What is attribution?

In marketing terms, attribution is the way to determine which brand touchpoints a customer interacted with before making a purchase. These touchpoints could be anything to do with the way a brand markets itself. It could be an email newsletter, a TV commercial, or an Instagram advert.

Attribution is a science. It’s intended to give data-driven insights into which marketing efforts are the most profitable and effective. 

Marketers need to master attribution to know if they’re getting a good return on investment (ROI) for their marketing campaigns.

What are marketing attribution models?

A marketing attribution model is a framework for attributing customer conversions to different channels, depending on when a customer interacted with them.

It used to be quite simple for marketers to find out how customers accessed a brand, but marketing methods have become a lot more elaborate in recent years.

There’s social media, for starters, as well as display advertising which comes in loads of different formats.

With all those marketing opportunities, models were designed to help add more meaning to the role of each channel. 

Each attribution model could fall into one of two categories: single-touch or multi-touch. 

Single-touch attribution

This format is becoming less and less popular. Attribution models that are single-touch look to attribute a customer conversion to only one touchpoint.

That touchpoint is usually the first or last place a customer lands before the conversion step.

With all the marketing channels available to shoppers, the single-touch format misses out all the possible places the customer may have been on their journey to a checkout.

Multi-touch attribution

So, how does a marketer get a more nuanced look at the customer journey? 

That’s where multi-touch attribution models come in. They consider every touchpoint that the customer engages with before they make a purchase.

There are several different multi-touch attribution models, and marketers can attribute value to channels in different ways depending on which model they use. 

Types of marketing attribution models

Marketers can use several different types of attribution models if they want to determine how much credit each channel gets for customer conversion.

Last-click attribution

Also known as ‘last interaction’ or ‘last-touch,’ this single-touch model attributes all credit to the final touchpoint in the customer journey before conversion.

• How it works. Say a customer sees your ad on public transport and searches for your business online. One week later, they see an ad on an Instagram story, swipe up, and make a purchase. All value is assigned just to that last channel – Instagram.

• Why it’s great. It’s a simple way to attribute value, and it’s always easy to find a customer’s last touchpoint. Last-click would work well for a business that’s focused on conversions and knowing what the deciding factor was.

• Why it isn’t great. It ignores all the touchpoints before that one final channel. These may be just as important in appealing to customers, or may be worth a share of the value, so last-click is too narrow for a lot of businesses.

First-click attribution

You’ve got it: this single-touch model is also known as ‘first interaction’ or ‘first-touch.’ Much like last-click, first-click attribution assigns all value to a single channel.

• How it works. If you run an ad on Facebook and it’s the first place someone becomes aware of your business, that ad gets all the credit. It disregards any follow-up strategies that the customer may come across in their journey.

• Why it’s great. Just as with the last-click attribution model, first-click analysis is easy to run. It would work for marketers who aren’t fixated on the success of conversions as an influence on their campaigns.

• Why it isn’t great. Again, this model completely overlooks any subsequent strategy the customer comes across before purchasing. A first-touch model makes it hard for marketers to determine what actually drives conversions.

Linear attribution 

This model has a little more nuance – it’s the simplest way to use multi-touch attribution. A linear model splits credit evenly between every touchpoint prior to conversion.

• How it works. A customer stumbles across your business on Instagram, then signs up for your mailing list. They then see a display ad while reading a blog post and click through to purchase. Every touchpoint is then worth the same amount.

• Why it’s great. Linear attribution is almost as straightforward as first-click and last-click models, but it’s working harder by recognizing every channel in the customer journey ahead of conversion.

• Why it isn’t great. It sounds fair, but using linear attribution could give too much or not enough credit to a particular touchpoint. The first or last channel a customer engaged with may actually have been more high-impact than any other.

Time decay attribution

Another multi-touch model, time decay looks at every touchpoint used but gives each one an increasing amount of value the closer they are to conversion.

• How it works. Someone sees your promoted pin on Pinterest. Then they see a display ad. They sign up to your mailing list and click on an email link to purchase. Every touchpoint gets credit, but the email link will get the most.

