Understanding Google Analytics' Data-Driven Attribution

Getting More from GA4's Data-Driven Attribution

In the realm of digital marketing, attribution models play an integral role in understanding customer behavior. They help businesses identify the customer touchpoints that contribute to conversions, allowing them to optimize marketing strategies accordingly. However, to be effective, attribution must accurately reflect the way each marketing element influences conversion and revenue. Google Analytics' (GA4) Data-Driven Attribution (DDA) is a vastly improved attribution model.

Read More

• How It Works
• DDA Setup
• DDA Marketing Impact
• Activating Marketing Data
• Attribution Reports
• Advantages 
• Disadvantages

What is Data-Driven Attribution (DDA)?

Only recently has Google Analytics made Data-Driven Attribution (DDA) available to everyone. Most digital marketers understand the limits of single-click attribution (first-click, last-click). Rules-based models, whether single- or multi-touch, are based on simple, predefined calculations that determine how credit for a conversion or sale is attributed to marketing touchpoints. For example, attribution awarding 100% revenue credit to the last click, or 40% each to the first and last click rarely matches actual customer experiences.

In contrast, data-driven attribution (DDA) is a dynamic model that looks at user history with the brand and conversion events and determines which are most important for a given conversion. In short, the DDA model is brand-, user-, and event-specific. 

How It Works

DDA can be thought of as a never-ending two-step process: building models and applying them.

1. Building the attribution model: DDA evaluates vast user historic data for the brand (the account), and considers numerous factors such as interactions (conversions and non-conversions), ads, assets, order, channels, devices, etc. to train the algorithm model.
2. Applying the attribution model: For any given conversion, DDA uses the trained algorithm to determine the effect of each touchpoint on the conversion, and credits touchpoints based on probability to convert. The model effectively evaluates how the presence or absence of each touchpoint affects the conversion. Data scientists will recognize this as a high-level description of the Shapley Value.

The model constantly learns and adapts to changing consumer behavior, making it a robust tool for understanding the nuances of the customer journey.

DDA Setup

Setting up Data-Driven-Attribution is simple. Navigate to a property’s Admin > Attribution settings > Reporting Attribution Model.  Select “Data-driven” from the drop-down menu. DDA is now the recommended and default attribution model for properties in GA4. 

As Google announced in April 2023, almost all rules-based models will be phased out by September 2023. As a result, first-click, linear, time-decay, and positioned-based will soon be unavailable. Instead, users will be able to choose DDA (recommended) or last-click. 

DDA Marketing Impact

Data-driven attribution provides a deeper understanding of how different marketing touchpoints contribute to customer actions, conversions, and revenue. DDA’s machine learning distributes credit across all touchpoints according to the role each played in the conversion path.

• Optimize Marketing Channels: Understand which channels are driving conversions and which are not, reallocating budget and resources to channels performing well.
• Enhance Content Strategy: Identify which content drives conversions, creating more high-performing content and reducing non-performing content objects and messaging.
• Improve Remarketing: Note obstacles and accelerators, improving performance through marketing intervention, such as specific ads and targeted emails where customers drop from the funnel.

As a cross-channel multi-touch attribution (MTA) model, DDA provides more accuracy and a more comprehensive view of the marketing mix components. As a result, marketers can elevate the impact of campaign optimization, forecasting, and even testing and experimenting.

Activating Marketing Data

GA4 provides multiple methods to extract and apply learnings from DDA results. While some are completely automated, others require more effort from marketers.

Automatic Insights and Application

• GA4 automatically generates insights from marketing data
• GA4 can drive performance in Google Ads by using DDA conversion data to improve bidding and placement. Marketers need only link their accounts.

Manual Insights and Application

• Users can manually connect other platforms to GA4 using additional code, such as Facebook Conversions API, to pull additional channel insights into the datastream. 
• Marketers can create custom insights triggers that serve perceptive information around key brand KPIs, campaigns, dimensions, and metrics.
• Marketers can export data to BigQuery, and then use a preferred analytics visualization tool to calculate insights.
• Marketers can access data from a number of built-in reports to identify insights to manually apply to marketing mix elements.

