Marketing attribution models, types, differences, and how to use them

In the ideal world of a marketer, a new user finds their website or clicks an ad, and immediately converts into a valuable lead or makes a purchase decision. 

Unfortunately, the reality is way too different. Users go through a buyer's journey and sales funnel, are targeted through various channels, and are subject to remarketing and retargeting before they are willing to make a decision. The question is, “Which digital marketing channel gets the credit and which tactics have been more persuasive?” Attribution modeling, used by marketing researchers to answer those questions, is a statistical technique to determine what influences consumers' buying decisions. Let’s dive into the world of attribution modeling. 

What is Marketing Attribution?

To explain marketing attribution, we will take a short and long way. 

The short definition: marketing attribution is the science of determining those marketing channels and tactics that contribute to conversions/sales.

The long definition: marketing attribution is the practice of evaluating the marketing touchpoints users experience on the buyer journey. The goal of this practice is to identify which channel(s) and message(s) have the greatest impact on customer decision-making. Popular attribution models are widely used by marketers to effectively get insights and optimize marketing campaigns based on marketing attribution. The goal is to meet customers’ desires and needs through marketing and improve marketing ROI. 

Why does Marketing Attribution matter?

Available attribution models are utilized to track consumer data from different channels and weigh each interaction’s influence on the buyer’s journey. Building marketing strategies around attribution is a robust start to digital marketing efforts aimed at saving resources. 

An example of attribution

Let’s take users A and B. They are both seeking a Z company’s product. Customer A converts in one step since they landed on your website knowing exactly what they needed and your site was able to supply it.  

Customer B, on the other hand, has a totally different journey before converting. Let's assume Customer B initially saw your company's display ad, then interacted with the brand on social media before visiting the company's website through an organic search listing, searched for services/products, and then made a purchase a week later.

As you can see, client B's user journey had numerous phases, or touchpoints, spread out across time.

With the attribution model, both conversion touchpoints are revealed and treated properly. In short, attribution modeling helps marketing efforts stop throwing spaghetti at the wall to see what sticks and instead enables them to work smarter, not harder, by harnessing the power of ROI data.

Here is one more example of a customer journey.

The customer's journey

The Benefits of Marketing Attribution

Aside from providing valuable insights about buyers’ journeys and conversions, marketing attribution has a whole bunch of benefits for helping businesses use time and resources more productively. 

Increased ROI: Effective attribution models or models chosen to serve business needs will help reach the right consumers at the right time, eventually bringing high marketing ROI.

Optimized marketing spend: Having effective marketing campaigns providing conversions means spending each cent with purpose and earning engagements. Marketing attribution will then help better optimize low-performance processes. 

Enhanced personalization: Understanding your customer's needs and expectations is the most vital piece of knowledge marketers strive to acquire. With attribution models and data about touchpoints, it is possible to more effectively target customers with actionable tactics.

Optimized creative: Marketing attribution assesses conversion rates and the content your customers encounter when interacting with your product or service. When the content is not well-tailored, the campaign will not have the expected outcome. The marketing attribution will help change and optimize the tone of voice, visual elements, and content to better communicate with users. 

Data-driven product improvement: Again, the attribution model works on understanding your audience; whether they like your product or not, and specifically, what is the message that plays a crucial role in their decision-making leading to improvements that can increase engagement and features to be added.

Biased attribution: Common Marketing Attribution Challenges

While attribution models have proven their efficiency in marketing efforts, there are common challenges the marketing team may face while running a marketing campaign. One of the reasons why marketing attribution may not bring results is biased attribution. It is the conflict of interests that may occur when attribution is implemented by one and the same measurement platform that provides traffic and conversions. To avoid biased attribution, it is important to use a third-party attribution provider separate from measurement tools, thus avoiding biased results and ensuring the attribution model will improve rather than hinder success.

In-market bias: The most common mistake of the marketing team is investing resources in attribution models to track those customers that were in the market and willingly purchased the product regardless of whether they had seen an ad or not. In this scenario, the ad will still get attribution although, in reality, it had no impact on a consumer. Check out the latest case study based on the performance of marketing attribution, where a similar conclusion of in-market bias was made. 

Correlation-based bias: The bias refers to the distorted assumption that one event causes another one in the customer journey, when, in reality, the event had no impact on conversion.

Digital signal bias: This biased attribution doesn’t take into account offline activity (applicable to those businesses offering online and offline services), leading to a bias towards online sales. 

Cheap inventory bias: This refers to low-priced products that eventually lead to conversions. Marketing attribution, showing better-performing campaigns, maybe just the result of low prices. 

Types of Attribution Models

There are several attribution models available, such as AI-based software, Facebook's own attribution tool, and regular Google Analytics attribution models. They are, however, classified into two types: single-touch and multi-touch.

Single-touch Models

Single-touch Marketing Attribution Models

As the name suggests, single-touch attribution models assign credits to one touchpoint on the customer journey. In its turn, single-touch attribution is divided into several types. Before diving into the details, let’s take an example of a single-touch model.

