It’s hard to overstate the problem posed by fraudulent and poor-quality traffic to digital marketing campaigns. Addressing this threat is impossible without 24/7 monitoring of digital activity and real-time response, something nearly impossible for marketers to do on their own.
It’s widely known that ad fraud is a massive problem in marketing today, one that experts estimate will cost companies $40 billion in advertising investment this year. By setting up fake websites packed with high-value keywords, fraudsters are able to fool the algorithms that drive programmatic ad platforms into buying units that aren’t visible to human users. Bots are then programmed to click the ad hundreds of times, generating an artificially high CTR and bilking the company out of thousands of dollars at a time.
Schemes like these are so difficult to root out that they’re becoming the second most common form of organized crime. As Eli Martin explains in Entrepreneur, “With the prospects of a high payout, low risk and relatively little effort needed, it’s easy to see why so many organized criminals are looking to profit from ad fraud.”
But while the term ad fraud may refer specifically to bot-driven traffic on programmatic display ads on desktop and mobile, tackling the malicious draining of ad investment by criminals is made more difficult by defining the problem so narrowly. To protect themselves against fraudulent traffic, both human and non-human, marketers must understand the characteristics of different fraud schemes and learn how to detect and react to fraud threats in real time.
Bot Traffic
$40 billion already seems like an outrageous figure to assign to siphoned investments by bot-driven traffic, but research cited by the World Federation of Advertisers (WFA) suggests the problem is even worse than it sounds. Among the WFA’s findings were that bots inflate monetized audiences by as much as 50%, that fraudulent bot traffic is experiencing year-on-year growth of 22%, and that such traffic accounts for up to 61.5% of website traffic generally.
Programmers typically establish rules for programmatic software to follow that help it to distinguish human and non-human traffic. The problem with this approach is that fraudsters are always looking for new ways to get around these rules, and each time they find one, programmers simply have to add more rules to prevent bad actors from exploiting that workaround again. And since these new attacks typically happen at hours during which there’s unlikely to be a human supervisor at the till, this inefficient process often reacts long after huge chunks of ad investment are sunk into fraudulent traffic.
Click Farms
If you’ve ever seen an ad offering you the opportunity to “buy likes” or followers on social media, it’s likely that what you encountered was a click farm. Workers at click farms, often enticed by scams promising the ability to “work from home,” are paid to click on ads and “like” paid promotions on Facebook. In the case of social media, these schemes hurt the visibility of your promoted posts by adding likes without engagement, in addition to draining ad budget.
Click farmers are harder to detect because computers have a difficult time distinguishing their behavior from that of real users. Marketers need smarter tools in order to detect fake human traffic and followers, observing which users actually engage with your web content and which ones only click through or like, blacklisting websites that appear to be associated with click farms and blocking suspicious accounts in real time.
Poor-Quality Traffic
While not technically qualifying as outright fraud, low-quality traffic is another factor that lessens or dilutes the impact of digital ad spend. There are plenty of websites that pose as real publications in order to get as much traffic revenue from advertisers as possible in a short amount of time, and succeed because most programmatic buying software isn’t adept at determining website quality.
Given that all these negative influences tend to impact programmatic advertising the most, some have suggested that programmatic be dropped altogether in favor of direct relationships with publishers. But the truth is that programmatic is still an indispensable channel for marketers, and that ad fraud is not an insurmountable obstacle if you have the right tools.
Artificial intelligence in programmatic marketing offers a solution to all these variations of fraud. Powered by machine learning, an A.I. platform from Adgorithms could monitor digital activity 24/7 and blacklist websites being affected by bot-driven fraud in real time. It could learn from fraudulent behavior from the past and anticipate it, rather than constantly reacting to fraud after the damage has already been done. It could optimize its media purchases to get the highest customer value with each buy, thereby learning which low-quality websites to avoid.
Ad fraud is one of the biggest challenges facing the marketing world, but it isn’t one that companies are powerless to stop. A combination of increased transparency from media companies, more rigorous standards, and high-powered marketing software will limit the risk that organizations face from this new threat.
