Goodbye Universal Analytics: Hello Google Analytics 4

Anyone spending even a moderate amount of time in Google Analytics will have seen the little pale-blue banner at the top of their dashboard, offering a ‘friendly’ message that Universal Analytics – is moving…

That’s right, as of July 1 next year, your trusty Universal Analytics property will cease to exist. To be clear, you will lose historic data after July 1 if it isn’t stored outside of Analytics.

In order to continue tracking (and analysing) what your customers do on-site, you’re going to need a brand new, extra shiny, somewhat bewildering Google Analytics 4 property.

To be fair to Google, they have been hinting not-so indelicately about GA4 since October 2020, but unless you really love data or were already unhappy with the tried-and-tested UA model, you may not have paid much attention to the newest kid on the block.

Maybe you’ve dipped your toe in, then immediately recoiled, slipping your foot back into your comfy, familiar, UA-branded slippers. Say again? They’re doing away with session-based tracking?! I need to sit quietly for a moment.

via GIPHY

Yes, there are some hefty changes to GA4 and it is an adjustment from what we’re all used to. But in all seriousness, once you get your head round those changes, it actually offers a lot more than Universal ever has.

In this article we’re going to break down the whether-tos, the why-fors and the how-tos, so by the end of this shit, you’ll be a total convert. So strap in, it’s about to get analytical.

via GIPHY

What, how and why? GA4 at a glance

Not to sound like the herald of woe, but the shift to GA4 comes with a myriad of changes for marketers and businesses to bend their minds around.

And whilst that may seem like an inconvenience when everything was ticking along nicely with traditional analytics, Google claims to have good reason and our best interests at heart. Bless.

via GIPHY

In their initial announcement, Google offered up a study that detailed how marketers currently struggle to gather actionable insights from their data and achieve a complete view of their customers.

GA4 is aimed at addressing this by providing us with the data we need to really move the needle on ROI, particularly in the long term.

So, what’s changing for analytics in GA4?

The main and most obvious difference between GA4 and its predecessor, is how users are tracked. Traditionally, users are tracked via sessions. Set time periods that collate everything a user does whilst on-site.

GA4 leverages event-based tracking. Rather than creating a new session when a user returns to your site, GA4 continues to track all their actions (or events) that they complete. Google defines this shift as follows:

The session counts in your Google Analytics 4 property may be lower than the session counts in your Universal Analytics property. This is because Google Analytics 4 does not create a new session when the campaign source changes mid session, while Universal Analytics does create a new session under that circumstance.

Brass tax? If a user ends their session by leaving, GA4 doesn’t start a new session when they return.

via GIPHY

UA creates a new session when the same person returns, whether through an ad campaign, through organic or from any other channel. GA4 doesn’t count a shift in source as a new session directive.

How will this change affect your data?

Event-based tracking actually allows for far superior accuracy than session or time-based.

Rather than focussing on just getting users on-site and clocking a visit, event-based will significantly reduce the risk of duplicated users, now that a new session doesn’t start with a campaign or source shift.

GA4 boasts an extremely high accuracy and low error rate when it comes down to user tracking (typically less than 2%). And as the emphasis is on measuring what your users are doing on-site, gleaning the useful insights to positively impact return on investment will undoubtedly be easier.

Because GA4 is so accurate, deduplication of users will mean that you’ll likely see a drop in comparative users between your traditional UA and Google Analytics 4.

Accepting this, not getting hung up on the pure amount of visits and utilising the buckets of additional data (there’s a lot, seriously) that GA4 can produce, will be the biggest challenge for some users. But the sooner you do, and the sooner you adapt – the richer you’ll be for it.

Why GA4? WHY?

It’s clear then, the exodus to GA4 marks a significant change in the industry as a whole. But aside from greater accuracy, why has Analytics 4.0 come about?

The main reason Google has really shaken things up and laid the groundwork for a future-proof platform is related to cookieless tracking. GA4 is touted as privacy-centric, specifically designed to work with or without cookies.

via GIPHY

Google is achieving this through leveraging the machine learning that their ranking algorithms are famous for and embracing statistical modelling, filling in the gaps left by the increasing number of us ‘opting out’ of cookie consent.

