Introduction to GA4: Everything you need to know so far

Google released a major update to their Web Analytics product in October 2020 called Google Analytics Version 4, or more catchily GA4. Do not be fooled by the unassuming, incremental name, this is a big change to one of the most popular web analytics platforms. In this blog, we will outline the updates that have been made and the sensible next steps you need to consider to be in the best position to utilise all the new features.

Why the new version now?

A lot has changed in the web space since GA was first rebranded from Urchin. Web sites are no longer a collection of pages, but are now described as web apps with journeys. As such, the old model of a pageview followed by another pageview scattered with a handful of events, is no longer fit for purpose. The powers that be at Google Analytics saw this coming a mile away and decided to act! Although, not before we all had to suffer the pain of measuring a pageview in a single page application. Mobile App analytics was another fun challenge, with older versions of GA being particularly difficult when trying to splice together App and Website analytics in a single view.

Google is also trying to reassure you that the scary third party cookie-less world (one that Google themselves are pushing for) won’t be so scary after all, well at least for analytics. Their answer to solving this problem is one they’ve borrowed from last decade, Artificial Intelligence.

What changed

The headline change in GA4 is the introduction of a brand new property type, unimaginatively called Google Analytics 4. This new property utilises the new event driven data model which has been in development since the release of Firebase analytics and later, the Web + App beta property. The GA4 Data model of events and parameters provides a more flexible and scalable method of tracking. Within this new property, there will be Data Streams, as opposed to views. A Data Stream represents a single source of data collection for a property: this can be a website, or an iOS or Android app. A property can be configured to have multiple data streams, allowing for web and mobile data to be easily captured within the same property.

GA4 leverages the many years of Google customers signed into their products on different devices, alongside their vast experience in targeted/personalised advertising networks, to provide statistical methods for more accurately identifying users. Something that has previously been considered a bit of a dark art. Google Signals, the name of this wizardry, will enable your organisation to utilise more accurate cross-device, Demographics and Interest reports. This wizardry does come with a few read-the-smallprint details that are important to familiarise yourself with, and that could have an impact on existing Privacy Consent policies. You can find these details here.

One particularly important note is that the Google Signals feature has major privacy implications, and it is important to ensure that customer privacy policies are updated and compliant with all relevant legislation before this feature is enabled.

Google Analytics 4 now also offers to plug the gaps left by your more privacy focused customers, by using Machine Learning to model the estimated behaviour of users who would otherwise prefer not to be analysed. Google claims to also use Machine Learning to identify important trends, providing predictive metrics such as revenue projections. It will be interesting to see how useful this feature ends up being. Something we’ll look to investigate in the future.

Enhanced Measurement, which is an out-of-the-box feature from GA4, provides a wide variety of tracking that can be enabled without the need for any code or tagging changes. Providing lots of new features, with minimal changes to the implementation. Who doesn’t want to do less, for more?

Is GA4 for you?

Well, if you want to continue using Google Analytics, then there is only one answer to this question. However, if you are thinking of jumping to GA from another Analytics provider, or vice-versa, then the release of GA4 provides an excellent decision point to consider a change, due to the organisational planning required to best utilise the GA4 data model.

If you’re new to GA, the new features included in the What changed section of this blog is a great start to understanding what is available to you, and the direction in which Google Analytics is going. If the new features and the changes required does not fit your organisation’s future plans, then it may be time to look for pastures greener.

So you’ve decided GA4 is for you. As previously mentioned, it’s not a straightforward change. There is a big difference in the data model between Universal Analytics and GA4. Whilst most UA features have an equivalent in GA4, not all will. Here we list examples of what you’ll need to consider: 

  1. If you’re an enterprise customer, using GA360, you’ll have to wait for GA4 to be fully supported within your 360 agreements and SLAs before it’s advised for you to use GA4 as your sole analytics measurement tool. Nothing stopping you from starting to plan out the upgrade process, however. Especially when your marketing friend at Company X is raving about how well Google Signals is identifying their customers.
  2. GA4 and UA properties are fundamentally different, and whilst a UA property can be upgraded to a GA4 property, there will be things missing. If there are UA specific features that your organisation relies on that are no longer available in GA4, then a rethink in your measurement framework will be required.
  3. Universal Analytics isn’t going anywhere yet. Google has not shared an end-of-life date for support, but it’s fairly clear that all there will only be more bells and whistles added to the shiny new GA4, with universal analytics being left to gather dust.
Migrating from Universal Analytics to GA4

