It’s not clear how many websites and apps use Google Analytics – but look at the source code of any web page and you’ll more than likely see that UA code snippet somewhere on the page so I would guess… most do. And it’s been working perfectly well, providing fascinating and useful insights over the years to millions of website owners, data analysts, UX and SEO professionals who have used these insights to fine-tune websites to meet business goals.
So why then, did Google announce in March this year they will be sunsetting Universal Analytics on 1 July 2023, with Universal Analytics 360 (the paid, premium version) following three months later on 1 October?
Because, quite simply, legislation and the way we consume digital content have changed profoundly since Google bought Urchin in 2005 and introduced free analytics to the masses later that year.
Back then Google Analytics was designed for online measurement in a simpler world, where the desktop was front and center of our digital experience and measurement was based on cookies. Fast forward 17 years and today we flip between phones, tablets, laptops and desktops in our fragmented user journeys – creating multiple user IDs along the way and providing major headaches to data analysts trying to build consolidated views of user journeys and behavior. Add to that evolving privacy regulation, browser and operating system updates that actively make measurement harder and individual choice in what data they decide to share – or not share – and suddenly gaping holes are appearing in digital measurement frameworks.
Unfortunately, this situation will not improve, with Chrome (the world’s most popular browser) announcing that it will deprecate support for cookies in 2023. The actual date has been moved out previously but it is certain to happen at some point.
Google started developing GA4 more than two years ago (then simply known as Web+App before rebranding)The sleek forwarded looking electric alternative to the dated, petrol-guzzling V8. Make no mistake, this is not simply an upgrade, but a total rebuild with a completely new data model. Where UA used a session-based model for measurement, GA4 uses an event-based model, so a straightforward comparison in data will not be possible. For example, in a session-based model, where a user watches three videos, it would count as one conversion. Using an event-based model would count three goals.
GA4 seeks to solve three major pain points that digital marketers are facing today:
- Measurement that is focused on respecting user privacy
- Future-proofing the measurement model
- Providing a unified view of user journeys across both web and app
Respecting User Privacy and Future-proofing measurement.
What the future holds in terms of privacy legislation and browser updates is unclear. The only certainty is that it will change. And it is important to appreciate the nuance: Government legislation will, to a large degree, influence how technology companies like Google, Facebook and Twitter develop their media buying platforms. Browser and platform providers, like Apple, Firefox, Chrome and Microsoft, while impacted by legislation, can – and have – made unexpected decisions unrelated to legislation that has significantly impacted the ecosystem.
Think about Firefox blocking cookies in 2019 and now Chrome indicating they will do the same in 2023. To future proof measurement, Google completely rebuilt its analytics measurement model, moving away from a cookie-based model to an event-driven model using machine learning to fill in the gaps. As an added benefit, the Machine Learning in GA4 voluntarily surfaces insights in the user interface that may not have been that obvious to data analysts previously.
To ensure privacy requirements are met Google has retained UA’s existing privacy settings and added a number of additional features, including not storing IP addresses and only making use of first-party cookies. Through Google’s Consent Mode, they also make it possible for users to decide what type of data they are willing to share, whether it be marketing or analytics data, or both. Where consent is not granted, Google will use ML to model conversion data as much as possible.
Unified view of user journeys.
With Universal Analytics marketers needed two separate properties, one for web and one for app, to measure user activity across their websites and mobile apps. With GA4 this is done in a single property, making cross-device user measurement easier.
While July 2023 may seem a long while off, it would be prudent for the forward-thinking digital marketer to set up a GA4 property sooner rather than later to run in parallel with their current UA setup. There are several reasons to follow this approach:
UA is no longer under active product development, so the UA you have now will be exactly the same one you’ll be using at its end of life. GA4 on the other hand has an active roadmap with regular features updates and product enhancements. The User Interface in GA4 is also brand new – use the next year to familiarise yourself and your team with new UI, features and ways of pulling reports.
Build up a record of historical data now so you are able to do a year-on-year comparison, rather than having to start from scratch on 1 July 2023.