Google Analytics 4 vs Universal Analytics.
If you haven’t had time to get your head around this big change, need help with the transition or just want to know more about GA4 this is the read for you.
With less than one year left before Universal Analytics (UA) sunsets, anyone who uses this tool (you probably do, right?) to measure the success of their website is thinking about making the transition to Google Analytics 4 (GA4).
You likely have a bunch of questions about what it does differently, how to manage the migration and what will happen with your historical data. We cover all of that and more, so keep reading to find out everything you need to know about the changes and why you should start tracking with GA4 now.
What is happening with Universal Analytics
We’ve all heard the quote “You can’t improve what you don’t measure”. For almost a decade Universal Analytics has been the industry standard for tracking and measuring website activity. In fact over 28 million sites—including 74% of the most popular 10,000—were using the tool as of April 2022.
After releasing a beta version of GA4 2 years ago, Google has decided to permanently sunset UA for standard customers on the 1st July 2023 whilst Universal Analytics 360 customers will be granted an additional 3 months of use.
From the dates mentioned above UA will no longer process any new hits and the properties will become read only so users will only be able to access UA properties for historical data. Google has stated that users will be able to access previously processed data in your Universal Analytics property for at least six months. It is strongly advised to export your historical reports during this time so you can still refer back to this valuable data. After this period the data in UA will be removed completely.
You may be wondering why the decision was made to remove UA entirely. There are several–kinda complicated–reasons for this. Google has said the main reason for the change is that UA is dated, and was not built to understand multi-platform journeys thoroughly. The session based data model is 15 years old - before devices like smartphones were widely used.
User privacy was another big reason for the decision. Entire countries have banned Google Analytics (GA), including France, who dropped the guillotine on GA this year, declaring it “illegal” for breaching EU GDPR.
So it is no wonder Google is taking these privacy concerns seriously. As such, GA4 is geared towards providing more privacy-centric solutions for users by ushering in the “cookieless era”, which is one of the (many) big differences between UA and GA4.
How does cookieless tracking work in GA4?
Cookies have been used on the internet for decades to track users by storing a small piece of data about a website user. They have been essential in Digital Marketing and Paid Advertising, particularly for remarketing campaigns which follow you across the internet. Remember that time you were thinking about buying a paddleboard? No? Well Google does.
The crux of the issue concerns these types of third party cookies which track users across multiple domains. Over time, they can learn alot about you by piecing together your personal data. Which is great for advertisers, but it is difficult to regulate how people’s personal data is used–or–abused.
The solution GA4 provides is called consent mode. It lets you adjust how your Google Tags behave based on the consent status of your users. You can indicate whether consent has been granted for Analytics and Ads cookies. Google's tags will dynamically adapt, only utilising measurement tools for the specified purposes when consent has been given by the user.
How does GA4 Identify users?
GA4 has 3 ways of identifying people who visit your website: User ID, Google Signals, and by device. You’ll need to select a default reporting identity to gather data about your users and have it applied to your reports. This is important for users that need to measure activity across multiple platforms. Here’s a bit more information about each way of reporting on identity:
User ID: This was first introduced with Universal Analytics. This ID is set by you with a unique identifier. With user ID, after the user has logged into your site, your authentication system can assign them this ID. This should be consistent across your mobile apps and website for the same user.
Device ID: This ID is basically a first party cookie that is set automatically. It’s stored on the user’s first visit and set to remain for two years. In the original version of Google Analytics, this was the only way visitors were identified.
Google Signals: This uses data from people who use Google Account and have enabled ads personalisation. However, the most accurate way of identifying visitors is User ID. But it works well only if people can log in to your website.
How does GA4 account for missing data?
So what about the users who don't consent for Analytics? Google announced a feature called Behavioural Modelling for Consent Mode which will combine the data from users that opted-in with other cookieless data to make total user data estimations. This type of data modelling uses very clever machine learning to fill in those gaps to estimate the actual total number of users, page views, sessions, transactions, revenue and other metrics that you would attract if you didn’t need consent.
Google has released improvements for this model since its launch so that websites with short data history or low conversion volume, can finally get access. Yay!
As these are black box models, we don’t actually know how they work, so we are putting our faith in Google to provide accurate estimations. Google has already stated that we can expect the un-consented conversion rates to be significantly lower than the consented conversion rates. They reported that consented users are typically 2-5x more likely to convert than un-consented users. Which makes sense, as someone that intends to convert, must have trust in the website and is therefore more likely to opt-in.
