Understanding the Multi-Touch Attribution (MTA) for digital projects
Multi-touch attribution (MTA) is a method used in digital marketing to assign value or credit to different touchpoints along a customer's journey, from initial awareness to final conversion (like a sale).
Multi-touch attribution recognizes that several interactions across different channels and touchpoints contribute to a user's decision to take a desired action.
The choice of attribution model has profound implications for digital marketing budget allocation and strategy, it's crucial to select a model that aligns with business goals, customer behaviour, and the nuances of the specific industry or market.
Last Click and First Click Distribution
Last-Click Attribution:
Definition: In this model, 100% of the credit for a conversion is given to the last touchpoint the customer interacted with before converting.
Use Case: Let's say a user saw a Facebook ad, clicked on a Google ad a week later, read an email newsletter from the brand, and finally made a purchase after clicking on a retargeting ad. In the last-click attribution model, the retargeting ad would receive all the credit for the conversion, as it was the final touchpoint before the user made a purchase.
Pros:
Simple and easy to understand.
Helps identify the final push that leads to conversions.
Cons:
Ignores all previous interactions, potentially undervaluing their impact.
This might lead to an overemphasis on retargeting and remarketing efforts at the expense of other touchpoints that help build awareness and consideration.
First-Click Attribution:
Definition: This model attributes 100% of the credit for a conversion to the very first touchpoint the customer interacted with in their journey.
Use Case: Using the same example, if a user first discovered the brand through a Facebook ad, then that Facebook ad would receive all the credit for any subsequent conversion, even if the user interacted with other marketing channels afterwards.
Pros:
Highlights the channels that are effective at introducing and creating awareness about a brand or product.
Useful for businesses focused on expanding their reach and attracting new potential customers.
Cons:
Overlooks the influence of subsequent interactions that might have nurtured the lead or pushed them towards conversion.
This might lead to an overemphasis on broad awareness campaigns and neglect of efforts that drive conversions.
In my point of view, while these models offer insights, they provide a limited view of the customer journey. This is why more advanced models like multi-touch attribution have been developed to give a more holistic understanding of how different touchpoints contribute to conversions.
Some common types of multi-touch attribution models:
Linear:
This model distributes the credit equally among all touchpoints. So, if a customer interacts with five touchpoints before buying a product, each touchpoint receives 20% of the credit for the sale.Time Decay:
Gives more credit to touchpoints that occurred closer to the conversion. For instance, if a customer saw an ad a month ago and then another ad yesterday before buying today, the ad from yesterday would get more credit.U-Shaped (or Position-Based):
Assign more credit to the first and last touchpoints and distribute the remaining credit equally among the other touchpoints. A common distribution might be 40% to the first touch, 40% to the last touch, and the remaining 20% distributed among other interactions.W-Shaped:
This is a variation of the U-Shaped model but includes an additional significant touchpoint, often the point where a lead becomes a marketing-qualified lead (MQL). Credit might be distributed as 30% for the first touch, 30% for the MQL touch, 30% for the last touch, and the remaining 10% divided among other interactions.Algorithmic (or Data-Driven):
Uses advanced analytics and machine learning to assign weights to each touchpoint based on how influential they are in driving conversions. This model is dynamic and can change based on data inputs.
The purpose of multi-touch attribution is to provide marketers with a more holistic and accurate view of how their marketing efforts are performing by understanding the contribution of each touchpoint, businesses can allocate their marketing budget more effectively, optimize their campaigns, and improve overall marketing ROI.
Key points of Multi-Touch Attribution:
Holistic View: MTA considers all touchpoints, not just the first or last one.
Better Allocation of Resources: By understanding which touchpoints are most effective, marketers can allocate resources more efficiently.
Complexity: MTA models can be complex, as they require data from multiple sources and sophisticated analytics to assign value to each touchpoint.
Data Requirements: Successful MTA requires a robust data infrastructure to track customer interactions across all channels.
Different Models: There are various MTA models, such as linear attribution (which assigns equal credit to each touchpoint), time decay (which gives more credit to touchpoints closer to conversion), and algorithmic models (which use machine learning to determine the value of each touchpoint).
Scenario
Imagine you sell luxury watches online.
A potential customer, named Alex, first discovers your brand when he sees a Facebook ad for one of your watches.
A few days later, he searches for reviews of your watches and clicks on a Google search ad that leads to your blog.
The next week, Alex receives an email newsletter from you showcasing a new collection.
Finally, after a couple more days, Alex sees a retargeting ad on a news website, clicks on it, and makes a purchase.
Touchpoints are:
Facebook ad
Google search ad leading to your blog
Email newsletter
Retargeting ads on a news website
Attribution Models e Distribution Calculation:
Last-Click Attribution:
This model would give 100% of the credit to the retargeting ad on the news website, since it was the last touchpoint before the purchase.
All credit (100%) is given to the last touchpoint before conversion.
First-Click Attribution:
100% of the credit would go to the Facebook ad, as it was the first interaction Alex had with your brand.
All credit (100%) is given to the first touchpoint.
Linear Attribution (a type of MTA):
Credit is distributed evenly across all touchpoints. So, each touchpoint (Facebook ad, Google ad, email newsletter, retargeting ad) would receive 25% of the credit for the conversion.
Credit is distributed evenly across all touchpoints.
Time Decay Attribution (another type of MTA):
Touchpoints closer to the conversion get more credit. In this case, the retargeting ad might receive 40% of the credit, the email newsletter 30%, the Google ad 20%, and the Facebook ad 10%.
This is a bit more complex as credit is given more to the touchpoints closer to the conversion. A simple way to do this is by using an exponential decay function. For simplicity, let's assume that the touchpoint closest to conversion gets 40%, the next one gets 30%, then 20%, and the first touchpoint gets 10%.
Formula: Predetermined percentages based on proximity to conversion.
Same conversion same scenario
Is quite important to understand our budget can be differently used depending on the method we calculate the attribution.
In the following example, I gathered conversion information from an e-commerce project. Numbers are expressed in percentages and single conversion.
The pie charts provide a clear visual of the proportion of conversions each touchpoint receives under different attribution models.
For instance, you can observe that the "Retargeting ad" touchpoint claims a large portion in the Last Click model, while the "Facebook ad" touchpoint dominates in the First Click model.
Here is the data breakdown:
As shown, this table provides a numerical breakdown of the conversions attributed to each touchpoint under the various attribution models; as you can see, analyzing this data there is an important difference in the effectiveness of conversion touchpoints across multiple attribution models.
Attribution models play a crucial role in determining how marketing budgets are allocated, and each model can lead to a different interpretation of the effectiveness of various marketing channels.
Challenges & Considerations:
While the ideal scenario is to use a model that best represents the actual customer journey, it's essential to recognize that no model is perfect.
Real-world decisions should be based on a combination of data insights, industry best practices, and business objectives.
It's also worth noting that frequent shifts between models can lead to inconsistent strategies and budget allocations, potentially confusing customers and stakeholders.
Multi-touch attribution recognizes that several interactions across different channels and touchpoints contribute to a user's decision to take a desired action, but my suggestion these are the reasons for not using just a single model but deciding a unique standard for each single project.