The digital advertising industry has a problem. And no, I’m not talking about privacy concerns or lack of transparency. Those topics are certainly on the list, but before those issues made mainstream discussion – years before programmatic or GDPR was even a thing – marketers were becoming obsessed with the sheer trackability of digital. If the question was which half of the money spent on advertising was working, the answer was clearly online, and shifts in advertising budgets have followed.
However, at the heart of the digital revolution is a major flaw in our foundation. The data – or at least the vast majority of it – is collected and presented based on the notion that “He who is last to be seen wins.” Seems to me, that’s the definition of a race to the bottom.
Think about it. Google Analytics is built to show last click conversions. Sure, they have conversion path reporting and assisted conversion values, but based on what? Only what Google Analytics can track on your site.
Ad servers do the same for media, but with a lack of visibility to site analytics data. They tout cross-device tracking and deduplication, but against what? Certainly not organic traffic or social engagement. Inherently, almost every tool or platform out there is built to take credit for the conversion, and everyone is constantly trying to figure out why one data set never matches another.
Let’s look at an example. Suppose Jane Doe is shopping for widgets this holiday season. She sees a few ads for a sale at widgets-r-us.com – one on her phone, one on her laptop, another on her phone and she clicks to find out more. A day or two later, she forgets the URL, searches on her laptop for the name of the widget she was looking for and winds up back at the site. She puts the widget in her shopping cart but decides to do a little comparison shopping. Later on, she revisits the site and checks out once she realizes it’s the best deal. Sale complete.
In this instance, there are three completely different ways to analyze the same conversion path. Let’s start with site analytics:
In the example above, none of the advertising is credited for influencing the sale. Even though Jane did click on an ad, the fact that the device where the conversion took place was different means that the mobile click event was statistically insignificant. It was nothing more than a mobile site session. It should be noted that at least the mobile click issue is being addressed with Google’s Cross Device beta, but that still leaves the impression impact lacking from the analysis.
“Well that’s not fair,” says the agency who bought the ad or the media company who served it. That’s because the agency’s ad servers would track the conversion path like this:
In this path analysis, you can actually assign conversion to two sources – either the view-thru impression in the second position or the mobile click in the fourth position. Add to this the fact that ad servers can’t track site analytics, and no credit is given to either the organic or direct engagement. Confused yet?
So who “wins” the conversion. If you’re responsible for SEO, you’d certainly argue it’s the organic search visit. The brand strategist would say it’s the direct visit (great brand recognition). The media buyer would argue it’s the media campaign. The smart marketer realizes it’s all of them. Enter scenario three and Attribution Modeling.
You might ask why every advertiser isn’t doing this, and the answer is as clear as mud. Centralizing data is no small task. It takes manpower, computing power, and frankly, money. It’s not cheap, and it’s not easy. The model is only as good as the data you’re able to feed it. For most marketers, the cost, the complexity of the data architecture and the infrastructure needed is a risk-barrier seemingly not worth crossing. But let me have you consider this: If we continue down the path of using data that only identifies half-truths, are we really any better off than we were before the digital media revolution?