author-avatar

As a part of the Marketing Intelligence team, Sylwia's role as Analytics Specialist allows her to combine a passion for problem solving and continual learning with data analysis and interpretation. Always eager to expand her perceptions, you can often find this bibliophile's nose buried in a book, occasionally coming up for air to play a game of chess or to head out into the wilderness for some back-country camping.

Other posts by:

Google Analytics 4: What (We Think) You Need to Know

Google Analytics 4 is the newest iteration of Google’s measurement solution, and with less than 7 months until Universal Analytics stops processing new data, the pressure is on for web administrators and marketers to implement GA4 and start learning its ins and outs.

As GA4 will become the default reporting option for many businesses, it is important to understand that this solution works very differently from Universal Analytics and thus the learning curve can be quite big.

Vovia has been setting up GA4 for clients since its launch, and the purpose of this blog is to shed light on our learnings so far when implementing and navigating this program. 

3 Key Differences Between Universal Analytics & GA4

While this section outlines the main differences between these platforms, I often find myself coaching clients to not think of one as an upgrade from the other. GA4 is not an upgraded version of Universal Analytics. It can be considered a whole new analytics platform, utilizing a very different data model from what we are used to. Looking at it from this perspective will reduce the headache down the road, though it will take time to retrain your brain into thinking within the GA4 platform.

Here are 3 key differences between UA and GA4:

  • Single Property to Track Web & App Data
    • Shifting from a hit-based model to an event-based model allows GA4 to combine the data across both web and app, where previously these had to exist as separate Universal Analytics properties. As a result, the ‘Views’ metric in GA4 is a combination of pageviews and app screenviews. 
  • Event-Based Measurement Model
    • The Universal Analytics measurement model is based on sessions and pageview data. In GA4 all user interactions, including pageviews, are now tracked as events. As a result of the switch to event-based tracking, you should take time to rethink your existing tracking set-up, and we recommend referring to Google’s Event Migration Guide for best practices on mapping events in GA4. 
    • Table from: https://support.google.com/analytics/answer/9964640?hl=en#zippy=%2Cin-this-article
  • The Addition of Session ID
    • Just like Universal Analytics, sessions in GA4 will end after 30 minutes of inactivity. However, unlike Universal Analytics, GA4 generates a session id at session_start. This allows GA4 to continue to maintain the same session even after midnight (Universal Analytics would reset at 12AM, resulting in a new session).
    • GA4 does not create a new session when the campaign source changes mid-session, while Universal Analytics does create a new session under that circumstance. As a result of these changes, you can expect GA4 sessions to be lower than what is currently reported in Universal Analytics.

TL;DR : What do these above changes mean for you and your stakeholders?

  • With all data housed within the same property, combined with the implementation of Session ID, GA4 can now capture a more accurate picture of the full user journey.
  • Increased likelihood of discrepancies between business data (first-party data) and GA4 reporting. This can include differences in reported eCommerce Revenue & Conversion numbers.
  • Year-Over-Year & Period-Over-Period data comparisons that combine Universal Analytics and GA4 dates become unreliable, as standard metrics like Sessions and Pageviews are no longer measured in the same way. Anticipate seeing drops in Sessions when switching to GA4. 

What We Like About GA4 So Far

  • Automatically Collected Events
    • Even if you get no further than following the GA4 Setup Assistant tutorial in your existing Universal Analytics Property, GA4 will still collect certain events by default (either as a part of automatically collected events or enhanced measurement events). These events include click page_view, (outbound) click, file_download. 
    • Before you begin to add recommended and custom events, take the time to explore GA4. Investigate the pre-defined reports and build custom explorations using the automatically collected events. Better understand what GA4 captures with each event by monitoring triggered events in realtime using the DebugView feature. 
  • Increased Focus on Engagement Metrics
    • Engagement rate is a new percentage metric within GA4 that reports the percentage of engaged sessions over total sessions. An engaged session is a session that lasts longer than 10 seconds, has a conversion event, or has at least 2 pageviews or screenviews.
    • With the adoption of this new metric, GA4 has also re-formulated its bounce rate calculation. Bounce rate is now defined as the percentage of sessions that were not engaged sessions.
  • Robust Default Channel Grouping Structure
    • Finally Paid Social has been peeled out of Paid Search, and new Google Campaigns like Performance Max are accounted for under ‘Cross-network’. To utilize this updated Channel Grouping to the fullest, we recommend using the  GA4 Default Channel definitions when generating UTMs to ensure your campaigns are getting bucketed in the correct Channel Group. 
    • Example below is the same client, but showing UA vs. GA4

Considerations When Using GA4

Sadly there is no one size fits all solution for analytics measurement, and when implementing and reporting out of GA4 it is important to remember that it too has limitations that will result in data discrepancies, especially when comparing it against your own company’s first-party data or against another analytics platform.

  • Reports vs. Explorations
    • Reports and explorations are each intended to provide a different view of GA4 data, with some dimensions and metrics being available in reports, but not explorations. Additionally, if you are working with large amounts of data, or are attempting a report with high granularity, both reports and explorations will result in sampled data.  If you have used Universal Analytics in the past, then the concept of sampled data is not new to you, sampling allows Google to aggregate and report on data in a timely fashion. However, reports and explorations use different methodologies to calculate the sample subset which results in differences in the reported data. 
  • Thresholding
    • GA4 applies thresholding to prevent anyone viewing a report or exploration from inferring the identity of individual users. As a result, data may be withheld from reports and explorations, again resulting in data discrepancies between first-party data and GA4.
  • Modelled Data & Conversion Attribution Delays
    • In response to limitations on collecting data and tracking the full user journey accurately, GA4 utilizes modelled data to estimate online conversions that can’t be directly observed. Modelled conversions look for trends on directly attributable events & conversions, and using machine learning apply those trends to unobservable conversions. 
    • When utilizing GA4 reports, keep in mind that conversion attribution data for a channel can take up to 7 days to be updated and attributed accurately after the conversion is recorded. GA4 uses this time to process the data and use it to train the model. When reviewing and analyzing your GA4 data, Google recommends selecting a date range prior to the previous week for increased accuracy. 
    • The result of GA4 using modelling is that while we will be able to see overall trends in user behaviour and conversions by channel, the smaller the data set or date range, the greater the inaccuracy. 

Final Thoughts

We can anticipate that Google will continue to add new features to GA4 over the next few months, until Universal Analytics stops processing new data as of July 2023. Keep in mind many third-party platforms that offer direct integrations with Universal Analytics, such as eCommerce and CRM tools, have not yet updated their functionality to integrate directly with GA4. As a result, I recommend implementing GA4 as soon as possible and running it in tandem with your existing Universal Analytics. Continuing to run Universal Analytics will allow you to still leverage the data collected from these integrations until they build out a solution, while at the same time allowing you to accumulate historical data in GA4.

While you may decide to not utilize GA4 as your final analytics measurement tool, it is still a necessary part of your website’s configuration. Getting comfortable setting up and using GA4 now will ensure a smooth transition come July 1st. If you need help with this, reach out to us.