Marketing attribution models - how to choose the right one

When you’re managing a cross-channel marketing strategy, balancing content, digital advertising, social media, affiliate marketing and more, it’s important that you know which channels are bringing in the most business and giving a good return on investment.

This is where marketing attribution models come in.

What is an attribution model?

Marketing attribution analyses the numerous brand touchpoints a customer comes across, or the actions they take, that contribute to them completing a conversion, and assigning a value to each of these events.

Depending on the attribution model you choose (we’ll get into some of your options shortly), the way you assign value to the different touchpoints in a customer journey will vary. But to put it simply, marketing attribution helps you understand which brand touchpoints are effective in encouraging customers to take action, and which should be given the credit for each conversion.

Why use attribution models?

Effective marketing attribution offers valuable insight into how your customers interact with your brand, and how your marketing efforts are influencing buyers. This enables you to alter your marketing tactics to better cater to the needs and behaviours of your customers, as well as to get full credit for the impact of your work.

But to do it properly, marketing teams must utilise advanced analytics platforms to aggregate customer data from across channels to ensure that each interaction in the sales journey is given the appropriate amount of credit.

Single touch vs multi touch attribution

Smaller companies, with simpler marketing and sales systems (if you only use a couple of marketing channels, for example) may benefit from using a single touch model, such as first click or last click attribution.

Multi touch attribution models, on the other hand, are more appropriate for businesses that use three or more marketing channels, or have a longer or more complex sales cycle. Most marketing teams will benefit from using a multi point attribution model, as this will give you more valuable insights into the effectiveness of your communications strategy.

Single touch attribution models

First Click Attribution

In a first click attribution model, all credit for a sale or conversion is given to, as the name suggests, the very first channel that a customer interacted with – their first click.

What does this look like in practice? Say you see an Instagram ad for a pair of shoes you like, so you click through on the ad to the website. You look around the site for a while, but don’t make a purchase.

A couple of days later, you receive an email from the same brand promoting their shoes. You click through on the email and this time decide to buy a pair.

In a first click model, your purchase would be attributed to the Instagram ad you clicked on, as this was your first interaction with the brand that kicked off your journey to purchase.

First click attribution

If your aim is to identify the channels and tactics that drive the most new customers to your brand, then a first click attribution model can be useful. However, when it comes to analysing effectiveness for conversion, the first click model isn’t typically the best choice, as it doesn’t take into account any touchpoints or actions taken after this initial interaction.

Last Click Attribution

Alternatively, in a last click attribution model, all credit is given to the very last channel a user interacted with before making a purchase.

What does this look like? Let’s go back to the previous example where you bought some shoes.

Again, you see an Instagram ad for the shoes you like, you click through but leave without buying anything. Then when you receive the marketing email, you decide to click through and buy the shoes.

This time, with the last click model, all of the revenue for the sale is attributed to the email, as this was the last thing you interacted with before completing the purchase.

Last click attribution

This model is useful if you want to identify which marketing channels are providing the most conversions, however, as with first click, it fails to take into account any of the other interactions a customer may have before coming to the decision to buy.

Last Non-Direct Click Attribution

This attribution model is similar to the last click model, however, the difference is that any direct traffic is ignored.

What does this look like? Again, we’ll go back to our previous customer journey example. So you’ve clicked through on a shoe brand’s social media ad, and then you received a promotional email and clicked through on that too. But this time, you left the website a second time without making a purchase.

Then, a little while later you decide you do want to buy the shoes, so you type the website URL into your browser and go straight to the site where you make a purchase.

In a last click model, the sale would be attributed to direct traffic. However, in a last non-direct click attribution model, the direct visit is ignored, and the credit is given entirely to the promotional email.

Last non-direct click attribution

Often, direct traffic are users who have already interacted with your marketing efforts numerous times. So this attribution model is useful for analysing the effectiveness of the final stages of your customer journey, without direct traffic diluting the data. But as with last click, this model also fails to take into consideration any other touchpoints prior to the final action.

