In this article, we will explore the topic of marketing attribution models. Which are a way of measuring the impact of different marketing channels and campaigns on customer conversion. The models can help understand which marketing efforts are most effective in driving conversions and optimizing the marketing mix. They can also help improve return on investment and the allocation of budget and resources more efficiently.
However, it’s not easy to set up and use. They require a lot of data, analysis, and experimentation. In this article, we will explain what marketing attribution models are? how they work? and how you can use them in your digital marketing strategy? In addition, provide some best practices for using different types of attribution models. Such as single source, multi-source, weighted multi-source, and custom attribution. We hope you find this newsletter useful and informative.
What is an Attribution Model?
It is a way of assigning value to the different marketing touchpoints that a customer interacts with before converting.
A touchpoint is any point of contact (or interaction) that a customer has with a brand. Such as an ad, an email, a social media post, a website visit, etc.
A conversion is any desired action or outcome that a customer performs. Such as a purchase, a sign-up, a download, etc.
A conversion path is the sequence of touchpoints that a customer follows before converting.
For example,
A customer may see an ad on Google. Clicks on it. Visits the website. Signs up for the newsletter. Receives an email offer, and then makes a purchase. This is a conversion path with six touchpoints.
The main difference between single source and multi-source attribution models is that single source attribution models assign all the value to one touchpoint (either the first one or the last one that the customer encountered prior to converting). While multi-source attribution models distribute the credit or value among multiple touchpoints, based on different rules or criteria. The pros and cons of each type of attribution model will be discussed in the following sections.
Single Source Attribution
The simplest and most common type. It assigns all the value to one marketing touchpoint, either the first or the last the customer encountered before converting. The first touchpoint is the one that did the introduction to the brand (or product). While the last is the one that persuaded the customer to purchase.
For example,
If a customer sees an ad on Facebook, clicks on it, visits the website, and then buys the product. The first attribution model will give all the credit to the Facebook ad (first touchpoint). While the last attribution model will get all the credit to the website.
The advantage of single source attribution models is that they are easy to implement and understand, as they only require one data point and one rule.
They can also help to identify the most effective channels or campaigns for generating awareness or closing sales. However, the disadvantage of single source attribution models is that they ignore the influence of other touchpoints along the customer journey, and they may overestimate or underestimate the value of certain channels or campaigns. They may also fail to capture the complexity and diversity of customer behaviour and preferences. For example, some customers may have multiple touchpoints before converting, while others may have none. Some customers may have a long and nonlinear conversion path, while others may have a short and linear one. Some customers may be influenced by different factors at different stages, while others may be consistent throughout.
Multi-Source Attribution
This is a more advanced and sophisticated type of attribution model. It distributes the value among multiple marketing touchpoints, based on different criteria. The criteria can be based on the position, the time, or the importance of the touchpoints in the conversion path.
For example,
A multi-source attribution model can give more credit to the first, the last, or the middle touchpoints, or to the touchpoints that are closer or farther away from the conversion, or to the touchpoints that have a higher or lower impact or contribution to the conversion.
There are several types of multi-source attribution models, such as linear, time decay, U-shaped, W-shaped, and Z-shaped attribution:
Linear Attribution
This model gives equal credit to all the touchpoints that the customer interacted with before converting. This model is fair and comprehensive, but it does not account for the varying importance of different touchpoints.
The advantage of the model is fairness and comprehension. As it accounts for the influence of all the touchpoints along the customer journey. It can also help identify the most effective channels to maintain and nurture the customer relationship, as it recognizes the value of each touchpoint.
However, the disadvantage of a linear attribution model, it does not account for varying importance (weight) of the different touchpoints, as a result diluting the value for some. It may also fail to capture the nuances of customer behaviour and preferences at different stages of the customer journey.
Time Decay Attribution
This model allocates more value to the touchpoints that are closer to the conversion and less to ones farther away. This model reflects the recency effect, but it may undervalue the role of early touchpoints in creating awareness and interest.
The advantage of time decay attribution models is that it reflects the recency effect. Which is the tendency of customers to remember the most recent information or experience more than the earlier ones.
It can help identify the most effective channels for persuading and closing the customer, as it recognizes the value of the last touchpoints.
However, the disadvantage of the time decay attribution model is it undervalues the role of the early touchpoints. Which may have created the awareness and interest in the customer. It can overestimate the role of the last touchpoints, which may have been only a final trigger. It may also fail to capture the complexity of customer behaviour. As some customers may have a longer or shorter conversion path, or influenced by different factors at different times.
U-shaped Attribution
This model allocates 40% of the value to the first touchpoint, 40% to the last and 20% to the touchpoints in between. The advantage of U-shaped attribution models is that they emphasize the importance of the first and last touchpoints. Which are often the most critical ones for generating and closing leads.
However, the disadvantage of U-shaped attribution models is that they may overlook the influence of the middle touchpoints, which may have played a significant role in nurturing the customer. Depending on the length and complexity of the conversion path.