Universal Analytics vs. Google Analytics
Over the past year or so of getting to know GA4, keeping up to date with new features and changes we can attest that there is a steep learning curve for GA4. Despite the changes, we are tentatively optimistic about this new marketing platform and the insights it can provide.
A few key differences between Universal Analytics and Google Analytics 4 are:
Different measuring models
As previously mentioned, the biggest difference between UA & GA4 is the measurement model that they use. UA uses a session based model which uses sessions and page views.
In UA a session (or hit) is measured with cookies, from the time one browser begins accessing a site, until it leaves and all the page views in between. A session can consist of multiple user interactions including; page views, events, social interactions, and e-commerce transactions. UA ends a session after thirty minutes of inactivity and If cookies are blocked, nothing is tracked at all.
In contrast, GA4 uses an events & parameters based model. Metrics are derived from an automatically collected session_start event, and every user interaction after that is captured as an event. The duration of a session is based on the time between the first and the last session.
If you’re familiar with setting up events in Universal Analytics you will be familiar with events having a category action and label and it being its own hit type. What can be confusing is that GA4 events have none of these, and every hit is an event that can have event parameters–these are additional pieces of info that can specify the action the user took, or add context to the event–but these are not essential.
For example, in GA4 there is an automatically created event called page_view collects the following parameters automatically; page_location (page URL), page_referrer (previous page URL), engagement_time_msec.
Types of events in GA4
There are 4 event categories in GA4:
Automatically collected events
Enhanced measurement events
Recommended events
Custom events
Automatically Collected Events
These events are automatically created and tracked when you install GA4.They are triggered on certain predefined user interactions such as page_view, first_visit and session_start.
Enhanced measurement events
These events are also automatically collected on certain predefined user interactions, but you can enable or disable these depending on your website functionality. These events can include interactions with web page elements such as video, button, image, external links etc.
Recommended events
These are events recommended by Google with predefined naming conventions and parameters that are recommended for different industries. These events are not triggered unless they are manually implemented. Check out the lists of recommended events that Google has published for the following industries:
Custom Events
These are events and parameters that you can create and implement yourself based on website requirements. Google recommends that before you create a custom event to make sure that there is no automatic, enhanced or recommended event that already provides what you need. The current limitations of custom events are there are 500 distinctly named events.
Data Streams
In UA each property may contain multiple views. Views allow for data filtering and configurations. Currently this functionality does not exist with GA4. In GA4 you have a single reporting view and data streams that feed into it instead of having a separate property for each. Data Streams are a way of collecting different sources of data under one property so you can add multiple web and app streams to each property.
In GA4, you apply filters at the property level to customise your single reporting view. So, instead of view filters like you had in Universal Analytics, you now have data filters. When you set up a data filter, you’ll apply it directly to a data stream.
This gives us the ability to view cross platform data in one place. However, if you currently rely on UA reporting at multiple views to distinguish each domain, you will no longer be able to do this in GA4.
User-Interface
The GA4 user-interface looks pretty different. Reports and metrics that are standard in UA are no longer available in GA4 because of the new measurement model. Universal Analytics was built as a comprehensive suite of pre-made reports that can be used out of the box to analyse online performance. Marketers and businesses should not expect to find many of the old default reports and even some features that they are used too.
For example, you no longer have a behaviour flow report in the new interface. However, there is a replacement of sorts in GA4 in the form of the path exploration report or the funnel exploration report - These GA4 reports are limited and sometimes bulky in comparison. As for other missing reports? The best option seems to be building custom GA4 reports in Data Studio.
If you were to compare UA sessions to GA4 sessions, you will notice a difference and that they do not match up. You might wonder why sessions are lower in GA4? The way Hits are processed can make sessions appear lower. Google also explains why that is by saying:
“some aspects of session counting differ between the two platforms. For instance, in Universal Analytics, a new campaign will start a new session regardless of activity, however, a new campaign does not begin a new session in Google Analytics 4. This may lead to lower session counts in your Google Analytics 4 reports.”
So should you make the switch to GA4?
Right now the best option is to Run UA and GA4 together. This can be done rather easily with Google Tag Manager. We recommend this as you can't go back in time with recording analytical data, and you can't really import from UA to GA4. So (if you haven’t already) start collecting data for GA4 now and take the time to learn about GA4 and what you can do with it. But keep Universal Analytics and your reporting until you are fully ready to make the switch.
Final thoughts
This is just some of the changes in GA4, with more likely to come before 1st July 2023. As it stands, there are issues and limitations. But we're optimistic that the good things like the the events based modelling and the focus on engaged sessions will give us a more accurate view of performance. If you haven’t had time to get your head around this big change, need help with the transition or just want to pick our brains then get in touch.
