This article was published on July 27, 2014

Why last click attributions breed bad decisions in a multi-device world


Why last click attributions breed bad decisions in a multi-device world

Sam Barnett is the CEO and co-founder of Struq.


The last click attribution model has always been flawed. It attempts to simplify complex conversion paths on a last come first served basis, and allows advertisers to ignore the vast amount of data available to them.

With the rise of the multi device user, it is now actively driving bad decisions. We need to rethink the way we measure effectiveness without resulting to traditional methods.

Measure your performance not your clicks

How we measure performance is the first crucial part of the issue; Click Through Rate (CTR) is one of the most widely used metrics for measuring ad effectiveness. This is despite the fact that a click doesn’t actually generate any revenue.

A recent study by eBay research labs into the effectiveness of Search Engine Marketing (SEM) stated as much – “clicks alone are not a source of revenue.” The main reason the industry is chasing clicks, not performance, is due to the traditional ways we measure and pay for performance in the advertising industry.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

Traditional attribution models are limiting your view of performance

Across the space the performance of each vendor is decided by attribution models, these credit a conversion or sale to whichever vendor is seen to drive that sale. Unfortunately the last click attribution model is the incumbent model in use in the industry.

To break this habit advertisers must focus on the way they view their users. Traditionally, when we look at our advertising space we see many devices, each of which is eligible to be shown an ad. We should, instead, see many users who own many devices – 90 percent of users will now use two or more screens a day to browse the net.

smartphone subway

Let’s say I start off browsing on my phone, seeing and clicking ads, visiting your site. When I make up my mind, I move to my PC because my payment information is saved there and search for the product. If my final click is an SEM result, you pay extra for that and the SEM provider gets attributed the sale.

Later, you review your vendor’s performance and make the decision to drop activity on your mobile channel because your attribution model says it does not provide a good return on investment, a lack of insight has actively driven a bad decision.

To gain insight, advertisers must move to a transparent relationship with their vendors both in their inventory and performance. If the business relationship is based on performance targets instead of clicks, then the vendor and advertisers goals are aligned.

This allows for strategic conversations so that there is a deeper understanding of the vendor offering and how it is contributing to revenue – these are the building blocks of a multi touch attribution model that measures the contribution of each touch point in the user’s path to conversion.

Incremental the key performance metric and can be tested

With the above in mind focus naturally shifts from clicks to “incrementality.” An incremental sale is one that would not have occurred without advertising activity taking place.

The same study concluded that eBay’s SEM activity that “paid-search expenditures are concentrated on consumers who would shop on eBay regardless of whether they were shown an ad”.

You do not need a team of academics to write a 35 page paper in order to measure your incrementality. But you do need a robust methodology. Your vendor and or third parties can help you avoid the pitfalls of incrementality. You can read more about that here.

Multiple devices and lost clicks drive negative behavior

Advertisers need to have a view of their users with many devices instead of just a pool of uniform devices. There is no excuse, with the rise of cookieless user tracking across devices both through deterministic and probabilistic methods, that advertisers should not be able to serve their user the right ad, at the right time on the right device and be able to measure the effectiveness of it.

Your providers probably already know all of this, but if you keep paying for clicks and not performance then your goals aren’t aligned and chances are they will choose to stay in their black box.

Get the TNW newsletter

Get the most important tech news in your inbox each week.