Attribution

Attribution in marketing refers to the process of attributing value or credit to different touchpoints in a customer’s journey, in order to understand the impact of marketing and advertising efforts on customer behavior and conversion. It helps to optimize and allocate marketing budgets by identifying the most effective channels and tactics.

What is attribution?

Marketing attribution is a scientific method to evaluate the efficacy of your digital marketing efforts, by determining which touchpoints your consumers are encountering and how those contribute to sales or conversions. 

There are several different attribution models used by marketers to attribute customer actions to particular marketing activities or content. Attribution modeling is an excellent way to understand the consumer’s journey and determine what’s working and where you can improve. It also shows you how various marketing channels are working together and where you should focus more effort and time in order to achieve your KPIs. 

How to measure attribution

Marketing attribution is typically measured using certain models. These models allocate a weight or value to different parts of your marketing campaign and use those to determine which of your efforts are more or less effective. There are a number of models available for you to use, and each comes with its own opportunities and challenges. 

Choosing the right model depends on what you’re trying to measure, and why, and it’s important to do your research before relying too heavily on a chosen model to give you results.

What is an attribution model?

A marketing attribution model is a set of rules that determines which touchpoints in your marketing campaigns should get the credit for a signup or conversion (or measuring brand awareness, if that’s your goal). The best models are those that will give you insight into what channels consumers have been exposed to, what messages they saw there, and what touchpoint(s) most impacted their decision to take action (such as making a purchase). A good attribution model also indicates how much impact – if any – the customer’s perception of your brand had on their decision to take the action, and even how much impact external factors had on their decision. 

What are the main attribution models?

As mentioned earlier, there are a few different attribution models you can use, generally broken down into two types: single-touch and multi-touch attribution models.

Single-touch attribution models

A single-touch attribution model assigns 100% credit to a “single” marketing touchpoint. This means that your customer may have seen three different YouTube pre-rolls before they decided to click on the website and make a purchase, but with the single-touch model, only one of those pre-rolls will get the credit for the conversion. 

Single-touch is usually divided into two classes: first-touch and last-touch. 

In first-touch, the model makes the assumption that a consumer chose to take action after the first touchpoint (for example, the first time they saw your ad online).

Last-touch, on the other hand, gives full credit to the final touchpoint before the action. It makes the assumption that the customer made a decision to buy or convert after interacting with the last touchpoint they were exposed to.

This model comes with some pretty obvious setbacks, like not being able to see the full journey a consumer takes. That’s why most marketers recommend using other methods in conjunction with this one, to get a more balanced view of where their marketing efforts are working. Of course, if your funnel is very simple, or you’re trying only to see what happens at the top or bottom of the funnel, this could be very useful.

Multi-touch attribution model

Multi-touch models look at every touchpoint a consumer has engaged with from the first to the last (that is, until they make their purchase or convert). These models are usually considered to be more accurate than single-touch and assign value, or weight, to each touchpoint differently.

These models are primarily differentiated by how they divide up the credit. Here are the four most common multi-touch attribution models:

• Linear: Divides credit between messages and touchpoints equally. It’s easy to understand and great if you only have a few marketing channels. 
• U-shaped: Allocates a lot of credit to both the first and last touchpoints, then divides up the rest across the touchpoints in the middle. This creates a U-shape, which the model is named for. This is a great model to use when you already understand your customers’ journeys and you’re trying to find ways to optimize the first and last interactions. 
• Time decay: In this model, any touchpoints that are closer to the point of conversion get more weight. This is because the model assumes that an earlier interaction has less impact on a sale. 
• W-shaped: Using the same kind of idea as the U-shaped model, this model comes with one extra stage – the opportunity – right in the middle. In this model, the three peaks, namely first touch, opportunity, and last touch, are each allocated 30% of the credit. The 10% that’s left is divided across any of the other touchpoints.

How to choose an attribution model

Deciding on the right attribution model depends on many factors from the type of business you have to what you want to learn about your customers. There is no “best” model to choose but there are a few pointers you can use if you’re picking your first model (and remember, you can – and should – change your model as your business grows).

First, consider how many stages there are in your funnel. Do you include lead generation, lead nurturing and sales? How about brand recognition? Your business goals will help you pinpoint how many stages your marketing funnel has, if you’re not 100% clear on them already.

Next, think through how many marketing touchpoints you have. No touchpoint is too small, so take note of all of your marketing efforts from targeted emails to online ads, offline marketing, content recommendations, and even social media. 

Also, a point to consider is what you want to achieve. Some businesses may be at a point where conversion is important while others just want to encourage sign-ups. You may even want to boost your brand awareness to keep your business name top of mind. 

It’s also valuable to consider any factors that are out of your control, and which may impact your results. Increased interest rates will definitely impact real estate and the sale of “nice to have” items, while a supply chain shortage could cause increased demand for your product. Ultimately, choosing the best attribution model depends on what you need. 

The challenges of attribution 

Attribution models come with certain challenges that should be accounted for. Though there are ways to handle these, be aware of them when you choose your model:

• Attribution data is not ‘exact’ data. Attribution is not a pure science and though it can give you some very helpful data, it can’t take everything into account. For example, you may have an existing customer base from which you earn recurring revenue, and the model can’t account for that. There are also some offline activities that a model simply can’t track. This means what you see in the model is not always based on reality. 
• Attribution can’t take offline-to-online into account. This is best understood with an example. Let’s say you run a car dealership with billboards all along a local highway. You also have employees distributing flyers at local malls and you sponsor a local event. Your attribution model simply cannot account for those marketing efforts. It doesn’t know how many of those consumers converted based solely on your online efforts, and how many saw and engaged with your offline efforts too. 
• Attribution can’t account for external factors. This goes beyond just your marketing efforts. External factors could include almost anything from natural disasters and pandemics, to cost-of-living increases, supply chain shortages, and even social media trends. 

So what’s the bottom line? Attribution models are not built equally and cannot take everything into account. This means the results you get from them will be good, but they’ll never be perfect. 

What you can do to mitigate this is to test and compare your results. A/B testing is a really smart way to help you narrow down how consumers are interacting with your brand. And the more you test, the more you’ll know which attribution model is right for you, so you get to know exactly the impact your marketing efforts are making.