Why Should You Care About GA4 Now?
I know many of us are fed up by all the significant changes we have gone through in 2020 so far. Of course I am talking about the evolving challenges of collecting first and third party data in digital analytics space! These changes create obstacles for our daily job as digital professionals, but they also challenge the ethical discipline of our industry and enforce the establishment of highly efficient analytics strategy based on consensual data.
How can one embrace these changes as opportunities to grow and to adapt? GA4 provides a framework to facilitate the re-design of your digital data architecture at this critical time. In the coming months, we will launch a series of blogs to share our testing results and insights of the individual features with you, but today I want to highlight the significance of GA4 in three areas: 1) Simplified event-based data model; 2) Intelligence centred insights; and 3) Cross device integrations.
An Evolution of Google Analytics Data Model
Unlike Universal Analytics’ data model, which groups user interaction data into 4 scopes (i.e. user level, session level, hit level and product level), GA4’s data model is based on hits only. All user interactions are tracked by individual events alongside extra attributes of the interaction.
As we gradually enter the cookieless world, we can no longer perceive the concept of ‘sessions’, ‘bounce rate’ or ‘landing page’ without diving into details. We cannot ignore the potential flaws of these metrics caused by the absence or latency of cookie consent. Businesses need to define what ‘a site visit’ means for their own audiences, rather than for general web users. This simplified event-based data model provides flexibility of processing user interactions in a business specific context.
Furthermore, instead of trying to shove all event data elements into the one-size-fits-all fields (i.e. event category, action and label) event data attributes are now stored in custom defined key/value pairs. This allows business to tailor the behavioural data structure to suit their business and marketing needs. I will explain the benefit of this point in the next section. To wrap up this section, here is a view of the GA4 data model
“Autobots, Transform, And Roll Out”
Please make sure you read the subtitle in Peter Cullen’s voice, and let us move onto the most exciting element of GA4 – the intelligent insights! Google has deployed an advanced machine learning model behind GA4. The data collected across web and app will feed into Google’s powerful algorithm, enabling the platform to learn and produce intelligent insights and predictive models.
In the previous section, I mentioned that events are no longer using category, action and label fields. In GA4, an interaction (e.g. add to basket) is given an event name, then all other useful data attributes (e.g. product name) are captured through custom dimensions with that event. Google provides vertical-specific recommendations to set up commonly used interactions:
- Events for all sites and apps
- Events for retail and ecommerce
- Events for jobs, education, local deals, and real estate
- Events for travel
- Events for games
These vertical based events will be used to enrich the intelligence for different industries. Hit level data functions as the raw ingredients of Google’s machine learning system, which generates real-time insights with minimum effort.
A Road to Cross Device Insights
GA4 natively combines data streams from multiple digital platforms into the same reporting property. A common use case would be combining the behaviours across web and app using the same event-based data model. By doing this, businesses can analyse user interactions in a device agnostic way and/or analyse full user journeys across devices.
Universal Analytics and Firebase Analytics have completely different models and reporting systems. In the past, to be able to bring these two groups of audiences together we had to push the two data sets into one data warehouse project and restructure the data from there so that we can engineer the visualisation in a reporting tool.
With GA4, app and web data will be visualised in the same reporting UI without further data manipulation. For the time being, a business account ID is recommended to be used as the common key to deduplicate user between devices. In the future, Google Signals will be used as the primary solution for identity resolution, which of course will be subject to appropriate privacy controls and consent management.
An Unexpected Journey
Remember that time when an ancient wizard showed up at your doorstep with a bunch of dwarves and asking you to go on a journey with them? And remember how you hated it in the beginning, and how it transformed your life in the end? Change can be scary, but it often surprises you with joy. If you have read this far, I am sure, like me, you are passionate about our industry, and you believe in using scientific methods to improve efficiency. This may be an unexpected journey for you, but it is time to take the first step.
Let us outline a few milestones to help you draft a roadmap for adopting GA4:
- Identify key consumer lifecycle stages
- Build business KPI structure based on the key stages
- Identify interactions across all platforms that facilitate the consumer’s engagement throughout their conversion journey
- Map data elements against each interaction that contribute to measuring the success of the KPIs
- Design event data structure and create a measurement matrix
- Implement GA4 in web and app using the same event data structure
Finally, I wish you all the best for this exciting journey. If you want to bring us along with you, get in touch with us, we are here to support.
Authored by:
