In an ideal world, a customer will spot your ad, click on it, and convert immediately to a sale. But it’s rarely that simple.
In fact, there are likely to be many stages before a customer spends on your website, or signs up for a subscription. This customer journey can be complicated for you to follow and frustrating to understand.
At Epsilon, however, we have spent decades developing the tools and skills needed to track customer behaviour so we can help you identify which touchpoints trigger a sale.
In this blog we will take you through the six most popular attribution modelling techniques and explain why a hybrid approach gets the best results.
What is Attribution Modelling?
Attribution modelling helps marketers understand an individual’s unique journey to a purchase.
It is a way of figuring out which touchpoints should get the credit for a sale. Each attribution model distributes the value of a conversion across each touchpoint differently.
For example, the 'last interaction' model assigns 100% credit to the final touchpoints immediately before a sale.
You can use a model comparison solution, which analyses how each model distributes the value of a conversion. It takes into account the customer touchpoints between devices, channels, walled-garden platforms.
This will give you a clearer picture of how different marketing channels contribute to a sale. And of course, this will help you shape your future digital advertising strategy.
When attribution modelling is well executed, it can save you money. Less advertising budget will be wasted. It can also boost sales conversions over time.
The beauty of modelling is that it allows you to focus on the buyer's journey, understand what's working well and what needs improvement. Attribution modelling solutions also show whether your marketing channels are working well together.
The Pros & Cons of Leading Attribution Models
Expert opinion is divided across the marketing industry over the best form of attribution model. Here we critique the six most popular models and offer our industry-leading advice.
1. Last Interaction Attribution
Last interaction attribution is the simplest to put in place and test. Called 'last-click' or 'last-touch' attribution, it gives 100% of the credit to the last customer interaction before the sale. This is the default attribution model in most platforms. Google Analytics, for example, uses last-touch attribution in its standard conversion reports.
With this attribution model, a visitor may find your website through organic search. A week later they see an Instagram Ad and click on it. The next day, they go to your website directly and make a purchase. The direct traffic gets all the credit for that sale.
It’s very easy to implement, but the downside is that 'last-touch' ignores everything before the final interaction. Many of the interactions before that last-click will be just as important.
But for many online brands, this model can be useful if you have a short buying cycle. If there are few touchpoints before the sale, 'last-touch' will give a good idea of your strongest channels.
2. First Interaction Attribution
As the name suggests, this model gives credit to the first user interaction with a brand.
A person may initially click on a Facebook ad for a sportswear brand’s website. They may not buy from that initial engagement, but they start seeing digital ads for the swimsuit on Instagram. A few weeks later, they buy the product after researching it more.
In first interaction attribution, the credit goes to that initial Facebook ad. None of it would go to the subsequent media buys, even though they had influenced the individual’s decision to buy.
This model is easy to integrate. However, it completely ignores valuable interactions a customer may have had after the first one. This model is also helpful if your particular product category has a short buying cycle. If customers tend to be converted immediately, then their first touchpoint is critical.
3. Last Indirect Click Attribution
Sometimes referred to as ‘last non-direct click’ attribution, this model assigns all credit to the last indirect interaction.
If a customer is looking for a new e-bike, they may engage in a few marketing efforts. For example, they may check the brand's app before using a direct search on Google to buy their bike.
Last indirect click attribution would assign total value to the app impression. That's because the customer already knew about the brand from previous marketing efforts. This would guide them to reach the website via direct search.
Eliminating direct clicks makes this a more insightful model than the last interaction. But it still assigns 100% of the value to one interaction. However, all the interactions before the last indirect direct click are ignored.
4. Linear Attribution
This model splits credit for a sale equally between all the interactions the customer had with your business.
If a customer finds you on TikTok, signs up for your email newsletter and later clicks an email link. The following week they go to your site directly and make a £240 purchase of a garden bench.
There have been three touchpoints in this situation. Each touchpoint gets equal credit of £80 attributed to the channel when the sale was made.
Unlike the other models, linear attribution offers a vast improvement in reporting accuracy. The problem with this approach, however, is that it assigns equal value to every marketing interaction. This is true even though some channels may be better at converting customers than others.
5. Time Decay Attribution
The 'time decay' model shares attribution across lots of touchpoints, but it also factors in when the interactions take place.
Interactions that happen closer to the time of sale have more value attributed to them. The first interaction gets less (or no) credit, while the last interaction will get the most.
One consideration is that this model minimises the effect of top-of-the-funnel marketing techniques.
It’s probably best to use 'time decay' when you're dealing with a particularly long sales cycle.
6. Position-based Attribution
This model gives the most credit to the first and last engagements. The touchpoints in between are given equal credit. It is also known as 'U-shaped attribution'.
For example, Google Analytics gives 40% credit to the first and last interactions and 20% for all the other touchpoints.
To illustrate this, imagine someone purchasing a handbag. Their first interaction with the brand may be an ad, while the last was a direct search to the brand's website.
Position-based attribution assigns most of the credit to the first and last interactions and splits the rest evenly.
This method reveals which channel first engages your customers and the last before the sale. It still does not accurately measure the influence of marketing efforts in between.
The benefits of a hybrid approach
Figuring out which interactions drive sales is difficult, but it’s a critical part of digital marketing. If you want to increase revenue it’s worth investing in marketing attribution software.
Each of the six attribution models has its own set of pros and cons. A combination of these models, however, often provides the most accurate view of your customer’s journey. This can be achieved by working with an expert partner.
By employing an attribution solution, you can work with ad media to make sure you use the right model, or a mix of models, according to a number of factors. One important consideration is where in the sales funnel your typical buyer is likely to be. If they are mid-funnel, and in that case ‘unsure buyers’, it’s worth investing in ad media to engage them. Whereas people who are 90% decided on buying your product are not worth spending on multiple interactions to convert. Segmenting your audience and optimising activity for each, in line with the most suitable attribution model, will pay dividends.
By using the Google Ads analytics tool, you will be restricted to ‘marking the homework’ of the Google channel in isolation. More sophisticated attribution reporting tools focus on enterprise cross-platform, and cross-channel advertising attribution. Such tools offer tailored attribution solutions for businesses across a range of industries.
It can be helpful if your chosen solution considers external factors, such as competition or changing industry trends, when pulling attribution data.
And, of course, it’s vital you can scale and flex around the product type and the value of the purchase in question. Attribution modelling for a low-ticket item like a pair of slippers will be very different from modelling for £5k luxury holidays. Your chosen solution must accommodate this.
Start by establishing clear business objectives
So, there you have it. At Epsilon, we believe a mix-and-match approach to marketing attribution will beat a cookie-cutter solution every time.
We start our approach by asking what your business goals are. You may want to increase your market share by 10%, or need to increase customer lifetime value by 20%. Armed with this information we can then set up an optimised attribution strategy tailored to your needs.
If you want to take digital attribution to the next level and boost your marketing campaigns, Epsilon can help.