If you work in real-time advertising, you shouldn’t just care about attribution models, you should be downright obsessed with them. Why? Because attribution means figuring out which aspects of
your campaign are working, and how you can do better. And if you don’t care about that, well, then real-time advertising might not be your true calling.
Let’s remember that the great
promise of real-time advertising has always been that it’s much more efficient and trackable than offline advertising. No, real-time marketer hasn’t turned advertising into a perfect
science as some of its early evangelists had hoped, but it has turned advertising into something much more rigorous and efficient than ever before. And attribution models are at the heart of this new
rigor and efficiency. A good model can make it much easier to understand how each dollar you’re spending is contributing (or not contributing) to the success of your campaign.
In other
words, saying you don’t care about marketing attribution is a bit like saying you don’t care about money.
Take the simple example of someone arriving at a software company’s
website after clicking on Facebook Exchange (FBX) ad. Let’s say this customer fills out a form on the company’s website. And let’s also say we know that the customer eventually
became a client after attending a webinar.
If you’re not using an attribution model -- or are using an overly simplified and misleading model -- you might conclude that the click on the
FBX ad and the viewing of the webinar drove the conversion. And you might then start putting all of your dollars toward search ads and webinars.
But what if the customer had seen lots of the
software company’s display banners before ever clicking on that FBX ad or viewing that webinar? What if the customer had also read a native article about the software, or seen a pre-roll on
YouTube? These might be critical touchpoints that moved the customer along. After all, we have mountains of data indicating that many types of conversions take place only after multiple interactions
withtouch points. In the absence of an intelligent attribution model, it will be impossible to know how those earlier touchpoints contributed to the conversion. It’s a bit like a winning
basketball team deciding that the best player on the team is the one who makes the last shot.
It might seem hard to believe that marketers today could still be so cavalier on the subject of
what’s driving conversions. And most marketers certainly know that the last-click approach described above makes little sense. And yet, amazingly, many real-time marketers (perhaps as many as
half) are still using last-click models.
That's not all. If attribution modeling has always been crucial for real-time marketers, it’s perhaps even more crucial than ever in
today’s mobile and cross-device world. While marketers used to think about different touchpoints on a single device, we now need models that allow us to understand the complicated relationship
between the given type of media and the given screen. And we need models that can account for the bridge that mobile provides between online and offline channels. If you’re trying to make sense
of all of this data without an attribution model, you’re essentially flying blind.
This doesn’t mean there aren’t fair criticisms of attribution modeling. Some models are
better than others. And some marketers have oversold what marketing attribution really can do. It can’t, as some people mistakenly think, give you a definitive, absolute truth. When you’re
building a model, you’re always making decisions about how much value to assign to different components of a campaign. For this reason an attribution model is more of a macroanalysis than a
microanalysis tool.
But such macroanalysis makes all the difference. According to a report in Forbes, 90% of marketing organizations that have begun to use attribution models
“are seeing significant benefits.” The benefits cited include a “better understanding of the customer journey” and the ability to optimize the media mix in a given
campaign.
It comes down to this: Attribution models provide valuable insights. Those insights can inform your next spend, which can save you money and increase conversions at the very same
time. How can anyone not care about that?