Can GA4 deliver the omnichannel data revolution?

With its stripped-back interface and fewer built-in reports, many found GA4 a daunting proposition. The change from UA brought its own challenges, including concerns over what historic data may be needed in the future, and where it can be most usefully stored and visualized once Google pull the plug on UA account access (1 July 2024), losing legacy data.

Inevitably, thoughts of wider reporting, business intelligence, data management and the processes that may be required have never been more front of mind.

Data-driven attribution

GA4’s switch to data-driven attribution is already making a material difference in marketers’ challenge to assign the credit for purchases in a coherent and actionable way – though, inevitably, that won’t cover all potential marketing and customer touchpoints.

Many engagements with consumers take place with a brand’s digital estate or offline entirely. Out of the box, GA4 facilitates the pipelining of web and app analytics data to its cloud data warehouse (BigQuery) as a free, native connection. This will be a key starting point for many in their journey to create an omnichannel reporting architecture, straddling on- and offline interactions.

This gives us the ability to bring the reporting of ad delivery from different platforms together, alongside web engagement and customer data to establish a holistic view of marketing activity. Removing marketing data and ad performance data from siloed platforms allows it to be set beside real-world data (sales figures, footfall, CRM database numbers, telemetrics) and then dissected and understood in a consistent and homogenized way: sliced by time period, channel, product, outlet, type of outcome, or basket size.

Combined activity

This is an approach many of our clients are developing: full marketing mix modeling (MMM), plus focused data storage and transformation projects, to create combined reporting layers that are geared toward measuring channel performance and their capacity to deliver against the organizational objectives that marketing is pursuing. This is a significant departure from previous attempts to understand the value of media investment by channel: methods like last-click attribution, which cannot tell the full story alone.

Building and refining models of combined measurement to see the overall impact on investment in certain channels, formats and campaigns is key in delivering an omnichannel attribution approach. Rather than identifying individual users and concerning itself with specific paths to purchase, this approach offers a more holistic understanding of performance, and is beneficial for maintaining a view of overall objectives and for understanding (and, crucially, predicting) the impact of additional investment in specific strands of activity on various parts of the customer journey to purchase.

Getting started with GA4

GA4 offers greater connectivity than its predecessor and as such is a solid starting point for many marketers. But it is truly only the first step. It’s unlikely that any ‘out-of-the box’ analytics/attribution solution is going to have the level of customization needed for data management and reporting to suit anything other than the simplest use cases.

Brands with multiple digital and offline touchpoints as well as a requirement to drill data consistently down into sub-markets or to product level will always need well-connected and properly structured data to achieve functional MMM and visualization.

Automation is vital here too: proper data hygiene is key. For MMM to be workable at scale, hygiene can’t be achieved through the application of too much human intervention. Manually loading data must be avoided wherever possible – it may be sufficient for some brands and can be achieved on relatively low budgets, but with scale and complexity, the real sophistication comes with the skills of data engineers in creating dynamic matching algorithms and data scientists who know how to work with the combined data set, using industry-specific knowledge to deliver insightful predictive modeling.

Models can (and undoubtedly will) grow in complexity when linked with more extensive data beyond what’s available to sales and marketing. From this will flow increased benefits to insights, marketing activity and budget management. Macroeconomic, socio-demographic, consumer and industry-specific data could all be factored into a more sophisticated model.

There’s a huge opportunity here for practically every advertiser, and the potential for enormous complexity for many. But the basics are relatively simple, and initial engagement with this space is consequently also straightforward and not overly costly. The key elements of this working effectively are organizational: a good understanding of what success looks like, and ensuring all stakeholders know how measurement models are constructed and viewed for action. With the support of an agency expert in the space, and with the foundation of an effective GA4 setup, the rest can be constructed over time.