Suppose your company sells bags. Your potential customer sees an ad on Instagram and visits your page. The next day, the user visits your website, surfs for a while, and leaves. It takes a week before the user is retargeted on Facebook, clicks the link, and makes a purchase. If you attribute 100% of the credit to a single touchpoint, you will focus on one of the touchpoints while the user has at least 4 interactions in the consumer buying process. How would you define the touchpoint for decision-making? 
 

• First Touch Model (first click attribution)

The first-click attribution model gives credit only for the first customer interaction with your business. It is an effective way to test what attracts new customers’ attention at the top of the funnel. This model has proved efficient in longer sales cycles for building a loyal customer base rather than driving one-time purchases. 
 

• Qualified Lead Model (Last Touch)

Unlike first-touch attribution, the qualified lead model gives all the credit to the last interaction between a customer and your business before making a purchase. Being the earliest marketing attribution model, it gradually fades into the backdrop of other attribution models. However, completely ignoring this model is not the best thing to do. It helps to understand which marketing channels drive customers to action.
 

• Lead Creation Model

In the lead-creation model, most efforts are put into the distinction between prospects and leads. It means that 100% of the credit is given to the milestone when a prospect becomes a qualified lead. It helps to determine the critical touchpoint at which the prospect converts.
 

• Last non-direct click Model

This attribution model gives credit to the customer’s last non-direct touchpoint, i.e., non-direct traffic from other sources directed to your website. Non-direct traffic sources can be social media posts, email campaigns, influencer campaigns, guest posts, etc.
 

• Last < most important touch > Model

The customizable last <most important channel> touch model is applied when the company has already determined which channel is the most important, with 100% of the credit given to the chosen channel responsible for all the sales.
 

Ready to test attribution models for your business? 

Although it might seem complicated at first, building the right marketing data infrastructure is an excellent place to start so that you can track every interaction your visitors have with your website, which will serve as the basis for future development. We can help to start with your attribution modeling.

Multi-Touch Models

Multi-touch Marketing Attribution Models

Customer journeys are complex, just like their purchase intentions. There is no single best approach when it comes to multi-touch attribution modeling. Depending on the business and customer profile, even having one efficient model may not be enough to map out customer journeys and determine the importance of all the touch points. That is why there are several multi-touch attribution models, and depending on which one you choose, the value will be assigned to channels in a different way.
 

• Linear Model

In this model, the credit is divided equally among all the touchpoints in the customer journey, no matter how many touchpoints are involved in the conversion path. This is probably the most straightforward approach to start from when you have set up the data infrastructure and can attribute different touchpoints to a single customer, but it lacks a lot of important nuances. 
 

• Time Decay Attribution

The model is in some ways similar to the previous one. Each touchpoint gets some credit for the conversion, but the closer to the last touchpoint, the greater the credit. In simple words, you lead the customer, and the more he is interested in you, the more you invest in converting him. In this approach, the customer is not always "convinced" that the product is worth purchasing. In fact, the last touch points and some re-targeting efforts might even be useless or even have the opposite effect on the customer’s purchase intent. 
 

• U-shaped Attribution

Also called position-based attribution, this model divides credit for conversion between the first interaction and the moment when the customer converts into a lead. Each of these touchpoints gets 40% credit. The rest of the 20% is shared among all the interactions that may take place between the first and last interaction. This approach is a consensus between the first-touch and last-touch advocates and may or may not work for a business depending on the customer journey. 
 

• W-shaped Attribution

The model splits credit between three touchpoints—the first, the last, and the qualified-lead milestone. Each of them gets one-third of the attribution credit. While this might be an appealing approach, it’s important to note that all the other touchpoints and efforts are fully neglected. 
 

• Z-shaped Attribution

The Z-shape is the one that divides credit among all the touchpoints, deciding the amount of credit depending on its position in the sales funnel. The four most critical touchpoints are First Touch, Lead Creation, Qualified Lead, and Customer Won, which get 22.5% credit each. The remaining 10% is shared among other touchpoints. 

How to choose the right attribution model for your business: Final Thoughts

Implementing a proven attribution model that was successful for one business doesn’t guarantee success for your business. Every service or product has its own target audience, with different conversion models and sales cycles. For example, you cannot compare a marketing campaign for a luxury car with a campaign for shoes. 

The choice of an attribution model should be based on extensive research of your customer's intent to identify the ins and outs of their buying journey.

• How do the leads become aware of you?
• Why do they drop off?
• How and why do they return to your site?
• Why and when do they make the final purchase decision? 

The choice of attribution modeling also depends on how complicated your customer journey is. Here is a quick hack. Large companies are using multi-touch models with hundreds of ads running across different channels. Their customers interact with the brand on different platforms and make decisions at a different pace. When it comes to small businesses with limited platforms, even a single-touch model may help track their behavior.