If your organisation is looking to upgrade to GA4, Google’s recommendation is to set up GA4 tracking in parallel with existing Universal Analytics tracking and to run both for an extended period. The advantages to this approach are:

● Existing tracking and reports will be unaffected and can be used while the GA4 implementation is tested and validated
● The dual run period can be used to gather experience and training with the new reporting and data model
● A period of historical data can be built up in the GA4 property in preparation for eventual switchover.

Pressing the Upgrade to GA4 button will upgrade an existing Universal Analytic property to a GA4 property. Upgrade has been italicised, because the UA property is not actually being modified in any way, instead a new GA4 property is created from the UA property. The new GA4 property will copy over everything from the UA property, except unfortunately for the data. That would’ve been too easy. Which is why it’s recommended to get a GA4 property setup and running ASAP, so when the transition does inevitably happen, you will have that all important historical data.

Analysts and report end-users are likely to notice many differences between GA4 and UA. To help ensure organisational buy-in to the new approach, organisations should plan for suitable training and knowledge sharing to understand and quantify the differences.

Privacy and Data Controls

GA4 does not offer additional privacy features or data controls over and above what is available with Universal Analytics.

One important change to keep in mind is that IP address anonymisation is always enabled for GA4 properties.

Aside from the above, existing privacy controls continue to be supported:

● Linking to Google Signals can be enabled or disabled for individual properties, and specific regions can be excluded.
● Data Retention period can be customised per property to control how long data should be stored by Google.
● Methods to opt out of data collection by end-users are available including the Analytics Opt- Out browser extension.
● Data sharing with Google support teams can be configured on a property basis.

As mentioned previously, Google Signals has major privacy implications, and it is important to ensure that customer privacy policies are updated and compliant with all relevant legislation before this feature is enabled.

Google provides information about the specific terms that must be included here.

Tips and Considerations
  1. GA4 includes some events with reserved names that are predefined. Using these will not impact your allocation of 500 events.
  2. GTM Container ID is now sent to GA4 in the network request
  3. Channel settings have been removed from GA4
  4. Referral Exclusions are not currently available
  5. Cross-domain tracking is now setup in GA4
  6. Measurement Protocol API can send extra events into the property with API secrets
  7. BigQuery linking is free with Hit Level data
  8. Standard reports aren’t sampled
  9. Custom reports may (but not always) be sampled
  10. GA4’s frontend is more like Firebase, compared to Universal Analytics
  11. Event names are generally snake_case
  12. If event names are misnamed, the incorrect event name will be kept. Be careful with spelling and formatting to avoid confusion.
  13. Landing Pages Report is missing
  14. Custom reports look more powerful with analysis explore – akin to Adobe Workspace
  15. Events take 24 hours to be processed before they are available in reporting
  16. Parameters of events need to be defined as dimensions or metrics, but not the event themselves
  17. DebugView is not turned on by default. The appropriate field name needs to be set to ‘true’ in the GA4 tag, which then turns on the DebugView for all devices. This setup only makes sense on a staging site, where it is possible to identify all devices that are sending through debug data.
  18. Tag Assistant can be used to turn on DebugView, which is useful when testing

What we lost:

● Views / Filtered views
● Channel groupings
● Referral exclusions
● Old behavioural metrics: Bounce rate, pages / session, time on site

What we gained:

● Debug view
● Analysis Hub
● New engagement reports
● Free BigQuery integrations
● Proper app and web reporting solutions
● A method for viewing unsampled data – either through standard reports or BigQuery for custom reports

Closing Thoughts

Google looks to have produced an update for GA4 with current and future evolutions of the web in mind. If you’d like any help or advice in making this transition, please reach out to info@lynchpin.com for a bespoke upgrade consultation.

Last reviewed: 04/02/2021