How To Avoid Dips In Marketing Performance With The Shift to GA4: 4 Things You Need To Know

In 2019, Google rolled out Google Analytics 4 (GA4) as an alternative to Universal Analytics (GA3, UA). Indeed, you already know that the latter will be phased out on July 1, 2023.

Whether you’ve followed Google’s recommendations about running two analytics solutions in parallel or not, chances are, on July 1, you will dive head-first into the murky waters of GA4.

The two platforms couldn’t be more different, primarily due to the event-driven shift. With the transition, metrics will change and could bear sustainability on your marketing analytics routines, if not wreak complete havoc over the first few months.

This article explores the pitfalls you might face as a marketer, even if you come prepared—it will help you gain a handle on four crucial changes between GA3 and GA4 that may overturn your current approaches to marketing analytics and reporting.

Event-Driven Shift and Metrics Disruption

The most significant shift in GA4 is with the data model. While UA tracks sessions (and events, such as conversions, within them), GA4 differs: each user action is considered an event.

How different is GA4 approach to data measurement?

Regarding data measurement, GA3 and GA4 are much like two different worlds. Within GA3 sessions, user interactions are captured as various types of ‘hits,’ where ‘event’ is one of those types. Google Analytics 4 shifts from ‘hits’ to ‘events.’ While GA4 allows for a more flexible analysis, it also complicates it.

UA had 115 pre-built reports available, most of which will become obsolete when GA4 takes the lead. If you doubt some data points in your master dashboard in a BI tool and decide to check some metrics by hand, likely, you won’t be able to do that in the Google Analytics 4 UI.

The inconsistency in metrics between the two analytics platforms also has quite unobvious implications. Remember the fear of losing audiences created in UA? Even though Google made the audiences migration guide, due to the lack of some metrics and dimensions in GA4, your migration process will likely still be complicated.

Since GA4 emphasizes events, it also pushes users toward creating custom reports. Here are the event types GA4 collects and supports.

The four types of events in GA4

1. Automatically collected events: GA4 collects them independently after analytics run (e.g., page_view, first_visit, and session_start).

2. Enhanced measurement events: Automatically captured interaction with content (among many others are site search, outbound clicks, etc.).

3. Recommended events: Google Analytics 4 defines a set of events to be tracked based on their importance (not automatically collected).

4. Custom events: Not being predefined, you are to set up names and sets of parameters to track and analyze, depending on your needs.

In GA4, up to 50 event parameters can be logged with each event (plus 25 for user properties and 10 for item properties).

GA4 Attribution and Channel Performance

Google Analytics 4 continues the Google Ads crusade on attribution models treated as typical by us marketers.

Firstly, the list of models changed, and you won’t find anything like ‘last non-direct click’ in Google Analytics 4. Furthermore, the rate of changes Google implements in their latest web analytics layer is pretty high. For instance, Google Analytics will completely sunset first-click, linear, time-decay, and position-based attribution models by September 2023.

On the one hand, nothing should change for Enterprise marketers exploring reports in a BI instrument, such as Tableau. Under the hood, your dashboard was previously extracting data from Universal Analytics, and then came the data engineering team and switched the data source to GA4. However, the picture of your marketing performance greatly depends on the attribution model you used with UA and the one set up in GA4.

For example, GA4 has at least two different last-click models: ‘Paid and organic last click’ and ‘Google paid channels last click.’ And, if, when comparing your YoY and QoQ marketing performance, you suddenly see a massive boost in paid ads acquisition (and related benefits such as unexpectedly better ROAS or lower CAC), make sure to check in with your analysts on the attribution model in use.

The Era of Custom Reporting and Analytics Maturity

Your marketing reports and metrics have changed significantly. They look different and also require higher levels of analytics maturity to operate.

For those who relied on standard GA reports for years, wiping most of them out of GA4 was a massive blow. Google Analytics 4 emphasizes custom reporting. It embodies a more proactive and mindful approach to marketing analytics. It does it at the cost of having a proper analytics strategy and more SQL capabilities.

Generally, you can use the ‘Explore’ mode in GA4 UI to create reports tailored to your requirements. However, for Enterprise marketers, this would mean crafting new specifications for data analysts to route necessary data through Google BigQuery to a BI instrument. It means that beyond optimizing, testing, planning, and whatnot, marketers will now have to explain (again) how they want to look at the performance data from their campaigns.

Another roadblock is that historical UA data cannot be directly merged with GA4. Within the Google Analytics interface, you can no longer compare your current performance with stats from a year or two. The solution exists, yet it is relatively complex: you can extract data from both platforms to a separate storage, normalize that data, and load it to your BI dashboard for analysis.

Your actual journey will likely be even longer: analysts will have to inspect the data model behind your current dashboard, discover UA metrics in use, figure out how they translate to the GA4 terms, extract data from both platforms, normalize it, push to your data model (or update it), and only then connect to your dashboards. But even this scrupulous and highly demanding method does not guarantee a perfect match–occasionally, you will trip over the mentioned n/a issues.

As marketers and analysts navigate the complex task of matching UA and GA4 data, Improvado sails in as their lifebuoy. Seamlessly map UA to GA4 with Improvado and focus on conquering the marketing landscape.

JOIN THE DOTS

Google BigQuery and Higher Time-to-Insight

In the UA ecosystem, direct data exports to Google BigQuery (or GBQ) were Google Analytics 360 users. With GA4, all users can stream their data to GBQ for free. And there are several ways of looking at it.

• With Google BigQuery, businesses of all sizes can be more agile and responsive to changes in the market landscape by leveraging data-driven decision-making.
• Marketers will have to undergo GBQ and Looker Studio training. Or, if they use a standalone data visualization tool, consult analysts and data engineers whenever they want to explore a specific insight.

In large enterprises, the latter will always require more than one person, a more complicated process, thus extending time-to-insight and diminishing the respective returns. And, if insight is uncovered from data you don’t entirely trust, it will be even slower: check the data first and then shift to exploring the insight.

Conclusion

Due to data models, reporting, and attribution differences, transitioning from Universal Analytics (GA3) to Google Analytics 4 (GA4) is a significant challenge. Despite potential disruptions, GA4 brings more flexibility and agility to data analysis, facilitating a more nuanced understanding of customer engagement and marketing performance. However, readiness to adapt is crucial. Anticipate a learning curve for SQL capabilities, data management, and analysis.

In essence, the shift from GA3 to GA4 is more than an upgrade—it’s a reorientation in data understanding and usage. Since you will still undergo the tectonic shift between the analytics platforms, consider upgrading your complete marketing analytics stack to prepare for better insights and the future of AI and predictive analytics.

Book a consultation to automate your entire marketing analytics and reporting with Improvado.