Preparing for Google Analytics 4.
In March, Google dropped a data bombshell by announcing it will be removing all previous analytic formats such as Universal Analytics (UA). From June 2023, for most organisations, the new Google Analytics 4 (GA4) will be the only viable interface with which to continue collecting data.
To rub salt into the wound, Google then followed up by announcing it will be deleting all data predating GA4 from Google Analytics in January 2024.
Focusing on the more immediate issue, we’re now less than 12 months away from when Universal Analytics will stop recording data. The clock is ticking.
At this stage we are advising to get GA4 up and running asap and also keep it running in parallel with Universal Analytics. This way there will be plenty of comparable data between the two, before UA is turned off.
Have you started preparing for the move? Is Google Analytics 4 still fairly new to you? In this blog, we’re unpacking exactly what GA4 is, how it differs from previous interfaces, and our tips to successfully make the move without jeopardising any data during the migration process.
A little side note: this article covers setting up from a more advanced, technical perspective. This will most likely beneficial for those in roles that involve managing websites, analytics such as analysts and web managers.
What is Google Analytics 4?
Interestingly, GA4 isn’t a completely new interface but a newer iteration of what was initially called ‘Google Analytics web+app’. The drive behind this new interface originated from the need for Data Analysts and Digital Marketers to view data in one central port.
Users typically move between apps and websites as they interact with various brand touchpoints, meaning data is created across many different mediums which can be difficult to capture in its entirety, hence the need for a more holistic interface.
With a growing focus on privacy in the industry, Google sought to develop more complex systems for ‘cookieless’ measurement, and behavioural and conversion modelling.
In Google’s own words, “GA4 is designed for the future of measurement”.
How is this different from previous versions?
This ‘app data’ focus really separates the way GA4 works from ‘Universal Analytics’ (UA). At its core Google Analytics collected data through sessions and pageviews. But apps don’t have pages, and people use them in very different ways to a typical ‘session’ on a website.
So this is where the key difference comes in, GA4 records everything as an event. Event-based tracking allows for greater insights to be derived about users and their interactions. Admittedly, as experienced users of Universal Analytics, we’ve found this to be the hardest part to adjust to due to familiarity.
The move to event-based tracking allows GA to automatically track the majority of engagement events marketers have been used to manually setting up themselves. Now with the click of a button, marketers can automatically track ‘automatic enhancements’ such as: scroll tracking, outbound links, site search tracking, video engagements and file downloads.
- Goals are No More
In what feels like a move to better align language, ‘Goals’ are no more in GA4, ‘Conversions’ will replace them. The process for making conversions has also been simplified in comparison to how you would have previously setup a ‘Goal’. Now you will easily be able to turn an event into a conversion, without having to remember the exact ‘event label’ and ‘event category’ you have used! The move to event-based tracking does mean that destination url goals will be confined to history, and not make the port into conversions. These goals will need to be switched over to events when ported across into your new property.
- A New Interface
The new analytics format also brings with it a new User Interface (UI). This replaces the old interface more visually inline with some of Google’s other products. This feels like an underlying theme in the more ‘front-end’ heavy changes you will experience with GA4, bringing one of Google’s old products inline with its growing product range.
There are changes to the default data retention period, shortening from effectively infinite retention to 2 months by default. This only affects user-level data (associated with cookies and advertising identifiers) so won’t impact basic reports, but will limit data reporting for any custom reports in the ‘Explore’ section. This change will likely see a reasonable difference when comparing repeat visitor reports between the two analytics types. Something to keep an eye out for.
GA4 also sees the ‘views’ function being removed. At a property level, you now add each website and app as a data stream. All settings that you would previously have set at a view level are now either property-level (IP filtering, conversions etc.) or view report-level filtering (domains etc.).
How to make the move:
Over time you will want to develop your use of more specific GA4 features but in the meantime the priority should be getting data collected, and in a way which is readily usable for fellow members of the marketing team and your organisation.
