Allo. Play Music. Answers. Lively. Buzz. Wave. Plus. Reader. Catalogs. Last month, Hangouts. Only a fraction of what has been killed by Google over the years. The tech behemoth has never been shy about launching and beheading their consumer products, but they’ve always played it a lot safer with their business users’ interests. Until now.
Why is this happening?
Privacy laws have changed, and the way Google tracks users on websites has to change too. UA relies on cookies and site pixels to track users around the web, so it isn’t compliant with mobile platform privacy policies or regulations like the California Consumer Privacy Act (CCPA) and the General Data Protection Regulation (GDPR)—the European Union’s set of laws against tracking people without their consent.
Google Analytics is straight-up illegal in major European markets because of a GDPR ruling in 2018, which was largely based on the collection of European citizens’ IP addresses. EU regulators banned GA in France, Germany, Italy, Austria and more. Not long after, New Zealand, Brazil, Japan and other countries passed their own privacy laws. The legislation was piling up.
Google launched GA4 in October 2020 as part of the solution, but at least a year after launch it was still in beta and missing tons of essential features. Barely anyone had migrated, and the sites that had were using it in tandem with UA. They had to get rid of it to enforce mass adoption. Google rushed into this because of legal necessity, and that’s reflected in how GA4 doesn’t feel like a finished product.
Google rushed into this because of legal necessity, and that’s reflected in how GA4 doesn’t feel like a finished product.
It is, however, compliant. GA4 was built for the higher privacy standards of a post-GDPR world, and only ever used anonymized IP addresses. Now, they’re removing them altogether. Not only does it work without cookies, it only uses first-party cookies (rather than third-party, which track users across multiple sites) and fills in the gaps with machine learning and statistical modeling. Google is anticipating a cookie-less future, and no doubt we’ll be grateful for the headstart when it arrives.
How different is GA4?
Short answer; completely. It’s a total overhaul in both structure and user interface.
Data Streams
How we think about the customer journey has changed completely since UA arrived in 2012. Instead of a funnel that ends at a sale, it’s a flywheel that retains customers and turns them into promoters. Brands and businesses need to attract, engage and delight consumers in an omnichannel landscape. GA4 allows you to track users across your site, apps, and multiple devices in one property using data streams.
Events
UA was about sessions, made up of hit types like page views, ecommerce transactions and so on. Sessions do exist in GA4, but everything is an event and passes through to Google differently. UA’s action, category, and label system is now parameters, enhanced measurement, and in-platform event creation. Whereas you would set goals in UA, you can now turn on enhanced measurement to collect everything from scrolls to downloads as events and then choose to mark any event as a conversion. You can also set up your own custom events using Gtag or Google Tag Manager.
Machine Learning
Machine learning tools aren’t new for Google Analytics. Launched in 2018, UA Insights helped people find the patterns and outliers in their data. But GA4 has ramped things up in a big way. Anomaly detection, predictive analytics, attribution projects—there’s too much to cover here, but suffice to say, the algorithms have arrived.
Reports
Plenty of familiar metrics just don’t exist anymore. The logic appears to be that UA had too many reports, which is arguably true. Personally, we didn’t mind them. They were all nested in different categories and maybe it was a little confusing, but c’mon, there was a search bar! Removing these reports has made it significantly more difficult to find the information you need within the default reports.
A lot of those missing metrics and reports are now hidden in an interface called Explorations, which is actually a powerful reporting tool. But it’s very hands-on and almost impossible to share the results. You can only share a “view only” version, so people can’t add metrics or dimensions. It’s also where they’ve hidden the user flow reports, which show how people travel through your website. Explorations is where you can actually use all the metrics and dimensions, assuming they’re compatible. With the default reports, it’s a case of “Here’s a table. Here’s a graph. Good luck.”
Our esteemed data analyst, Jake Johnson, added an event to a GA4 account recently and had one hell of a time just trying to figure out where the traffic was coming from. That’s someone who works with and teaches people how to use GA tools for a living. To say that GA4 is a steep learning curve is an understatement.
However, the higher powers do appear to be listening. GA4 feels like an escaped beta, but Google is adding things back in as people complain. After initially getting nixed, the UA “behaviour” metrics (like bounce and conversion rates) were brought back in July, as were some of the missing UTM (Urchin Traffic Monitor) tracking codes. Carly Boddy, Product Manager at Google Analytics, just announced two new metrics last week: views per session (number of screens/pages in a single visit) and average session duration (time spent on the site). Metrics are defined differently in GA4 though, so expect to be surprised by the numbers.
What’s going to happen to my data?
Just for fun, let’s do some math. On March 16th of this year, Google announced that Universal Analytics would stop recording hits on July 1st, 2023. If someone wanted to maintain accurate year-on-year reporting, they would’ve had to have GA4 up and running by July 1st this year. That gave people just over 3 months, which is a mammoth sprint for large companies where decisions take a while and implementation takes even longer. The real barrier is that setting up GA4 is one thing, and trying to import a complicated UA account is another. Import isn’t really the right word, since you can’t.
UA will stop collecting data on July 1st, but after 6 months, you won’t be able to access your historical data at all. Well, they’ve said “at least” 6 months, with the actual date to be announced “in the coming months”. Unless you can find a way to export your data, it will just be gone. And you can only export individual reports, unless you’re using the GA Reporting API, which has its limitations.
