Moving off Last-Click Attribution for Startups
Most companies fail to maximize the returns from paid media because they fail to attribute appropriate revenue to the spend.
Revenue maximizing attribution can be achieved relatively easily by the marketer and without much input from the analytics parts of the organization.
Companies Struggle to Move Beyond Last-Click Attribution
As much discussion as there is on attribution models, today still 91% of agency clients use a single-click attribution system. Further, according to AdRoll, 59% of marketers say 'lack of knowledge' is the main obstacle to implementing a smarter attribution system.
Moving Beyond Last-Click with a Thought Experiment
When you are on a single-click attribution system, moving to literally anything else will be more effective. Generally speaking, revenue and efficiency will improve ~15% in moving off a last-click system. So how do we begin? Let's start with an experiment.
For simplicity, imagine you are a startup with neither offline spend nor an app. Less mature organizations, tend to lack the funds needed to execute on large-scale TV, radio, out of home, and other forms of offline branding. Further, they likely have just a website and while there may be plans for an app one day, today they likely are web-only.
Now, imagine a channel revenue breakdown like the following:
Next, some reflection.
How did you generate your email list?
Most email captures are generated from other channels (unless you are buying lists). Given this, shouldn't we immediately "re-attribute" the 10% of email revenue to the channel from which the email was captured?
If you are not already capturing this in your database, ask your engineers to start today.
Even without this precise data, one can simply re-attribute the 10% of email revenue proportionally to all channels. Otherwise stated if organic accounts for 30% of your revenue, simply attribute 30% of email's revenue to the channel. Yes, it is imprecise. But it is definitely better than what we were doing before.
Now, the attribution model looks like this:
Next, consider the source of direct revenue
When someone types in our URL direct into their browser and arrives on our site, it is because they are familiar with our brand (how else would they know our URL?). Most often users coming direct have previously engaged with our brand.
This trend becomes clear with a quick analysis of paid clicks versus direct and brand clicks. Quite often, direct and brand traffic moves proportionally to the amount of paid or non-brand traffic.
This intuitively makes sense. A user may start with a generic search like "best gaming laptop" and land on bestbuy.com. After browsing the available inventory the user returns a week later directly to bestbuy.com and purchases the laptop. Clearly, the purchase was influenced in its entirety by the previous non-brand search.
As such, for the organization that engages in no offline marketing we can re-attribute nearly all direct revenue with a previous touch-point.
Now, we get a view of revenue by channel having re-attributed both direct, and email revenue:
Re-Attributing Email and Direct Revenue Provides Paid Media Boost
When we've successfully gone through the thought experiment of "where did our email and direct revenue originate from?" we are able to spend more to drive the top of the funnel.
Originally, paid media looked like it drove 25% of the revenue in the example. After re-attributing email and direct it accounts for nearly 35% of revenue!
Tactically, this means that we can now bid up ~50% knowing that what we don't make off the click we will make up for in follow-on email and direct traffic.
