Attribution Models

An In-Depth Look at Attribution Modeling in Digital Marketing | by Bill Su  | Analytics for Humans | Medium

Attribution modeling is an important step in the checklist for making data-backed decisions. It is essential to utilize appropriate attribution modeling for what you want to accomplish. Attribution modeling is an analysis tool, and you can use it to compare performance under each model to gain a better understanding of multiple touchpoints in the customer’s journey. This information can help you better gauge where your business should focus its efforts to reach customers. Some of the common attribution models include first interaction, last interaction, linear, time decay, and position based.

First Interaction Attribution

Attribution Models: What is attribution modeling and which model should I  use? - AgencyAnalytics

First interaction, also called the First-Click, rewards all the credit for a conversion to a business’ first interaction with a customer. This is similar to the last interaction which gives credit to a single click/interaction. For example, if a customer learns about your business on Instagram, then Instagram gets all the credit for any sales that occur after that first interaction.

Last Interaction Attribution

Attribution Models: What is attribution modeling and which model should I  use? - AgencyAnalytics

Last interaction, also called the Last-Click, rewards all the credit for a conversion to a business’ last interaction with a customer. This is similar to the first interaction which gives credit to a single click/interaction. For example, a visitor does a Google search about your business. A week later, they see an ad on Instagram and click the ad. Later that day, they go to your website directly and purchase something. In this scenario, direct traffic would receive the credit.

Linear Attribution

Attribution Models: What is attribution modeling and which model should I  use? - AgencyAnalytics

With linear attribution model, it gives credit to everyone involved. This model splits the credit for a conversion equally between all the different interactions that a customer had along the way involving your business. For example, a customer finds you on Facebook, signs up for your email list and later clicks on the email link. The following week, they go directly to your site and make a purchase. Each touchpoint gets equal credit for contributing to the customer’s purchase.

Time Decay Attribution

Attribution Models: What is attribution modeling and which model should I  use? - AgencyAnalytics

Time Decay attribution is similar to linear attribution in that it gives some credit to all the channels that led to the conversion. However, the amount of credit being less (decaying) the further back in time the interaction took place. This model takes into consideration the time frame that each touchpoint occurred. Interactions that took place closer to the time of the purchase have more value attributed to them than those that took place further away.

Position Based Attribution

Attribution Models: What is attribution modeling and which model should I  use? - AgencyAnalytics

The position based attribution model, also called the U-shaped model, gives credit for the sale to the first touchpoint and the last touchpoint. Both the first and last touchpoints receive 40% of the credit while the remaining interactions receive the remaining 20%, divided equally. This model views the first and last touchpoints as important in driving sales and attributes more than a single interaction for the success.

Takeaways about Attribution Models

  • The attribution modeling measures the impact that each channel has on conversions across the consumer’s journey
  • It is key to determine which attribution model best fits your organization
  • Each model has its own unique benefits and shortcomings

To explore more about the different attribution models, check out these two sites About attribution models – Google Ads Help and Types of Attribution Models – How and When to Use Them? | LeadsRX

Leave a Reply

Your email address will not be published. Required fields are marked *