We’re still trying to find a solution. Some clients easily have 5 years’ worth of data, and without a built-in mechanism for exporting or migrating the historical data, the timeframe looks even more untenable. And while free accounts may not feel entitled to grumble, 360 members paying thousands of dollars a month only got an extra 3 months to save their data from the void.
But Russell Ketchum, Director of Product Management at Google Analytics, recently announced they’d pushed that sunset to July 1st, 2024. In response to backlash? Couldn’t say. Ketchum says that 360 customers have the most complex implementations, and need more time to collect data in both platforms so they have better comparisons post-transition. Translation: headache big, change tough.
It seems that Google recognizes the migraine in their wake, so they’re also updating the Setup Assistant. Whereas it used to just provide tools to help you migrate, their Jumpstart Initiative (starting early 2023) will automate the Setup Assistant so everyone has a functioning GA4 property in place.
While some hold out hope that Google will announce a way to migrate your data to GA4, it’s almost certainly futile.
While some hold out hope that Google will announce a way to migrate your data to GA4, it’s almost certainly futile. UA and GA4 use completely different data models, and the difference in schema and dimension definitions/calculations mean that merging the data is essentially impossible.
What should I do now?
Focus your efforts on GA4. A lot of people are trying to stick with UA for as long as they can, and we get that. It’s been established for so long, all your data is there, and you’re more comfortable with it. But the clock is ticking.
A basic GA4 setup needs to be done and any new tracking added there, so you can start generating data as soon as possible. Then think about how your existing UA setup can be transported to GA4, and develop a plan for saving your historical data. We can help, so talk to us.
How do I decide what data to export?
The most important work to do upfront is narrowing down what data you need to keep. There are just too many report combinations and dimension variables to save everything. With your team, figure out:
- What dimensions and metrics are most often needed or requested by your stakeholders? What segments and filters do you apply to your data?
- Do you really need more than 2 years of historical data? And what time intervals; are monthly comparisons enough? Quarterly?
- What format do you need the data stored in for future reporting and visualization?
What are the options for exporting my data?
- Cloud-based Storage: If you’re a 360 customer, you can use the native integration in GA4 to build a data export to BigQuery (BQ). If you’re not an enterprise customer, you could try a third-party solution like Supermetrics to connect your GA to BigQuery. You can use Apple or Microsoft cloud products, but BQ integrates with other Google products. Database integrations like the GA Reporting API also let you extract, transform and load (ETL) large data sets, but you’ll need the technical capability or help to run code in an application. The downside to both is that it might not take long for the investment involved to outstrip the data’s value.
- Manual Export: Open the GA report you want to keep, customize it for any segments or filters/secondary dimensions, then export it to PDF, Sheets, Excel or CSV. It’s easy and free, but limited and tedious.
- Google Sheets Add-On: Install the GA app in the Google Workspace Marketplace, then use the extension in your Sheet to create a report. You’ll be able to select your property, metrics, segments etc. and import it from UA into the Sheet.
- Third-party Tools: We already mentioned Supermetrics, and Locomotive has released an open source GA4 to 3 converter for use in BQ. There’s pre-built data pipelines by companies like Hevo that export historical data from GA to a database file or data warehouse of your choice. There’s also alternative analytics platforms like Matomo, if you’re ready to exit the Googleverse.
How about a silver lining?
The new tracking methods are more sophisticated, and GA4 is going to be an improvement in the long run. We’ve seen a few cases where GA4 caught a bunch of bot traffic that UA didn’t, keeping the user counts accurate. It’s a rushed and rocky transition, but Google is implementing user feedback quickly and releasing new features and functionality regularly.
The new tracking methods are more sophisticated, and GA4 is going to be an improvement in the long run.
Still, there shouldn’t have to be so much customization that goes into creating a basic account, and the information you need should be readily available without all the tinkering. But the saving grace here is Data Studio (now called Looker Studio), Google’s data visualization platform. UA pointed you to Data Studio when you tried to create custom reports, and really, Explorations is a hybrid of the old UA custom dashboard and the Data Studio UI.
It natively connects to any Google product, including Analytics, Search Console and Ads, and third party connectors for things like Facebook Ads. It’s free, powerful, and the visualization platform we use the most for analytics. There aren’t really any competitors. It lets you sculpt the data in a way that makes sense for the people ingesting it, and is invaluable for the end clients who don’t want to dig through layers of interfaces to find the metrics they’re looking for.
For the most part, we’ll fall back to Data Studio so people don’t have to deal with the GA4 interface. Unless Explorations become more shareable, that’s where we’ll curate data. However, not everything is in there. Bounce rate may have been added to GA4, but they didn’t add it to the Data Studio connection.
Google promotes and updates Data Studio regularly, and has been making relatively substantial changes to it every few months (exhibit one: new name). This gives us hope that it’s going to stick around. But with the track record, it’s easy to wonder. Will Data Studio also end up on the kill list? It’s a free product. If it seemed like it wasn’t worth the effort, if the functionality was covered elsewhere, if for any other reason, there would be nothing stopping Google from axing it. It’s their ecosystem, we just live in it. So keep calm, carry on, and let’s cross our collective fingers.
