What are attribution models?

  • Written by

It’s important as an advertiser to understand which online channels are driving sales and conversions. Attribution models can help analyze the data.

Share this

What are attribution models?  

It’s important for advertisers to understand where their sales and conversions come from. An attribution model is a set of rules that helps gather and analyze this information. Attribution modeling can come in a few shapes and sizes, so let’s explore some of the different ways to make use of marketing attribution models. 

Why is marketing attribution important? 

There are many pathways a customer may take before ultimately making a purchase or conversion. They might simply click an ad, visit your website and make a transaction – but in most cases, there’s a lot more to the journey than that.  

For example, a customer might first click on a banner ad and visit your website but not make a purchase. Then, days later they might be served a retargeting ad, which they click, but again does not lead to a sale. While on your site they may have visited your blog and learned a little more about the brand and product, and they might then follow your business on social media. The following week they see a post with a link that ultimately leads to them finally making the conversion.  

Which marketing channel should be credited for the sale? Was it the original ad that introduced the brand? Was it the retargeting ad? How important was the blog? Did it all come down to the social media post?  

As a business, understanding which marketing channels have the greatest impact is essential to your continued success. It helps you align your strategy with the data, so you can place more emphasis and budget on what you know works.   

Marketing attribution models allow businesses to analyze each of these channels, or touchpoints, and determine which of them had the most significant impact on gaining a conversion.  

What are the different marketing attribution models? 

There are a number of different attribution models that each distribute the value across touchpoints differently. Touchpoints describe any point where a potential customer interacts with your business on the sales journey: it could be display advertising, SEO, social media, email marketing, paid search, and so on. 

The most common marketing attribution models are:  

Last-click attribution models

Also known as last interaction or last-touch attribution, last-click attribution places all of the credit on the last interaction the customer had with the business before converting. In our earlier example, the social media link would receive 100% of the credit, as this was the last touchpoint before the sale was made.  

This is the simplest attribution model and is often the default model for analytics platforms. Because there are so many journeys that can lead to a conversion, many feel the most accurate way to look at it is to simply give all the credit to whatever the final interaction was.   

However, this disregards all of the other steps that led to that final interaction. In our example, the customer would not have seen and interacted with the social media post had they not initially been served the ads and visited the site, so you may feel it’s illogical to give all of the credit to the last touchpoint.  

First interaction attribution models  

On the other hand, some might feel the very first interaction was the most important. First interaction attribution gives 100% of the credit to the first click. In our example, the first banner ad the customer clicked on gets all the credit – it initiated the journey, and without it, none of the other touchpoints would have been hit. 

With this model, whatever happens between the first interaction and the conversion is insignificant; the first interaction sets the ball in motion and therefore deserves all the credit.  

Last non-direct click attribution models  

Last non-direct click is another attribution model that places 100% of the credit on a single touchpoint, but in this case, any direct interactions are disregarded. Direct traffic refers to a customer manually entering your URL to visit your website, or clicking a bookmarked link. The assumption is that if they’ve bookmarked your website or are typing in your URL, they already know your company, and they likely already expect to make a purchase.  

Linear attribution models  

A linear attribution model places equal value across every touchpoint in the customer’s journey. While it does provide a more balanced look, this assumes that every touchpoint was just as effective as the other – which isn’t necessarily the case.  

A linear attribution model is a very straightforward model and won’t really help you identify which marketing channel had the most significant impact.  

Time-decay attributions  

In a time-decay model, the value is spread across each touchpoint, but not evenly. Instead, the touchpoints closer to conversion are assigned a higher value. This is a great way to show the importance of customer relationship building. The first interaction has the lowest value, but with each touchpoint, the customer relationship grows stronger, before finally leading to conversion after the last, most valuable interaction.  

Position-based attribution models  

Position-based attribution gives different weighting to different stages of the journey.  One of the most common types of position-based attribution is called U-shaped attribution. In this model, the first and last touchpoints equally share the most value – 40% of the credit each, while the interactions in between share the remaining 20%. So, in our example, the first banner ad takes 40% of the credit, as does the social media channel, while the retargeting ad and blog each take 10% of the credit.  

The first interaction is valuable as it introduced the customer to the brand, the last interaction is equally as valuable as it closed the deal, while the other touchpoints still carry some value, but not as much.  

Attribution and De-Duplication in the Affiliate Channel

While attribution remains essential in the affiliate channel, affiliate programs also rely on de-duplication to allocate commission across the partner base.  Most commonly, affiliate programs run on a "last click wins" basis, meaning the last affiliate in the customer journey is attributed the commission for the sale, but can also be used to accommodate other digital channels, if an advertiser should choose to not pay for a sale if the affiliate channel was not the last referring channel.  The IAB Affiliate Marketing handbook indicates that all advertisers should provide complete transparency on their de-duplication practices so publishers can understand how their traffic is reviewed against other online channels.  Awin’s recommendations for de-duplication can be found on the Wiki.

Which multi-touch attribution model works best? 

Each model is useful in its own way, and businesses would benefit from using each one to analyze their advertising strategy.  

First click is useful when analyzing which channels are best for bringing in new customers and raising awareness. At the other end, last-click attribution is best when analyzing conversions specifically, and which channels best bring potential customers to the bottom of the sales funnel.   

Last non-direct click attribution is similar to last click and might be more useful, as you can argue that the last non-direct click is what really sealed the deal.  

Position-based models still offer a broad view of each touchpoint’s value while placing more importance on the start and finish of the journey. This is useful for those who want to look closer at which channels are most successful for acquiring audiences and which are most successful for converting audiences, without completely discounting the steps in between.  

Think about the goals of your marketing campaign. If you’re fully focused on sales, perhaps a last click or last non-direct click model provides the best insight. If you’re more focused on increasing brand awareness and gaining new leads, a first click model will help most. Perhaps you’re interested in seeing how well you retain interest and awareness across the entire sales cycle, in which case a linear model is ideal. If you’re running a short campaign, like a special promotion, then a time decay model will be most useful.  

How to use multi-touch attribution  

So, now we know the different ways we can analyze the sales journey, but how do we use this information? The attribution model will show you how much revenue should be attributed to each touchpoint.  

Let’s imagine you were focused on a single sale of $100 from a customer that interacted with four touchpoints. If you used a single-touch attribution model, the entire $100 would be credited to either the first or last interaction. If you used a U-shaped position-based model, then $40 would be attributed to the first click, $40 to the last click, and the two touchpoints in between would each be attributed $10.  

You can then use this information to guide your marketing and advertising spend going forward. Perhaps you find through attribution models that hundreds of people engage with your blog without making a conversion, but that social media ads do lead to conversions. You might then choose to increase spend on social media, or you might use this info to reshape and improve your blog content. 

Awin’s advertiser tools are making multi-touch attribution more accessible and accurate than ever. Using the SingleView solution, advertisers on the Awin network have a user-friendly, impartial, and comprehensive platform to gain important marketing insights, helping you measure the precise value of different affiliate partners and marketing channels.  

SingleView, available via Awin’s Advertiser MasterTag, allows you to track performance across some of the most popular ecommerce platforms. You can then use this data to compare different attribution models, with in-depth analysis tools, data-driven recommendations, and features that take the guesswork out of attribution.   

Attribution modeling will help you form stronger partnerships with affiliate partners and better align your budget and strategy based on the channels that drive maximum value. With a robust approach to attribution modeling, you’ll gain a true understanding of your marketing strategy and how best to serve your target audience and conversion goals. If you’d like to know more about SingleView, you can get in touch with the team. 

Related articles