of the previous beta suggests, can be perfectly used for measuring not only websites, but also Android and iOS applications. In the case of mobile applications, traditional Page View statistics are difficult to interpret.
In so-called Single Page applications, where all pages of a website are technically one page, there are no page loads between viewing different content.
In the past, the bounce rate was primarily the indicator that many people, and thus many of our clients, used to analyze how interested site visitors were, in the site's content.
With that, the bounce rate was easily manipulated and thus not exactly the perfect indicator. With the introduction of different engagement metrics in GA4, we can analyze in a more nuanced way whether the users coming to the page are really interested in the content of the specified destination page.
One such engagement metric will be:
Another change in GA4 is the rules for setting up "events," or GA Events. As in the case of Google Analytics 3, it is mandatory to ensure that the event has category and action values, and optionally label and value values. In Google Analytics 4, this changes. When configuring GA4 Events, we can define the parameters of the Events completely freely.
As we described above with statistics, bounce rate was previously not the best metric for measuring engagement, and will no longer be available in GA4 either. In our opinion, this was a good decision by Google. While many people liked to use the bounce rate to measure the success of a specific campaign or channel, few knew how the bounce rate is calculated. And with it the ease of manipulating the values of poorly set events.
Chances are many will miss it, but overall it's a positive move by Google, helping to reduce the number of bad analytics based on misinterpreted data.
Access to predictive indicators can also be a useful new feature of Google Analytics. These predictive statistics are determined based on Google's machine learning algorithms and aims to give specialists useful insight. An example of such a predictive indicator in GA4 could be:
As websites become better developed and users become more aware of privacy terms and, as a result, more likely to opt out of tracking their sessions, Google Analytics has made some major changes to continue to provide Web site owners with relevant insights.
GA4 addresses this issue by using AI to create models based on personal data. In doing so, it is important to provide sufficient insight into how users who do not want to be tracked interact with the websites.
Another new feature is DebugView. The DebugView feature allows you to detect errors in individual events or other technical settings in GA4
In the previous GA version, integration with Big Query was possible, but only for paying Google Analytics 360 users, not for free Google Analytics users. With GA4, this changes and now you can analyze the data of the respective account with BigQuery even with the free version.
On July 1, 2023, standard Universal Analytics pros no longer process data. After July 1, 2023, you can continue to view your UA reports for a certain period of time. However, new data will flow only to GA4-properties.
On Oct. 1, 2023, Google Analytics 360-property's processing extension ends. Therefore, it is a good idea to use GA4 now and see what benefits can be gained.