As we make the shift into a privacy-focussed, cookieless future, GA4 ensures we’ll still have the resource to measure and monitor user behaviour.

What’s up with cookies anyway?

In essence, nothing. A singular cookie is simply a file that stores a small amount of your data to be used at a later date.

Cookies allow for auto-fill information, they remember login details for you and store products, like limited edition Star Wars socks (looking at you Dave), in your basket after you’ve left and returned to site. Even though you were definitely, never, no chance, ever really going to buy them…

…That’s what I keep telling myself, anyway.

However, cookies can store more personal information, like names, addresses and bank details. So when you consider that these innocently-named folders enable remarketing and retargeting campaigns to track wherever you go on the internet, the risk of a breach in privacy becomes more of a concern.

Not all cookies are baked equally

There are two types of cookie; first-party and third-party.

First-party being those used on just one domain (the one you’re currently visiting), third-party being those that follow you around like a bad smell, tracking your behaviour and building up a digital picture of ‘you’ online.

Because of the sketchy-at-best sharing of data that third-party cookies pose, the EU introduced GDPR, requiring websites to give us the option to opt-in or out of both types of cookie.

The relevancy to GA4 here is that the platform relies on first-party cookies to track data and stay compliant with GDPR and not dissimilar CCPA. Interestingly, Google announced in 2020 that they would be blocking all third-party cookies by default, with the roll out taking ‘til 2022. Incidentally, that date has moved to 2023… the same year that UA will stop tracking and GA4 wholly takes over. Coincidence? Probably not.

via GIPHY

How is GA4 privacy focussed then?

If the relatively recent iOS14 update is anything to go by, anyone who’s anyone in data is already onboard the cookieless train. And, iOS14 has already started to make its presence felt across other channels, like paid social. GA4 may rely on first party cookies to track data for the moment, or at least while people are still opting in, but they have an ace in the hole.

Remember we briefly mentioned statistical modelling earlier? It’s this approach to what Google have dubbed ‘blended data’ that will enable them to traverse the cookie phase-out, and still offer us the tracking we need into the future. In their own words:

What the FLoC are you on about?

The modelling approach to data is not a new concept for Google. Back in 2020 they introduced FLoC or Federated Learning of Cohorts for a snappier title. FLoC uses Chrome to track user behaviour anonymously, assigns users to an ‘interest cohort’ where Google then creates predefined buckets that people fit within.

As part of their Privacy Sandbox initiative, Google went on to present a data set in 2021 that displayed how the innovation around statistical modelling can deliver results nearly as effective as cookie-based approaches, laying out their pathway to eliminating third-party cookies and replacing them with viable privacy-first alternatives.

Granted, both FLoC and Privacy Sandbox are related specifically to Google ads, but as we’ve said, GA4 adopts the same principles in ‘filling in the blanks’ from missing cookies.

Additionally, the GA4 platform is intrinsically linked with the other Google suites that rely on data manipulation, including Ads, Shopping and Tag Manager.

Whilst the loss of cookies entirely looks to be a way off, it seems ultimately unavoidable. Whilst the cookieless model definitely has its hurdles to clear, as we’ve already seen with iOS14 and Facebook Ads, Google’s statistical modelling offers the best solution we have in the long-term.

So where does GA4 fit into all this?

Because of the machine learning placed at the heart of GA4, how it learns about user behaviour anonymously and builds behaviour patterns specifically for your property from multiple data streams, shifting now to begin gathering your own data and strengthening your learning models is essential.

In the words of Winston Churchill;

Get on the GA4 train now, before it leaves the station

After a lot of heavy information, outlining the general need for GA4 and the relative immediacy that we all need to feel in adopting the new platform, it’s important to keep in mind that the change offers far better access to previously untapped data.

We’re going to write about the specifics of how the platform has changed, the shift in certain metrics, the increased integration with ads and more granular insights for analysis and reporting in another article. Cuz this one is long enough.