Noisy Little Monkey’s rundown
The audience is frightened of GA4. They fear it. They tremble at the thought of a loss of their marketing insights, and at the prospect of a cold and heartless Google expecting them to learn a completely new and dizzyingly unintuitive tool on top of—and in order to continue to effectively do—their jobs. And they do need those jobs. They need them to pay for food, and heat, and in some cases, to find a sense of worth in a thankless, uncaring universe.
Enter Jill Quick, who actually knows how the thing works. She could, you imagine, probably solve all of your problems for you, if you could just find a way to get her in front of your laptop.
Here are my takeaways from Jill’s talk.
Google Analytics 4 is fundamentally different from Universal Analytics (now considered “GA3”)
Where some previous Google Analytics sequels have taken the same basic codebase and built from there, GA4 is built from Google’s Firebase product. The underlying engine is different, and it arrives at its numbers in a different way.
So far, so good. So what?
For marketing professionals, this has a number of ramifications. Wherever you are able to get the thing to cough up the same metric as before, the actual numbers will probably be different, having been gathered and processed in a different way. Cue an apoplectic big cheese behind an oversized, real mahogany desk, frothing at the mouth, cheeks puce. What do you mean the traffic numbers are lower this month?!
In other cases, the metric you relied on in stakeholder reports can no longer even be found.
Rather than slavishly recreating your comfy old reports in the new Google Analytics, you might need to step back and evaluate what questions you are trying to answer with your reporting in the first place. What were the numbers supposed to show us exactly? According to Jill, once you are armed with the right questions, you can start to understand what answers might look like under the new regime.
Forget Sessions and think about Users
To be fair, this one has been coming for a while. Single session metrics ignore the wider circumstances of a site visit, and GA4 has a notable lack of emphasis on sessions. GA4 is mostly about the user, and understanding their experience as a persistent, multi-session thing.
Events are not special
In Universal Analytics, events were something special that you went out of your way to measure. You had a 4 tier system for assigning them with categories, actions, labels and values. It was a simpler time.
In GA4, all measurements are conducted via events. Pageviews are events. So are scroll depth triggers. So are form submissions, video plays, pauses, rewinds and completions. Want to know how often somebody downloads 4 PDF documents on the fifteenth Sunday after Pentecost? That’s going to be an event triggered by another event. Meta.
Unlike the events we once knew, GA4 events have up to 25 parameters, and they are simply key-value pairs that can represent anything you need them to. An “add-to-cart” event might include parameters like “name” or “price”; a “download” event might include parameters like “filename” or “download_page”.
GA4 is not as “plug and play” as previous versions. You’ll have to build everything out yourself
Universal Analytics had a lot of reports about all kinds of stuff. Some of it you probably never used, but it was nice to know it was there. GA4 has a handful of automatically collected measurements. Everything else you will need to build out yourself. This extends to reports, as well as the aforementioned events.
Google categorises its new GA4 events in the following deliciously tautological way:
Automatically collected events:
- Automatically collected events
- Enhanced measurement events
Events that require implementation:
- Recommended events
- Custom events.
Want an event to be triggered by successful form submission? You’ll have to create that event yourself. Sort of. Confusingly, Google has a bunch of predefined event names (and parameters to go with them) for various common purposes which they recommend that you use, but will require some implementation to do so, and you may have to do some legwork to dig out the document that lists out all the predefined event names. They are in snake_case, and deviating from that is not the done thing.
All this talk of deciding what events you need to have under GA4 brings us back to a previous key point and allows us to go further;
- What questions are you trying to answer with your reporting?
- What do you need to measure in order to answer them satisfactorily?
- What events will you need to measure those things?
- Is there a pre-defined event for this purpose that will work for you, or will you need to create and implement a custom event?
You’re going to need to budget some time to prepare your business to use GA4
There’s no getting away from it. The new Google Analytics is a powerful technical application and the intuition and muscle memory you may have with previous versions will not get you far. If your business wants to continue to use Google’s Analytics products, you will need to make a serious investment of time, not just to redevelop your own skills to produce useful reporting insights, but to scope and implement an instance of GA4 that effectively serves your business.
You wouldn’t expect to learn the ropes of another type of professional software overnight. How long would you expect to spend in order to achieve competency with a piece of CAD software, professional video editing software, digital audio workstation software, or an Apache helicopter flight simulator?
The alternative is to use another Analytics package. Jill listed off a number of them, including Plausible and Matomo. If these better suit your needs, they may well be the way to go.
Final thoughts
Google Analytics 4 boasts more powerful features than previous iterations, most of which I haven’t had space to cover in this round-up, including some of Jill’s favourites. Powerful segmentation tools that you can use to create Audiences and cohorts, flexible report building based on custom measurement metrics. GA4 offers surgical precision to answer specific questions about your marketing and website performance in a way that the previous version couldn’t.
The downside is that Google expects us to be power users in order to use the thing. It isn’t plug and play. It isn’t ready to go out the box. It will require proper thought and a good migration plan, and that isn’t for everyone.