• Why it’s great. Time decay recognizes every interaction but gives emphasis to the channel that gave the customer that last push to purchase. This method could be useful if a marketing campaign lasts several months.

• Why it isn’t great. Marketing strategies at the top of the funnel – meaning where the customer first learns about a business – will never get much credit, even if they were just as impactful as the last touchpoint.

U-shaped attribution

Also known as ‘position-based’ attribution, this model is multi-touch but it assigns the most credit to two touchpoints: the first and the lead creation.

Wait, what’s ‘lead creation’?

A lead is a person who has shown interest in purchasing from your business. The term ‘lead creation’ refers to when someone takes the first step in preparing to purchase by handing over their contact details.

• How it works. If a customer first sees you on social media, engages with a few more of your strategies, then makes an account with you, those first and last touchpoints get 80% of the credit. 

• Why it’s great. Marketers can identify the strongest touchpoints for introducing brands to people, and for converting them into leads. It also acknowledges that other marketing channels played a role.

• Why it’s isn’t great. It doesn’t always make sense. The first touchpoint may be really insignificant to the conversion path, and there may be additional points after the lead creation that are relevant. 

Full-path attribution

Also called ‘z-shaped’ modeling, this is the most extensive non-custom model in marketing attribution. It’s a format that will give you an overview of the entire customer journey.

• How it works. Full-path attribution gives equal credit to key touchpoints. It acknowledges first-click, lead creation, opportunity, and the purchase stage – they each get 22.5% credit. All in-between touchpoints share the remaining 10% credit.

• Why it’s great. This is a complete measure of the sales funnel, right from the moment a person becomes aware of a business, to when they finalize a transaction. It can really help businesses understand how marketing efforts impact sales.

• Why it isn’t great. Because it is so complex, a full-path model can be costly to implement. It also still isn’t the most nuanced way to analyze attribution, as again you’ll be giving set amounts of value to different touchpoints.

Custom attribution

This could be the best option if your needs are specific or complex. With custom attribution models, you can build a format that is tailored to your business.

• How it works. Businesses or marketers can build a model that allows them to adjust the weight of value given to different channels. Google Analytics allows for custom modeling. It’s a useful way to optimize attribution to suit specific business needs.

• Why it’s great. Custom attribution can be the most accurate way to measure channel values. For a multi-channel business with a sophisticated data tracking setup, this model is appealing for its efficiency.

• Why it isn’t great. Even if you use Google Analytics, it’s a complex, time-consuming model to implement. For businesses just starting out with marketing attribution, a custom model is probably too much work.

The benefits of attribution modeling

Attribution modeling can be a challenging way to work. Marketers are tasked with connecting all the dots of each customer’s journey, which is no easy task. 

But attribution also has some really clear benefits that can make all that effort worthwhile:

• It helps you find the most effective channels. Multi-touch models can give businesses clear data-driven insights into what is and isn’t working for them.
• It increases the ROI of marketing. When done properly, marketing attribution will show marketers how to refine their strategies and increase conversions.
• It reveals the customer journey. Anyone who’s ever wondered how other people shop will gain fascinating insights with attribution modeling.
• It improves product development. Attribution is customer-focused, so the data and insights gained can play a big role in influencing product innovation.
• It improves marketing creativity. When marketers are able to understand what is and isn’t working, they can perfect the visual elements of their strategies.

How to choose the perfect attribution model for your business

There isn’t really a perfect attribution model. 

Attribution is a complex task and every model has its flaws. But you can find the best format for your business by considering a load of relevant factors:

• How many channels you can market strategies through
• The typical length of selling cycles for products in your business
• What percentage of customers buy online compared to offline

While a business with a short sales cycle and fewer channels may prefer a single-touch model, the multi-touch approach will give you way more data insights.

We already said that we should define attribution as a science, but there’s an art to it too. A lot of your success with attribution modeling will come from having a great intuition.