Multi touch attribution models

Many marketers may not find value in using an attribution model that only gives credit to one channel. A customer will likely interact with a number of different brand touchpoints before converting, so each of these channels should be given credit for the sale.

Linear Attribution

In a linear attribution model, each interaction is weighed evenly, and given equal credit for a conversion.

What does this look like? So imagine a customer first clicked through to your website through a Google search ad, then a social media post, an organic search result and then finally a display ad before completing a purchase.

With a linear attribution model, each of these touchpoints would receive 25% of the credit for the sale.

Linear attribution

A linear attribution model is beneficial if you are looking for a holistic view of your customer journey. However, linear attribution is often not advisable as if every interaction is given equal credit, how can you determine which channels are the most effective?

Position Based Attribution

A position based model (also sometimes referred to as U-shaped attribution), is a multi touch attribution model which prioritises the first and last interactions before a sale, but still gives some credit to any clicks in between.

So 40% of the credit is attributed to both the first and last touchpoints, and the other 20% is split evenly between the rest.

What does this look like? To keep things simple, let’s stay with the previous four point customer journey: the customer clicks through on a Google search ad, then a social media post, then an organic search result, and then finally a display ad.

With a position based attribution model, the Google search ad and display ad are seen as the most significant touchpoints in influencing the sale, so 40% of the credit is attributed to the search ad, and 40% to the display ad.

10% is then attributed to both the organic search result and the social media post.

Position based attribution

A U-shaped attribution model is useful for giving insight into which channels are best for attracting new customers, and which are best for conversions, while still giving an overview of the customer journey as a whole.

However, because all interactions between the first and last are treated as equal, this may not give you a fully accurate picture of how much each channel contributed to the sale.

Time Decay Attribution

The time decay attribution model acknowledges that different interactions have different values, so each interaction is given a different amount of credit for the sale. Brand touchpoints that were interacted with closer to the time of conversion are given a higher percentage of the credit.

What does this look like? Again, let’s go back to our four point customer journey. In this scenario, the click through on the display ad which prompted the purchase session, would be given 40% of the credit for the sale. The organic search result before it would be given 30%, the social media post 20% and the initial Google search ad would be assigned 10%.

Time decay attribution

The time decay model is beneficial in identifying which marketing or sales channels are most effective at converting customers, and which are primarily acquisition channels. However, if your marketing activity centres around top of the funnel campaigns, a time decay model won’t be the right option for you as top of these actions will be the furthest from the point of conversion.

Data Driven Attribution

Data driven attribution is a Google attribution model which assigns credit based on how users search for a product or service using collected data. Google compares the click paths of customers who convert, and those who don’t, to identify patterns and assign attribution based on this.

Data driven attribution uses data from your Google Ads account to determine how people searched for your business, how each campaign, ad group, ad and keyword actually contributed across your customers’ conversion path, which had the most influence on the purchase decision, and which are the most likely to generate more sales for your business.

This is typically Google’s most accurate attribution model, as it’s specific to each ad account and takes multiple factors into account when attributing credit. However, data driven attribution requires a certain amount of data, and is only available to high traffic ad accounts that get a large volume of conversions (15,000 clicks and 600 conversions) each month, so it’s not an option that’s available to all marketers.

Choosing an attribution model

Unfortunately, when it comes to marketing attribution models, there is no correct choice or ‘one size fits all’ – each model has its own advantages and disadvantages that you’ll need to consider.

As a general rule, a multi point model is typically your best choice for insight into the different stages of your customer journey, with a position based model proving valuable to many marketers. However, if your main goal is new customer acquisition and brand awareness, you may find it more useful to focus on a first touch model.

Ultimately, what it comes down to is, does this model give you the information you need to meet your goals?

Take a look at the different marketing tactics and channels you use, and what your overall objectives are, and try to find the attribution model that will most effectively help you achieve your goals. Each model works, it’s all about trying them out and finding which one fits your strategy best.