Universal Analytics (UA) has been around since 2005, so we can all agree that the upgrade to Google Analytics 4 (GA4) was long overdue, given how fast things are changing in the tech world. Although, this doesn’t make the transition any easier since many of us have gotten used to UA, and GA4 brings around some significant changes that’ll require some reading up and training. While the new tool has great potential, it also has several imperfections and bugs that hopefully, Google is going to address very soon.
Why is GA3 (UA) being replaced by GA4?
Our team here at Mountain Laurel Digital is curious by nature, and we like to confront things before they confront us, so we started getting our hands dirty as soon as the news came out. We found GA4 better addresses the more contemporary online user behavior encompassing multiple channels, devices, and properties like websites and apps. It is also equipped with data modeling to account for conversions and behavioral data for users who decline cookies based on users who accept them. It’s great news for companies operating in Europe bound by GDPR regulations, as well as several US states like California and Virginia. Modeling is available within the Blended Reporting Identity option, but it is only available to websites meeting certain requirements, such as having at least 1,000 events per day for at least seven days, coming from visitors declining cookies. As a result, many websites with fewer visitors (that have consent mode) will have trouble meeting those thresholds.
Finally, GA4 is also more accurate with counting users due to its improved cross-device tracking capability and the use of Google Signals (you can expect to see lower user data in GA4 than you saw in UA, thanks to that cross-device deduplication).
Some of the important differences between GA3 and GA4 include:
Views are gone
GA4 only has one Data Stream (aka data source), which is powerful, because it combines data from websites and apps and helps understand user behavior in a more holistic way. For example, this will be an adjustment for companies that used several views for various geographical areas. In GA4, you will need to apply segments to imitate that setup unless you have GA4 360, which allows sub-properties.
Channel Groups are no longer customizable
In Universal Analytics, you could define a Channel Group, but in GA4, it is no longer possible. Channel Groups are predefined and can’t be adjusted. You have to use UTM tags (we recommend using this URL Builder) to make sure your traffic falls into the desired Channel Group. A good practice we follow is adding UTM tags to all URLs we have control over, including organic social media posts. This will diminish the Direct or Unassigned traffic in GA4. You will notice that the Channel Groups in GA4 are expanded compared to what GA3 offered, so it gives the opportunity to have a more comprehensive understanding of the traffic sources.
Custom Events
Universal Analytics (UA) operates with users, sessions, and views, and while GA4 is user-based, everything now is an event. Events can be triggered whenever a new session is started, a page view is recorded, or when a user scrolls to 90% of the page. This is the first group of events that are tracked by default.
The second group of events is called Enhanced Measurement, and you can choose whether you want to track these events or not. Tip: we recommend disabling the event called “Form interactions” for now until GA4 resolves the bug (MeasureSchool has a great blog post about this).
The third group is Custom Events, which require additional implementations and might require a developer. But they are also very powerful, because you can track anything your website allows you to track (including e-commerce events).
Each event can have up to 25 parameters that describe an event and provide more context. For example, an outbound click event could have a click URL parameter (destination URL) or a page URL parameter (URL where the click occurred).
More flexible and robust reporting
This is still a work in progress and some reports leave a lot to be desired, but GA4 allows for extensive report customization to whatever a business needs. It’s a really fun feature once you start digging in!
First, we have the Overview Reports – this is similar to what GA3 offered, e.g., in the Acquisition Overview report. They provide a high-level overview of performance, as the name implies.
Then you can go a step further and create Detailed Reports. You can build these reports around specific metrics that you want to focus on. For example, you can build a report dedicated to Google Ads by customizing the report available in the Acquisition Overview, and you can even make it show up in the Acquisition category on the left-hand side for faster access.
If you want to dive deeper, you can create reports based on Custom Events and Parameters. We have a couple websites that have maps where visitors can search for physical stores or retailers. We set up additional reports for custom events that track outbound clicks to see where people click through the most on those maps. You could also pull out individual parameters like a click URL and create reports solely around that parameter (secret tip – they need to be defined as a Custom Dimension first!).
And lastly, there are also Exploration Reports, which give you the most freedom. You can analyze User Paths, Purchase Funnels, Purchase Cohorts, and more. You can see where people drop off your page, which steps of the checkout process need improvements, and even analyze specific segments and how they navigate the website, e.g., look at a path of only those people who submitted a contact form. If you are a data nerd, this is definitely where you want to live.
All data from GA4 can also be exported to Big Query for analysis. Make sure to export all your historical UA data before January 2024 if you want to be able to reference this data in the future since all UA will then be deprecated, and you won’t be able to access data anymore.
Bounce Rate
This is not a huge change compared to the points mentioned above, but we want to highlight it since many people are married to this metric. It has always been skewed data since so many users open multiple tabs and forget about them (a reminder to focus more on trends than metrics alone). GA4 is introducing new metrics to replace the bounce rate – Engaged Sessions and Engagement Rate. An Engaged Session happens whenever a user:
- Spends at least 10 seconds on the website (you can change that number, and we recommend you do so)
- Triggers an event
- Views more than one page
This is a much better metric to represent user engagement, but we recommend coupling it with other reports available in GA4, as well as additional tools like Hotjar or VWO, to get a full picture.
Things are still in motion, which means new features are being released, discoveries are being made, and new plugins and integrations are being rolled out to make all of the tools out there compatible with GA4. If you want to stay abreast of all the changes, here are some blogs we recommend following:
Author: Karolina Dróżdż of Mountain Laurel Digital
Discussion about this post