Here we detail our process that we have been using with our clients. But it is worth noting that our process includes two assumptions:
- You have an existing Google Analytics account (using Google Analytics Universal Analytics).
- You are using Google Tag Manager on your website to trigger your Google Analytics.
On this basis we have a 5 step process:
- Audit your existing Google Analytics data and goal setup
- You have the opportunity to start from scratch without any legacy issues. So this means you can leave old views and goals behind.
- When auditing the goals our checks cover 4 key elements:
- Is it still relevant to you?
- Is it working correctly?
- Is it recording data?
- Is it transferable to GA4?
- Get a second opinion. Before you decide to leave a goal behind, just make sure no one else is currently using this in their reporting.
This gives you a thorough understanding of what your data recording situation is and the scale of work needed for your migration to GA4.
- Now you can create your new GA4 property
- Google Analytics has a great wizard to help you (at a top-level) create a new GA4 property from your existing Universal Analytics property.
- Create new data streams for each website or app that you will be using the new property for. You will find, for each data stream you get a new set of pretty useful settings, as well as extra reporting uses.
- Make sure you match up some key settings for each data stream such as IP filters, with your corresponding settings in UA. So you can keep the data as actionable as possible without diluting with internal traffic sources (like employee site visits).
- This is also the place to enable automatic enhancements. Which we would definitely recommend doing (for beneficial reasons mentioned above in the events section)
- Then head on over to Google Tag Manager, and enable your new GA4 configuration tag.
- Copy across your new measurement ID and enable this tag to fire a pageview.
- Then utilise the same triggers as you currently use on your existing Universal Analytics pageview tag.
- Do make sure to use the exact same triggers as you currently use, including any connected to your cookie control management.
- Data streams can take 24 hours to start showing data coming in so you will need to wait a day (or two) to check the data is coming in accurately.
- Now that data is coming into the property, head over to the events report and check how many of your old goals are being automatically reported by GA4’s new events report.
- For any events which are not yet being pulled through, you will need to create new GA4 Event tag in Google Tag Manager.
- For the new GA4 tags, you just need to mirror the existing UA tags, but with the new GA4 Events tags. This means, utilising the same triggers (including any cookie consent requirements).
- We always recommend previewing and testing those events before publishing the tags on site – just in case.
- Then as before, wait around a day to see if that date is now pulling into the event report.
Once your old goals are pulling into the event report it’s time to ‘upgrade’ some of those organisationally important ones into conversions.
If you are currently using ecommerce tracking through Google Tag Manager then you may be able to port across to GA4 with limited technical support. GA4 ecommerce utilises events too, in specific, any event which is named ‘purchase’ GA4 will deem as Ecommerce and pull that data into its ecommerce report.
GA4 does come with more sophisticated ecommerce tracking, as standard it is similar to UA’s enhanced ecommerce. Currently you can’t utilise GA4 ecommerce (fully) and UA ecommerce, at the same time and we definitely wouldn’t suggest binning your UA ecommerce yet. As a compromise, you can gain the same data as standard GA4 ecommerce into GA4 through a couple of custom variables. Truthfully this will be enough for most people, for now.
- GA4 in tag manager utilises the ‘data layer’s’ “purchase event” push. So you will need to adjust your trigger to utilise a ‘purchase’ event.
- You then need to create some new custom variables in order to translate your existing datalayer into the relevant information for GA4. Then pull them into a new GA4 tag, so they look like this:
- In our experience the currency code is necessary to feed that data into the GA4, although (at the time of writing) this is currently missing from Google’s own support pages. In our tests, without it the value data fails to push to GA4.
- As always, you should then go into preview mode and run a test donation to check that the new ‘parameter names’ contain the relevant information.
This should cover the essentials for you right now. As you go about the port to GA4 we recommend utilising preview mode on Google Tag Manager as much as possible, to ensure that you can see any issues prior to publishing.
If you do encounter any issues or want to talk through getting some support on your migration to GA4 drop us a message, we’d be happy to answer any questions, chat through the process, or see how else we can help.
Share this article: