It’s Here: GA4 FAQs and Guidelines
Google Analytics 4 marks a significant milestone in the evolution of web analytics. Google has marketed its newest iteration as the answer to “major shifts in consumer behavior and privacy-driven changes”. GA4 is AI-driven and uses machine learning to automatically alert you to trends and changes in data, unlocks powerful new reporting and integrations, and helps bridge data between mobile and desktop devices by launching the App + Web measurement model into primetime – all in the pursuit of improving your marketing ROI.
This sounds well and good, but what does it mean for you and your business? Since the GA4 announcement in October, we’ve been working closely with our clients to implement GA4 and have put together this list of their most frequent and important questions. We hope this FAQ will help you navigate the changes coming to Google Analytics, but if you have any other questions, feel free to reach out.
GA4 is a far-reaching paradigm change and a big topic, so keep an eye out and as we share more about it over the coming weeks!
What is GA4?
GA4 is basically the new, updated version of Google Analytics. It’s a REALLY big and exciting change in the UI, analytical features, and the underlying data model compared to prior versions. While most sites still use the standard “Universal Analytics” flavor of GA that has been live and evolving for almost 10 years now, you might notice that GA4 is a direct evolution of a number of recent releases. The App + Web properties that first emerged in beta in 2019, Firebase Analytics, Google Signals, and GA’s 2017 gtag.js tagging framework are all apparent in GA4’s DNA.
It’s not an option, an addendum, or a different platform – this is the new Google Analytics. You can still use your old properties for now, but moving forward, the new GA4 version is what will be actively upgraded, developed, and maintained as a part of Google’s Marketing Platform product offering. The existing property types we’re using today are being replaced, and eventually everyone will have to make the transition, across the board, to stay current.
What is the benefit to using GA4?
You can think of it a bit like upgrading from a 2012 car to a 2020 car – it’s still a car, it’ll still do all the basic car things you need… but it will also add some shiny new features, meet new fuel economy standards, and maybe even offer self-driving in some circumstances!
The GA4 release comes with some very cool new features out of the gate, and even more has been revealed on its product roadmap to come. The official line on the main updates is:
- Smarter insights that use machine learning to identify trends.
- Deeper integration with Google Ads.
- Customer-lifecycle-centric reporting.
- More granular data controls.
For those who haven’t yet taken advantage of App + Web properties, the integration of web and app data is big too!
In practice, you’ll see a massively revised user interface with new analysis features, more ML-powered insights generated (ex. Forecasting surging demand for a product you sell so you can shift advertising budget and/or supply chain resources appropriately or predicting churn rate of leads at every stage in your defined user journeys so you can target tests and optimizations accordingly) and better Ads/YouTube data sharing.
On the technical end, some basic event tracking will be done automatically (though custom events will still be in high demand), interacting with consent/data deletion will be easier and we’ve been told to expect some sophisticated features to compensate for data sparsity in the near future (ex. a drop in measured traffic from implementing GDPR consent/opt-in). This will be a critical source of competitive advantage as the technology landscape matures, and it’s clear that Google is moving aggressively to address privacy and data issues with recent announcements like Federated Learning of Cohorts in the Ads ecosystem.
How does GA4 differ from prior versions of Google Analytics?
You should think about GA4 as the new “regular” Google Analytics. We’ll create properties, implement data collection, and use the reporting UI a little differently, but it’s not a different product option. Most users will primarily notice the significantly revised user interface and new feature additions mentioned above, but analytics specialists and developers will need to adapt to updated JS libraries, data models, and implementation approaches as well.
There are many hours worth of fine points to understand and absorb (particularly on the technical end) to be an expert, but an overview of the different data models underlying UA and GA4 can be found here: https://support.google.com/analytics/answer/9964640?hl=en
I have GA360: how is GA4 different? Do I need both GA4 and GA360?
If you need GA360’s enhanced features and data limits now, you’ll continue to need them – you’ll just be migrating to the new features and implementations of this release over time. If you have GA360, it would still be valuable to start collecting data and getting familiar with GA4, but be aware that Google hasn’t fully integrated these changes with the GA360 product yet. However, they have indicated that new GA4 updates will be coming soon. SLAs and more advanced integrations with tools like BigQuery have been hinted at.
What happens to my existing GA properties and views?
For now, old properties and their constituent views will remain safely in place, and we would recommend maintaining them in parallel with new GA4 properties for at least 6 months for the sake of stack stability, user transition, and allowing cross-comparison in the historical data record. To the last point, you’ll shut off data collection to the old profiles eventually but even then will want to keep them for historical data comparison.
In GA4, there is actually no such thing as a view (formerly referred to as a Profile in classic GA); the data model and collection routes Data Streams into a unified property view that you can break down in reporting as needed.
It’s possible Google may force-upgrade old properties and views in future as it has in other cases, but no announcements have been made to this effect so far.
How do I upgrade?
It’s mostly a pretty simple process for the back-end basics in terms of creating a new property and tag! Google makes it easy with a built-in setup assistant to upgrade sites with existing GA implementations. Bear in mind, however, that additional custom events, eCommerce implementations, and so on may take a bit more work.
Making the transition as a reporting user will just be a matter of learning the new UI – but don’t underestimate the lift here! Many go to metrics like bounces are gone or substituted, and you’ll be learning to replace segments with “Comparisons”, custom reports with the Analysis Toolkit, and so on.
If you’re seeking assistance in upgrading to GA4, please reach out to our Wheelhouse Analytics team at any time.
Is there a cost?
Nope! Standard GA will remain free to use, and GA360 will remain a paid product. GA4 does offer the paid option to use custom Server-Side data collection and configure raw data exports to a BigQuery data warehouse (which we’ve done to great success for several of our clients).
Will it be more difficult to use?
In the short term, you can expect a bit of a learning curve in adapting to the new UI and changed/added features, but ultimately the goal is to create a more capable, efficient user experience. The stronger your understanding of analytics fundamentals the easier the transition will be!
Just about everything is at least a little different in GA4, from the reporting interface and functionality (ex. custom reports and segments are replaced by an evolution of the beta Analysis editor and Comparison applications to reports) to the data schema and implementation methods. Some changes could be a bit of a mind-bender for more traditional marketers who aren’t native to data-driven approaches or less analytics-savvy users that are used to logging in for just a few conventional metrics (ex. the misleading and long maligned Bounce Rate is gone – replaced by ‘engaged sessions’).
As a metaphoric example, if you just know a couple travel-book phrases of Spanish, learning Italian will be just as difficult as starting from scratch, but if you’re near fluent in Spanish, picking up a facility in Italian will come fairly easily (extremely similar grammatic structures, etc) and be more an experience of adjusting your automatic responses and habits that are different in the new context.
Who is GA4 for?
Everyone! GA4 is the new standard and likely will be implemented for all users and sites within a year or two (new accounts already default to it). Think of it like an operating system update – you may want to hold off for a little while to make sure you have all the compatibility you need and to see if there are any bugs yet to be worked out, but you have to switch over eventually to avoid getting left in the dust.
Google hasn’t discussed a formal sunset for Universal Analytics (and it’s not uncommon to see websites still operating on analytics.js over three years after the move to gtag.js), but we recommend spinning up a new GA4 property as soon as is practical for everyone – the more time you have to compare side-by-side and adjust without pressure the better!
Who shouldn’t use GA4? (i.e. are there certain industries/sectors/types of business models for which GA4 might be counterproductive?)
No one – it shouldn’t be considered optional! That said, there is the question of when to get started with GA4 – and a lot of flexibility there for now if you need it. eCommerce implementations will change pretty substantially and appear to have some upcoming updates in queue. In light of this, we recommend that eComm businesses begin to work on understanding and planning for GA4, but hold off on transitioning to a production implementation until current foibles related to eCommerce tracking in GA4 are resolved.
How does GA4 address GDPR/CCPA & other privacy laws?
The new version of GA is built with GDPR and CCPA in mind and offers a number of options and controls to better facilitate actions like data deletion. However, it is important to bear in mind that GDPR, CCPA/CRPA are heavily focused on governing data collection. GA4 provides features and settings to handle user requests and data handling for compliance once it’s in the platform, but you’ll still need to implement consent management and linked controls on when and how data is sent to GA (or any other platform) on the front-end.
Do all of my goals & events migrate to GA4?
There’s some setup to be done and conversions/events do work a little differently, but at the end of the day you’ll still be able to get all the same interaction and conversion information you do today.
Is there a minimum amount of traffic I need to have for GA4’s machine learning functions to be valuable/statistically significant?
As long as you’re getting at least a couple thousand sessions per month, you should see anomaly detection notifications and insights from the behind-the-scenes ML machinery in GA (in GA4 or on the current version). Bear in mind also that Google (and the market in general) increasingly favors Bayesian approaches to statistical inference as with Google Optimize – the traditional concept of significance and problematic alpha levels and p-values isn’t even applicable to these types of models. While most “machine learning” or “AI” features in web analytics or CRO products today are pretty user friendly for laypeople, if you’re not comfortable with the distinction between a Bayes factor and a frequentist alpha/p-value, you could definitely be getting more out of your data (or at least more effective insights from what your tools are actually telling you) by reaching out to an analytics specialist!
Wheelhouse Is Here to Help!
If you have any further questions about GA4, let us know in the comments: we’ll do our best to get back to you ASAP, and may even update the FAQ with an answer. You can also reach out any time if you’re interested in having our Analytics team help with GA4 or with any other tracking, tagging, reporting, or data analysis needs your team is currently facing.