The topic is a hot one. For at least a couple of weeks now, both trade and mainstream press have had stories covering the subject. Maybe you've heard about it. "Fake news?"
If you don't already know -- and I can't imagine if you spend time on Facebook, or pretty much anywhere across the internet, you don't -- fake news is the catch-all for websites that publish a host of hoaxes, propaganda, and outright false information for the purposes of driving traffic that can in turn be monetized in automated digital media buying marketplaces.
Fake news as a category has been around for some time. In the dark, or sometimes not-so-dark corners of the internet, sites with content that was at best dubious have long wormed their way into ad networks and Google AdWords and other "bulk" media spaces as an easy if questionable source of revenue for the owners of those sites.
But in the wake of the recent presidential election, fake news has come to the fore of our attention. So much of how the election turned out, and how so many of us -- and the media -- conducted ourselves in relation to digitally distributed content surrounding it has led to a grand questioning of not only what is content, but what is truth.
We aren't talking about that, though. The question before us is what kind of impact does fake news have on digital advertising?
The primary impact is not specific to "fake news" itself as a category but to the quality of advertising supported content in the ecosystem in general.
The question of fake news leads to a reexamination of what kind of inventory is in the marketplace that is sourced and served programmatically.
For a long time, there has been no real incentive to address the issue of quality in the value chain of programmatically sourced and served inventory.
- Technology and platform providers make money on the transactions, both coming and going.
- Publishers make money on market presence.
- Agencies make money on committing the tasks of transaction, and for holding companies with a trading desk interest, there is the opportunity to double dip, making money on both transaction and the spread on inventory.
- And the client gets to show positive results on the perennial desire to manage per unit costs.
But everyone loses here if for no other reason than that the data being generated from these questionable -- or non-existent -- sources is dirty, creating a body of false positives against phantom signs. The data are all noise and no signal. This leads to making decisions with the wrong inputs because of false outputs. What the fake news hubbub does is shine a light on the deeply shadowed places of an ecosystem that rewards a variety of bad behavior, whether it's fake news, headless browsers, bots, etc.
The other impact concern over fake news will have on digital advertising is that placement matters.
For all the reasons above, brand has been decoupled from context. In the quest to manage risk in the form of costs -- costs of inventory, costs of managing it -- advertising was untethered from place. Machines somewhere in the ether-sphere exchange data and inputs, and turn them into outputs. So long as those outputs are consistent with a numerical objective, that is all that mattered. And that would be all that mattered if the concern were only machines and the numbers they crunch. But where a brand appears does matter when humans get involved. Not all human behavior and human reaction can be rendered into machine-readable form. We feel things. We react. We are irrational, only to rationalize later. Environment is something we feel, react, and can be irrational about. Editorial adjacency used to be a default setting on all advertising. This was true in the early stages of digital, too, when planning online was a lot like planning print. But when the data got more voluminous and the technology to manage it all got both better and more complex, thinking shifted to a place where a machine-construct of who you were talking to mattered most. Maybe this is the inevitable conclusion of data saturation running into a faith in one-to-one marketing.
But there is such a thing as the halo effect. The halo effect is cast onto a brand from the content it appears with. Sometimes that halo is light, sometimes a shadow. Knowing which kind of content relative to what kind of brand is necessary in knowing which.
The blight of fake news isn't going to be salved overnight, and it's not up to us to address it. But it is up to us to think harder and more seriously about what we are feeding on in the digital advertising ecosystem and make sure that its truly healthy for our brands.