Most marketers use different channels to communicate the same message. In today’s digital marketing it is very common to run search ads, send an email to an existing database, run display ads for remarketing, write a blog post on your website and post on social media websites. However, the challenge comes from analyze your marketing performance to understand what marketing initiative drove your final conversion which could be either a sale or for B2B marketers an inquiry or form submit. In order to solve that problem, marketing analysts have come up with an attribution model.

What is an Attribution Model:

Attribution model is a model which determines what marketing channel or touch point gets credit for the sale. Typically, it is using a rule or a set of rules to provide the credit to the right source or sources. While providing marketing analytics consulting to a company, it is usually the first task in setting up your conversion tracking and before analyzing sources of data.

Why should you be using an attribution model?

Without an attribution model it is impossible to understand what works and what does not in marketing. For example, you can be analyzing email marketing campaigns by looking at open rates of your emails. If the trend is that open rate is increasing you may think that your email marketing is working. However, if you were to attribute sales from your email marketing you may realize actually the amount of sales that are generated from email does not justify the effort and any cost associated with it.

Another benefit to having proper attribution by source of conversion, is the ability to compare sources and optimize your marketing spent, calculate ROI and quantify your marketing efforts to upper management and other departments.

Finally, not all sources are made equally so make sure you choose the right model for you. Some sources are great at finalizing a conversion. While other sources may create awareness, but they are either hard to measure in terms of magazine ads or TV ads. Therefore, make sure you pick the right model for you as not all models are created equally.

What are the different attribution models and their benefits?

There are four different attribution models that you can choose from:

  1. Last Touch Attribution Model

Last Touch is easy to calculate as the last marketing touch before the conversion is being made gets 100% of the credit. For example, if someone has seen an ad on Adwords, signed up for a newsletter and three weeks later received an email advertising the product that they originally saw. After clicking on an e-mail they make a purchase. Email marketing would get 100% of the credit, before email marketing was the final marketing channel that the customer has taken.

The benefits of using this model:

  • Is that you truly understand what channels help drive the final sale or form submit (conversion).
  • Easy to implement.
  • Great for companies that have low margins and are focused on sales only.

Because it is only a part of the journey, it is not an ideal model to use for sophisticate marketers.

  1. First Touch Attribution Model

First touch attribution model is the exact opposite of last touch. The source that made the prospect aware of your brand or company is the channel that receives 100% credit for the sale or conversion. Recall our previous example, where the prospect clicked on AdWords but purchased a product a few weeks later after an email was sent. Using first touch attribution model AdWords would receive 100% of the credit. As in last touch attribution model it only looks at one channel at a time, making other channels seem less

The benefits of using this model:

  • Great at understanding what channels help drive customer’s awareness.
  • Easy to implement.
  • Ideal for companies that are focusing on driving traffic to website and not on lead conversion or eCommerce.
  1. Equally Distributed Attribution Model

Also known as liner attribution model, this model takes into account all channels that contributed to a conversion. In our example of both AdWords and Email channels being used to help generate a sale, 50% of the sale would go towards email marketing and 50% towards AdWords.  While this model is a bit harder to implement, it helps justify all your marketing efforts.

The benefits of using this model:

  • All marketing efforts are accounted for.
  • Easy to implement as there is no sophisticated math required.

However, the flaw of this model is that it treats all channels as equals. Not all channels are created equally someone liking a Facebook post may have less impact than someone clicking on a link in an email.

  1. Weighted Attribution Model

Using algorithms or creating custom rules created a weighted attribution model. One way to create it is to assigned different values based on length of time between the original source and until the final conversion took place. Another way is to assign a higher percentage to the last and first source and a smaller attribution to all the marketing touches in between. Another way is to put a marketing bias based on each channel, while it is less mathematical it could take into account marketer’s logic.  Finally, it could be a model which takes different elements and creating a custom model. For example, you may decide that the last touch always gets 50% of the attribution model no matter what, while, the other channels gets attribution based on length of time from the purchase, and if the channel is x it would get only 50% of the attribution that the model has predicted. While, it may sound confusing, hard to implement and may require changes overtime, the weighted attribution model is the most accurate model that takes into account the whole customer journey and not just parts of it.

So which model is right for you?

While, you may decide that weighted attribution model is the best model to use, it may not be right for your business. First attribution or last attribution models may be sufficient for your business. Weighted custom attribution model maybe critical for businesses that are running many different marketing efforts and really need to understand their marketing spent due to budget constraints. Keep in mind that while mathematically a model maybe more accurate, it may have no impact on your business but can increase your cost of implementing the attribution model. Another thing to keep in mind is that while some channels do not contribute to sales, they can increase engagement and overall brand awareness. The key is not what model you use but rather being consistent in your analysis.

Happy Analyzing!