Inventory quality (IQ) might be advertisers’ most pressing concern when it comes to programmatic, and with good reason. Recent years have seen reports of ads from reputable brands appearing next to objectionable content, as well as instances of bad actors ripping off advertisers with non-human traffic and other forms of ad fraud.
Until now, most people have discussed those two issues – objectionable content and fraud – as two distinct parts of the larger IQ problem. But by analyzing data from the more than 200 billion impressions that come across our platform every day, we’ve observed that ad fraud and fake news are inextricably linked – and they meet at the intersection of viral content and traffic generation.
When you strip away all the noise around fake news, you’ll find that the publishers behind these sites have a similar business model to the ones who flood your social media newsfeeds with clickbait videos and listicles: acquiring impressions as cheaply as possible and selling ads against them. In addition to relying on social shares, many of these sites resort to shady, sometimes fraudulent strategies to achieve that goal, and advertisers are left bearing the consequences.
Using this insight as a jumping off point, we’re working to identify fraudsters and remove them from our platform. Below is a sneak preview of our findings – to get the full report, along with tips on how to avoid fraudsters, you can download our latest white paper.
The viral content business model
Viral content refers to any content that’s designed to acquire traffic by having readers share it with others, usually on social media. Most web sites, including reputable ones, dabble in viral content to some degree.
But problems arise with publishers who make viral content the centerpiece of their business model. While most publishers are looking to build up a long-term audience by consistently putting out high-quality content, that requires more of an investment than viral publishers are willing to make. Instead, they’d rather acquire a constant stream of new users as cheaply as possible so as to maximize the profits on their ad sales. In that way, they function more like arbitragers – pay $1.00 to acquire a visitor, sell an ad on the impression for $1.50, and pocket the difference.
That need for new readers incentivizes viral publishers to make their stories as shareable as possible, and science shows that the content most likely to be clicked and shared is that which gets an emotional response from users. In fact, the data suggests negative emotions like anger are the most effective.
Fake news and incendiary political content fit that business model perfectly. After all, people care deeply about political and social issues, so content that plays on that is likely to go viral and attract the traffic these sites rely on. The modus operandi for these sites is to take a kernel of truth from a real news event, wrap it in hyper-partisan commentary, and pray that it makes readers happy or angry enough to read it and share. We’ve observed several publishers who run both conservative and liberal-leaning versions of these sites, suggesting that they’re motivated more by money than ideology.
How viral content can lead to fraud
Viral content is a boom or bust proposition. Business is good as long as there’s a steady stream of successful, shareable stories bringing in those cheap impressions. But when these sites hit a cold streak, the audience quickly disappears – remember, most people click these sites once, read the story, and then leave, never to return.
When the impressions start to dry up, viral content publishers often look to buy traffic from third parties. While there are certainly legitimate ways to do that – PPC search ads on mainstream search engines, for instance – many traffic generators are actually sending publishers non-human traffic.
How much is difficult to say. Some providers, such as low-end traffic exchanges, are virtually guaranteed to send bots rather than people – publishers who work with them almost surely know that they’re getting non-human traffic. But others, like content discovery networks for example, operate in a grey area. These providers sometimes package non-human traffic together with legitimate impressions, and it’s often unclear whether or to what degree any of the parties involved — the traffic generator, the publisher, or the advertisers — can tell the difference or know exactly how much of the traffic is legitimate. That uncertainty is what enables the bad actors to profit from ad fraud, and it’s ultimately caused by the overall murkiness of the ad tech supply chain – since impressions go through several middle men before finally being sold to an advertiser, it’s difficult for any one of them to confirm their veracity.
How can we address this issue?
Viral content is just the latest front in an ongoing battle between fraudsters and reputable ad tech providers like AppNexus. Our industry is constantly coming up with effective new tactics to eliminate the latest scam, and we’re confident the same will hold true for viral content fraud. Industry-wide efforts like the IAB’s ads.txt initiative are already showing promise in keeping shady middlemen out of our supply chain.
As the world’s leading, independent ad tech platform, we’re in a unique position to learn how these fraudsters operate, identify them, and ultimately remove them from the digital advertising ecosystem. We log every single impression that comes across our platform and have processes in place to flag those that appear to be fraudulent. You can check out the video at the top of this post to learn more about the work our data science team is doing on this front.
We also recommend you download our latest white paper for a deeper dive on the connection between viral content, fake news, and ad fraud. Inside, you’ll find a comprehensive breakdown of what a suspicious viral publisher looks like, how fraudulent impressions make their way to reputable advertisers, and what publishers and advertisers can do to start protecting themselves now.