Attribution Reports

The attribution results from the DDA model can be accessed in various reports within Google Analytics. These reports provide insights into the conversion performance of different channels, campaigns, keywords, and other touchpoints. Armed with this information, marketers can make informed decisions about where to invest their resources to maximize ROI.

Note, some reports will not reflect DDA, while others allow users to explore multiple models from within the report.

• Traffic Acquisition reports use the last non-direct click. For example, when a user last visits organic Google Search before reaching a brand touchpoint, organic Google Search is the traffic source. This report shows how people arrive for a given session.
• User Acquisition reports use first-click attribution. The first brand touchpoint clicked is considered the user acquisition. 
• Advertising Attribution > Model comparison allows comparisons between multiple models in columns. For example, marketers can compare conversions and revenue across touchpoints for last-click vs Data-Driven. 
• Advertising Attribution > Conversion paths estimates attribution within three main journey buckets (Early, Mid, and Late touchpoints), and allows different models to be selected.
• Engagement > Conversions report calculates attribution for any selected conversion event.
• Advertising > Performance > All channels may be the most effective way of reviewing revenue attribution by channel and touchpoint combination.

Advantages

Compared to prior rules-based models, GA4’s DDA offers great advantages to marketers with a relatively simple marketing mix. 

• Accuracy: By leveraging machine learning, the DDA model provides a more accurate representation of the customer journey than traditional rule-based models. It factors in the complexity and interplay of multiple touchpoints, leading to a more nuanced understanding of what drives conversions.
• Touchpoint Inclusion: DDA leverages up to 50 touchpoints for any conversion instance, ensuring that maximum touchpoints are considered for attribution. Earlier versions considered only 4.
• Dynamic Adaptation: The DDA model continually learns and adapts to changes in consumer behavior and market dynamics. This means it remains effective even as marketing campaigns evolve and customer behavior shifts.
• Optimized Resource Allocation: By accurately assigning conversion credit, the DDA model helps marketers identify the most impactful touchpoints. This information allows for more effective resource allocation, potentially leading to improved ROI.
• Comparisons: Currently, brands can compare multiple models and their differing allocations across touchpoints.

Disadvantages

While Google’s DDA is a significant improvement over prior models, it is still a small component of a Google-centric platform. Along with many advantages, Google’s DDA is not without shortcomings.

• Data Requirements: The DDA model requires a significant amount of data to function effectively. Smaller businesses or those with low traffic volumes may struggle to meet these data requirements, limiting the model's usefulness.
• Complexity: The DDA model's algorithmic nature makes it more complex than traditional rule-based models. This can make it difficult for some users to understand and interpret the results. For example, multiple reports reveal attribution in different ways, creating interpretation challenges. Additionally, GA4 uses different dimensions for traffic sources (e.g, First User Source, Session Source, Source).  Each dimension provides different results for brand data.
• Limited to Google Ecosystem: While the DDA model does an excellent job of attributing credit within the Google ecosystem, it may not fully account for interactions on other platforms or offline touchpoints. This could potentially lead to an incomplete understanding of the customer journey.
• Black Box Modeling: DDA’s opaque system of attribution does not expose how calculations operate in a given conversion journey. As a result, marketers cannot identify how clicks and interactions are calculated to measure impact. 
• Aggregate/Sampled Data: GA4 uses data sampling and machine learning to better connect touchpoints to conversions. Moreover, data is most often provided in aggregate, making it difficult to extract specific journey information for analysis. 

While the advantages of GA4's DDA are tempered by the drawbacks, today's marketers no longer need to depend on Google or channel partners for attribution clarity. Arcalea's Galileo is a full-funnel omni path attribution platform that multiplies insights and revenue.

Conclusion

In an era where customers interact with businesses through a variety of touchpoints, the importance of a sophisticated and accurate attribution model cannot be overstated. Google Analytics' Data-Driven Attribution model offers a potent solution, leveraging machine learning to provide a more holistic and nuanced understanding of the customer journey.

However, like all tools, it has its limitations. It's essential for businesses to understand these limitations and consider them in the context of their unique needs and circumstances. Brands with sophisticated marketing mixes or with the need to activate data across platforms should consider full-path, transparent multi-touch (MTA) solutions.