Digital advertising fraud will cost marketers as much as $40 billion annually by the middle of this decade, says industry analysts such as Juniper.
But one of the world’s leading independent ad fraud experts, Dr Augustine Fou, says the industry is not incentivised to solve the problem.
“We haven't solved ad fraud, because we don't want to,” says Fou.
“If you're an Ad Exchange, you make more money if you clear more transactions, just like Wall Street, right? So a lot of the middlemen have zero incentive to solve the fraud. And in fact, there's a cottage industry of fraud detection companies who depend on fraud to continue so that they can keep making money.”
Marketers are tolerating millions of dollars in loss through digital ad fraud because the only goal, according to Fou, is “higher volumes and lower prices.”
“If we didn't buy from programmatic exchanges in the first place, if we went about buying media from good publishers, almost like as if it were 1995, then most of these fraud issues would not exist. But yet we've gone down this path and so we're basically throwing more and more tech at it, trying to solve a problem, but the tech is not going to solve it until we solve the incentives issue, because everyone wants it to continue.”
Fou describes ad fraud as a “multi layered problem” that has some tell tale signals.
Businesses can examine their own analytics through tools such as Google Analytics, and the campaign may show a dramatic increase in visits to the site, sometimes as high as a 100,000 per cent increase. But, the increase will be only from one type of device – such as only from Android.
“How's that possible? Right? Why are all the Android devices clicking and coming to your site? So in those cases, you can start to see something's wrong, and then delve deeper into figuring out why.”
Fou also highlights log data as an example of an ad fraud signal.
“If you look at log level data, you can start to see, okay, well, this visitor looked at 1000 pages within one minute, a human can't physically do that, right. So then you can start to see some of these tell tale signs.”
As the first step towards solving this problem, Fou encourages businesses to involve their IT teams in the marketing programs, in order to locate fraud that may be affecting the campaigns.
“At a bare minimum, if more of the IT folks with IT expertise get involved in some of the marketing programs, they can actually start to pick out some of the fraud that's impacting their campaigns without any further specialised tools, without any additional cost. And I think that kind of partnership with the marketing department would be very fruitful.”
While some companies are investing in fraud detection services, Fou suggests that many of these companies are only looking for invalid traffic, and missing the fraud occurring in mobile and CTV for example.
“In these newer channels, the completeness of detection is actually pretty low. And so in those cases, if the bot detection technologies are tuned for only looking for IVT, they're going to miss other forms of fraud, simply because they're not looking for it.”
Fou also explains the crossover from ad fraud into cybersecurity, where malware writers are no longer seeking to harvest passwords, but to instead load ad impressions while mobiles are turned on, but not in use.
“People don't turn off their mobile phones, right, they turn off their laptops at night, but they don't turn off their mobile phones. And the mobile phones typically have constant internet access either through their home Wi Fi or their wireless. So in that case, the malware writers can just load ad impressions or load web pages in the background to continuously make money through ad fraud.
“I think if we have more cybersecurity people or infosec people involved in helping solve the problem of ad fraud, it's going to help dramatically, because that's how those two topics actually intersect.”