W-shaped Attribution
This model allocates 30% of the value to the first touchpoint, 30% to the last touchpoint, and 30% to the touchpoint that marks the transition from lead to opportunity. The remaining 10% is distributed among the other touchpoints. It recognizes the key milestones in the customer journey, but it may not capture the nuances of different touchpoints within each stage.
Z-shaped Attribution
This model gives 22.5% of the credit to the first touchpoint, 22.5% to the last, 22.5% to the transitional touchpoints (from lead to opportunity), and 22.5% to the touchpoint that converts (from opportunity to customer). The remaining 10% is distributed among the other touchpoints. This model covers the entire customer journey, but it may be too complex and rigid for some scenarios.
The advantage of Z-shaped attribution models is that they cover the entire customer journey. From awareness to loyalty. They recognize the key milestones in the customer journey, which are often the influential ones for creating, nurturing, and converting.
However, the disadvantage of Z-shaped attribution model is that it can ignore the influence of the other touchpoints. Which may have played a significant role in reinforcing the customer decision.
Weighted Multi-Source Attribution
This type assigns different weights to different touchpoints, based on their relevance, impact, or contribution to the conversion. The weights can be determined by using statistical methods, machine learning algorithms, or human judgment.
For example,
If a customer sees an ad on Facebook. Clicks on it. Visits the website. Signs up for a newsletter. Receives an email offer. Requests a demo. Then makes a purchase. The weighted multi-source attribution model can give 10% of the value to the first touchpoint. 20% to the second. 15% to the third. 25% to the fourth. 10% to the fifth. 15% to the sixth, and 5% to the last touchpoint. These weights can be based on the historical data, the predictive model, or the expert opinion of the marketer.
The advantage of weighted multi-source attribution models is that they are flexible and customizable. They allow marketers to create their own attribution model, based on their specific objectives, data, and assumptions. They can also help to account for the importance of different touchpoints, and to reflect on the complexity of behaviour.
However, the disadvantage of weighted multi-source attribution models is that they require more data, resources, and expertise to implement. They involve some subjectivity and uncertainty in determining the weights. They may also vary depending on the type and size of the business, industry, market, or customer segment.
Custom Attribution
This is a type of multi-source attribution model. Allowing marketers to create their own attribution model, based on their specific objectives, data, and assumptions.
The custom attribution model can use any combination of rules, criteria, weights, or methods to assign value to different touchpoints. Depending on the needs and preferences of the marketer.
For example,
If a customer sees an ad on Facebook. Clicks on it. Visits the website. Signs up for a newsletter. Receives an email offer. Requests a demo, and then makes a purchase. The custom attribution model can allocate 15% of the credit to the first touchpoint. 25% to the second. 10% to the third. 20% to the fourth. 5% to the fifth. 15% to the sixth, and 10% to the last touchpoint. These weights can be based on the marketer’s own judgment, experience, or intuition.
The advantage of custom attribution models is that they are the most flexible and customizable type of attribution model. They allow marketers to work freely with the available data and previous knowledge.
However, the disadvantage of custom attribution models is that they require the most data, resources, and expertise to implement. They involve a lot of subjectivity and uncertainty in creating and testing the model. They may also be difficult to compare and validate, as they may not follow any standard or common framework or methodology.
Tagging & UTM
Now that we have discussed the different types of attribution models and how they work. We can use tagging and UTM (Urchin Tracking Module) parameters to track the performance of different digital marketing channels. Tagging and UTM are snippets of text added to the end of a URL to track the journey. They can contain up to five parameters: source, medium, campaign, content, and term. These parameters help in identifying the origin, channel, name, variation, and keyword of your campaign.
For example. To track how many visitors come to a website from a Facebook post. The UTM tag can be included to the URL, as follows:
[https://www.yourwebsite.com/?utm_source=facebook&utm_medium=social&utm_campaign=summer_sale]
This will tell Google Analytics that the visitor came from Facebook (source). Through a social media channel (medium). As part of a summer sale campaign (campaign).
UTM tags can be created manually, or with the use of tools like Google’s Campaign URL Builder, Google Tags or HubSpot’s Free UTM Generator to generate them easily. You can then use these tagged URLs in your ads, emails, social media posts, or any other digital marketing channels.
Tagging and UTM can help to improve the accuracy of attribution models, as they allow tracking and measuring performance of different digital marketing channels and campaigns, and to link them to the conversion path and the conversion.
It can also help avoid double counting or missing out on any touchpoints, as they provide a consistent and standardized way of identifying and reporting the touchpoints. Tagging and UTM can also help when comparing the results and insights from different attribution models, as they provide a common and clear basis for assigning value to different touchpoints.
Conclusion
Marketing attribution models are a powerful tool for digital marketers, but they are not a silver bullet. They are only as good as the data and assumptions that underlie them, and they should be used with caution and creativity. They are also constantly evolving and improving, as new technologies and methods emerge. Therefore, we encourage you to keep learning and experimenting with attribution models, and to share your feedback and results with us at info@mena-review.com.
See you in the next issue 😊