Attribution Data-Driven

Move to data-driven attribution and discover the real contribution of your advertising partners on the generation of conversions.

What is Multi-Touch Attribution (MTA)?

Faced with increasingly complex customer journeys, comprising multiple points of contact between a brand and its prospects, it is essential to discern which elements of this journey really influence conversion.

While single-touch attribution offers a simplified view by assigning credit to a single touchpoint, multi-touch attribution (MTA) offers a more nuanced and comprehensive perspective.

Multi-touch attribution is a technique that attributes credit for a conversion (such as a sale or lead) to multiple touchpoints or interactions the customer had with the brand before converting.

She recognizes that the customer journey is often complex, with multiple interactions all playing a role in the final decision.


What are the benefits of MTA?

Multi-touch attribution allows brands to:

Understanding the customer journey: It shows which interactions have the most influence on the purchasing decision.

Optimize spending: By understanding which interactions are most valuable, businesses can allocate their budget more effectively.


What are the main MTA models?

When it comes to analyzing marketing mix effectiveness using MMM, here are the commonly examined components:

Data-driven: Uses algorithms to assign credit based on the actual effectiveness of each touchpoint.

The most commonly used algorithms in the context of marketing measurement are based on the Shapley value or on Markov chains.

Linear: Each touchpoint receives equal credit for the conversion.

Time decay: Touchpoints closest to conversion receive the most credit.

Understanding MTA models:

Shapley: Shapley value, which comes from game theory, is considered a data-driven attribution method when applied to marketing.

It is used as a multi-touch attribution (MTA) model to distribute credit for a conversion across different touchpoints or marketing channels.

Shapley value assigns a "value" to each player (in this context, a touchpoint or marketing channel) based on its contribution to the whole.

This method considers all possible combinations of channels to determine the marginal contribution of each channel to conversion. In this way, the Shapley value gives a balanced view of the importance of each channel, avoiding biases that could result from simpler attribution methods.

By using the Shapley value, brands can gain a better understanding of the relative performance of their different marketing channels, allowing them to optimize their spend and efforts.

Markov: Markov chains describe a sequence of events in which the transition probability of each event depends on one or more preceding steps.

In the context of marketing attribution, each event can represent a marketing channel (a click or an impression), and transition probabilities represent the likelihood that a customer will move from one channel to another.

The goal is to assess the relative importance of each channel in the conversion journey.

This is done by looking at transition probabilities and identifying which channels have the greatest impact on conversion.

Markov chains can help identify not only which channels directly contribute to conversion, but also which ones play a crucial role in the journey, even if they are not the final touchpoint.

So, much like the Shapley value, Markov chains offer a data-driven and more nuanced perspective on attribution compared to more traditional attribution models.


What are the challenges of the MTA?


Complexity: Implementing multi-touch attribution models can be complex or suffer from the black box phenomenon.

This is where Eulerian provides an answer, by facilitating the "click" creation of Predefined (Markov, Shapley) or Custom MTA models, with total transparency on how the algorithms manage the weights.


How does MTA compare to other marketing analysis methods?

MTA is typically deployed to advertisers who:
  • Historically drove their marketing investments using advanced single-touch models
  • Are eager to get a more reliable reading of the profitability of each marketing channel and each advertising partner
  • Have a rich media mix (many advertising partners activated throughout the year)