Attribution: Multiple Assignment Models

Multiple assignment models allow you to distribute attribution credit across several campaigns or channels, providing a comprehensive view of the customer journey. This approach helps you evaluate the combined impact of multiple marketing touchpoints.

Why Use Multiple Assignment Models?

• To analyze complex journeys where multiple campaigns contribute to conversion.

• To allocate credit more fairly than single assignment models.

1 Overview of Multiple Assignment Models

We differentiate between manual models, where the user configures weightings or distribution rules, and the automatic model, which dynamically calculates channel contributions using data-driven insights.

The table below summarizes the key characteristics of the multiple assignment models, helping you decide which one best fits your needs:

2 Data-Driven Attribution (DDA)

Data-Driven Attribution (DDA) provides a transparent way to fairly distribute credit among all marketing channels in a customer journey. By analyzing historical journeys and applying Shapley Values, this approach ensures accurate attribution that reflects each channel's unique role and sequence.

2.1 Why Use DDA?

Marketers face complex customer journeys with multiple touchpoints. DDA simplifies this complexity by:

• Fairly distributing credit among all contributing channels.

• Using clear, data-driven insights instead of black-box algorithms.

• Continuously adapting to the latest customer journey data.

2.2 How DDA works

DDA follows two main steps:

1. Shapley Values Calculation

   Analyzes historical journeys to determine each channel's fair contribution.
   

2. Attribution to New Journeys

   Applies these calculated values to credit channels in future customer journeys.

Step 1: Shapley Value Calculation

Shapley Values are calculated by analyzing past customer journeys and assigning credit to each channel based on its marginal contribution. This process involves two key elements:

1. Equal Contribution Assumption: Each channel in a customer journey is assumed to contribute equally to the total conversion value.

2. Summing Contributions: Contributions across all journeys are totaled for each channel, and their percentage share is calculated to determine the Shapley Value.

Example Calculation:

The Shapley Values are then calculated by dividing each channel's total contribution by the overall conversion value (1000 €):

• SEA: 400 / 1000 = 40%

• Email: 450 / 1000 = 45%

• Display: 150 / 1000 = 15%

These calculations include adjustments based on the channels' positions in the customer journey. Channels closer to the conversion often receive more weight, reflecting their increased importance.

Step 2: Attribution to New Customer Journeys

Once Shapley Values have been calculated for each channel and position, they are used to determine the relative contribution of channels in new customer journeys. Here’s how this works:

New Cases – Example Calculation:

Attributed Values:

The percentages in the table above are normalized for the relevant channels in each customer journey to ensure the total equals 100%. For example, in the Display > Email journey, Email's contribution of 45% and Display's 15% are recalculated as 75% and 25%, respectively.

This step-by-step approach ensures that credit is distributed fairly to each channel based on its role and position in the customer journey.

2.3 Additional Information for Marketers

• Clear Visibility:
  Use the attribution metrics for analysis and the Smart notification to understand the calculated shared each channel has.

• Dynamic Updates:
  Values are recalculated daily using the latest 90 days of data, ensuring accuracy.

• Fair and Transparent:
  Shapley Values provide clear and logical attribution, avoiding the complexity of opaque machine-learning models.

• Consistent Attribution:
  The method ensures consistent results for similar customer journeys while adapting to changes in customer behavior.

3 Distribution across all Ad Media

This manual model assigns attribution credit to all campaigns recorded in the customer journey. Users configure weightings manually, allowing precise control over how credit is distributed across touchpoints. This approach is ideal for straightforward use cases where you want to ensure that every campaign in the journey receives a share of the credit, regardless of its position or impact.

Example: Configuration including percentage-based distribution per position for metric Order Value

• Customer Journey
  

• Analysis
  

If less than 5 Ad media are available, the weightings will be distributed.

Traffic source Weighting
In addition to the position within the Customer Journey, the campaign type can also be used for analysis purposes with traffic source weightings.

4 Distribution across the last five Ad Media

Focusing on the most recent five campaigns in the customer journey, this model distributes credit based on user-defined percentages. If fewer than five campaigns are recorded, the weightings are normalized. This model is particularly useful for scenarios where recent interactions are more influential in driving conversions.

Example: Configuration including percentage-based distribution per position for metric Order Value

• Customer Journey
  

• Analysis
  

If less than five Ad media are available, the data will be normalized.