GA4 also marks a fundamental change in how data is collected and processed. For example, the default data retention periods will be two or 14 months which will limit businesses' ability to “look back” over the periods of time they’ve been used to with UA. This helps Google’s ability to handle vast swathes of data, but is a hugely significant change for analytics users.
Marketers may feel frustrated by the changes, but it is a reminder that it is Google’s ‘ball’. They always have and always will control what data we see on the free platform.
Why is GA4 happening?
Sunsetting a popular product is never easy.
But platforms need to evolve and respond to new challenges. The new version of Analytics has been in operation for 18 months, as the ‘App + Web’ feature and so we’ve been testing its iterations to date as Google rolled out the product. If your analytics property was created after 14 October, 2020, it is likely you have been using GA4 anyway, as GA4 was promoted as the default option. However, it was still possible to create a UA property or run both alongside each other.
According to Russell Ketchum, Director of Product Management at Google Analytics, GA4 was primarily introduced ‘to understand the complex, multi-platform journeys of their customers – all while prioritising user privacy’. Those are the two main drivers behind GA4 – a different user tracking model and the need to adapt to a cookie-conscious online environment.
One of the biggest talking points about this change and this iteration of Analytics are the implications it has for legacy data. Previously Google has provided long term support for legacy GA platforms, and has not restricted access to historical data collected on an old methodology. It is clear that won’t be the case with the move to GA4. We have some solutions later in this article.
July 1, 2023, is the key date for your diary. That is when data collection for standard Universal Analytics properties will cease. An extension has been granted for Universal Analytics 360 properties until October 1, 2023, during which time new tracking data will still be processed.
What else has changed with GA4?
Goodbye sessions, hello events.
Google Analytics 4 is event and user-based in its tracking model, rather than session-based. This is to reflect the fact that conversions are more likely to take place across multiple touchpoints, rather than a standalone desktop or mobile session. Google also believes user behaviour data is more meaningful than that which comes solely from page or URL-based activity.
There is also the added bonus that event tracking (where you can track things such as downloads, or clicks on a form) does not require any new code or additions in Google Tag Manager. Some basic events like ‘first_visit’ or session_start’ are automated. In GA4, there is a wider suite of ‘Enhanced Measurement’ event tracking metrics which can be turned on easily from within the ‘All Events’ dashboard.
The potential for event tracking has increased beyond anything possible in previous iterations. In UA there was a standard template to follow and populate – event ‘category’, ‘action’ and ‘label’. In GA4 these restrictions no longer exist. For example, 25 additional parameters can be sent to GA4 with each event, as opposed to four in UA.
A key point to bear in mind is that structure is hugely important. Whereas in UA the structure was predefined, in GA4 it needs to be put in place before tracking is set up and then followed for all future tracking. This will help reporting and optimisation of parameter name limits.
Goals are now conversions.
There are now only ‘conversions’ in GA4. ‘Goals’ are no more. Other changes you will notice from the UI is the reconfigured homepage, streamlined navigation and new report structure. Instead of the five top level Reports (Real-time, Audience, Acquisition, Behaviour and Conversions) and their collapsible sub-menus, Reports now works with a widget-based design. These are called ‘scorecards’ or summary cards.
Some metrics have gone or are accessed differently. Bounce rate and average pages per session no longer exist (but engagement rate does). Site speed is not collected by default, but requires specific tracking. Landing page traffic data is now accessed via the Reports snapshot and ‘View pages and screens’ widget. Purchase data is via the Reports snapshot, ‘View Conversions’ widget and purchase event name.
Data for a website and app can now be tracked together in GA4, rather than in two separate properties. The same tracking code is added into a single property and creates one data stream through desktop, tablet and app activity.
The move to property-based reporting requires a change in approach to how data is viewed. It is no longer possible to use multiple UA views within a property in order to segment data at a high level. You can add multiple data streams in a property, so GA4 properties make roll-up reporting easier.
If you're a brand with multiple digital platforms, this new reporting approach should make it simpler to understand the user journey between them.
Machine Learning and ‘Predictive Insights’.
This is the first major iteration of Analytics after GDPR and the rise of other privacy legislation such as California’s Consumer Privacy Act (CCPA) or Brazil’s LGPD. These initiatives, and the rise of online privacy advocacy, mean that GA4 has to reflect that a significant amount of data will not be available because of greater user control over cookies, or other privacy settings such as Apple’s IOS changes. So, Google Analytics has turned to machine learning to effectively fill the gap of the missing data with insights about user behaviour. It creates patterns from users with similar behaviour, and can generate ‘predictive insights’ about datasets, enabling marketers to create forecasts based on these predictions.
During our testing of GA4 over the previous 18 months, it’s clear you will notice that session traffic will be lower in GA4 – that is due to a change in session attribution. A user journey within a 30 minute period will no longer be split into different sessions: for example, if a user entered on organic, left, then entered again via social, these will now be treated as one organic session. We see this affecting clients in some industries more than others, such as ecommerce with short purchase windows seeing larger reductions.
In Universal Analytics there was a limit of 10 million ‘hits’ per month, per property. GA4 has done away with hit limits and increased the flexibility of data capable of being tracked. Sampling also only occurs in GA4 when the data in advanced reporting exceeds 10 million hits.