Today’s marketing landscape is run by data. But to make any sense of that data, you need to make sure you understand where it’s coming from. Enter marketing attribution. Marketing attribution helps marketers understand where their conversions are coming from and how to use that info to make better decisions.
Marketing attribution is a way of determining what marketing communications contributed to a conversion. In other words, it’s the process of finding out what marketing campaigns pushed someone to make a purchase.
For example, let’s say that the same customer saw both Facebook ads and LinkedIn ads. Marketing attribution would tell the marketer which ad led to the purchase. From there, they can restructure their campaign to account for this.
Ultimately, marketing attribution helps marketers run more efficient campaigns. Plus, it makes sure their marketing budgets are going straight to the most successful touchpoints.
A few years ago, most people only had one device that they would use to browse the internet: their computer. Eventually, that grew to a desktop and a laptop, then smart phones were added, then tablets, and so on.
Now, the same consumer might browse the web from a desktop, laptop, work laptop, smartphone, tablet, and VR headset all in the same day. Each of those might look like a unique visitor on your analytics tool. As you might imagine, it’s a tough challenge to parse out which marketing touchpoint led to a conversion.
The first step is to set up an analytics tool like Google Analytics or WordPress. Once you’ve got that all ready, you can move on to choosing the right marketing attribution model for your needs. These include:
As the name implies, first-touch attribution means that the first ad that a prospect interacts with or sees is the one that gets the full credit for the sale.
The theory behind it is this: no matter how many ads the consumer saw, they ultimately made the unconscious decision to convert after seeing the first ad. So, that ad gets 100% of the credit for the conversion.
Of course, things are never this clear cut, but first-touch attribution makes this assumption for practical reasons. This type of attribution is easy to set up on Google Analytics, but it also leaves a lot of data out of the picture.
Last-touch attribution is basically the opposite of first-touch attribution. Instead of giving the credit to the first ad that a consumer saw, this form of attribution gives the credit to the last ad that the consumer saw before converting.
Essentially, last-touch attribution assumes that the last ad you saw was the most convincing, and that’s the one that pushed you over the edge and got you to convert. Consequently, it gives it 100% credit and pays no attention to any touchpoints before it.
Unfortunately, just like first-touch attribution, this model ignores a big part of the picture by only taking into account the last touchpoint.
Last non-direct click attribution is similar to last-touch attribution. However, it gives 100% credit to the last ad that the consumer clicked on outside of your website before making a purchase.
For example, if a consumer saw an ad, clicked on it, didn’t make a purchase, saw another ad, and then made a purchase, the ad that they clicked on would get the credit, even though the second ad came later.
While the last few attribution models we looked at only look at one touchpoint, multi-touch attribution models take all touchpoints into account. As a result, they are generally considered to be more accurate.
To make sense of all this data, multi-touch attribution models typically weight the touchpoints differently. So, you have some models that weight later touchpoints more than earlier ones and vice versa.
Linear attribution assigns credit to each of the touchpoints equally — no preference is given to clicks, proximity to conversion, or anything else.
For example, if you have 20 touchpoints, each one will get 5% of the credit. It’s a bit basic, but it can be useful for certain circumstances and advertising models.
One of the good things about linear attribution is that it allows marketers to take the whole picture into account. However, it also doesn’t provide any differentiation between touchpoints.
This model is most similar to last-touch and last non-direct click attribution. Unlike linear attribution, which divvies up the credit equally, time decay attribution gives more credit to touchpoints that are closer to the conversion event.
Essentially, the closer the touchpoint is to the conversion, the higher its weight. Ultimately, the last touchpoint will get the most credit, and the first will get the least.
This model succeeds in helping marketers identify the touchpoints that led to conversion events more easily. However, it doesn’t provide any information about how the customer found the business in the first place, which is important to know.
This attribution scheme tries to find a compromise between first-touch and last-touch attribution. In short, it gives 40% to the first touchpoint, 40% to the last touchpoint, and then divides the remaining 20% between however many touch points came in between. So, if there were 20 touchpoints, each would get 1%.
In this model, marketers attempt to give the most weight to the first and last-touch points — theoretically, the first time the customer was introduced to the business and then the last ad before making a purchase.
Even once you’ve hammered out which attribution method you’d like to use, there are still mistakes that you might make and challenges you’ll come up against. Making these mistakes can compromise the integrity of your data and insights.
One of the pitfalls of marketing attribution is that it doesn’t properly take into account the value of a brand and how people react to it. When marketing is reduced solely to weighted touchpoints, it can be difficult to get the full picture.
If you’ve ever taken a science or math class, you’ve likely heard the phrase “causation is not correlation.” Despite being such a famous phrase, few people remember it when it matters.
In short, it means that you can’t infer that one thing caused another just because there is something seemingly linking the two together. When you are looking for patterns, your mind can sometimes start making them up, seeing them where there aren’t any, or favoring a specific result, which is where correlation-biased bias comes from.
Essentially, as you look through your data, it’s important not to get too carried away with what you’re seeing. Don’t assume that just because several people converted after seeing a specific ad that there must be a causal relationship. It’s possible there was merely a correlation and that the true cause was a different ad entirely.
In-market bias is problematic because it includes all the people who saw your ads but were going to convert anyway. For example, this could include someone who heard about your product from a friend, decided to buy it, and then just happened to scroll past your ad on Facebook before they had the chance to make a purchase.
Unfortunately, there’s no easy way to account for this, so you’ll need to keep it in mind as you analyze the data. Just remember that not 100% of the numbers that show up are going to be coming from your ads, and you should be fine.
Marketing attribution is important because it helps marketers keep track of which communications and campaigns are working. By understanding what leads customers to make purchases, you can extract valuable insights and dial in ever more efficient campaigns.
Understanding market attribution is both an art and a science. You’ll need to experiment with different types to find the best solution for your business.