From the perspective of the modern marketer, never has it been so necessary to understand where every pound is spent - and the kinds of return it brings to the business. Media consumption habits have moved from mono to omni-channel in a short period of time - and keeping up with the always-on consumer is a challenge all marketers face today. The display advertising space has seen a shift in how it is measured and understood - specifically because it’s a channel with so many trackable touchpoints. This has resulted in the emergence of a number of attribution vendors and some notable acquisitions of some of these vendors, by the likes of Google, Rakuten and AOL.
The fact that so many businesses have cropped up (and are viewed as valuable by the tech elite) is testament to how much brands are embarking on a period of discovery to test and learn the best solutions out there in order to better get to grips with the data being fed back from their marketing activities. The double-edged sword of this approach to measurement is that data analysis is almost infinite in its possibilities. By using an attribution solution there is a common trap that advertisers fall into, which is in trying to measure everything. Yet finding a root cause and its effect isn’t always easy to identify - whereas correlating data is incredibly easy. The reason brands still use last touch attribution is not because of a lack of understanding about why it is flawed - that’s particularly well-documented - but because a true multi-touch model is so hard to achieve.
Building an attribution solution that measures every single influence on every single touchpoint, for each unique consumer is an incredibly difficult big data problem, and (at the moment) an almost impossible task.
Let’s take a step back; there is one real reason we use attribution - to work out the impact that each tactic is having so that budget can be allocated correctly. At the moment performance is based on whoever provides the last touch before a conversion, this normally results in too much retargeting and not enough prospecting, meaning that vendors end up crowding ads onto the consumer when they are nearer a conversion because there is no value given to influences further up the funnel.
Insight into your vendors is key
There are two overarching ways in which display media is purchased today: Prospecting and retargeting. In theory prospecting is used to attract new customers that have not interacted with the brand. Once they have, retargeting is used to drive a conversion. Both are very different tactics but are commonly measured using the same metics - and here lies the challenge.
The fact is that many prospectors will also undertake some retargeting as well because, due to the state of performance metrics these days, they need to run additional retargeting activity to ensure they perform as per the plan. It is therefore important that advertisers take the time to gain insight into what each provider is actually doing, otherwise they may well end up over-valuing the wrong providers, accidentally pushing too much budget towards retargeting without even realising and also bidding against yourself on the ad exchanges because a provider you thought was solely prospecting is bidding against your retargeter.
The insight you gain from having those conversations is not just for the purpose of realising what you’re doing wrong, but also so that you can end up running a more achievable attribution model by measuring not just one but two touchpoints; a remedy to last-touch attribution that is easier to create than a multi-touch model. We call this split-funnel attribution.
The key to running split-funnel attribution is introducing the first site visit as a key metric. By introducing just one additional metric, you can immediately improve the accuracy of your attribution and consequently your campaign performance.
Split funnel attribution
Split funnel attribution means that advertisers will be able to take charge and stop encouraging the wrong behaviour from their providers. For example, if I base the performance of my prospecting provider on the number of first site visits they achieve and then, measure the number of these visitors that convert, then I can see not only how effective my prospecting is at driving traffic but also how relevant that traffic is.
This also then helps to reward the retargeting provider based on how many conversions they provide. If a conversion comes from a site visitor provided by my prospecting vendor then that will split the attributed value of the sale. It is an efficient and logical solution to the flaws presented by last-touch attribution, since each provider is judged on how well they do the job they are meant to be doing and consequently focus on their own role instead of all aiming their guns at the last click.
Where are we heading?
It is also true that these same efficiencies outlined previously can be achieved by having one vendor that can provide both prospecting and retargeting. Full-funnel solutions represent a new ‘one-stop shop’ approach for the industry that provide efficiencies from using the proprietary data and technology allowing for a more efficient campaign.
These synergies are hard to ignore. As capabilities start to merge across the industries and advertisers become as device agnostic as their consumers, it is hard not to imagine an industry of one-stop shop advertising providers. These shops will not only solve the issue of budget allocation but also, in a particularly circular way, the issue of attribution itself.