• by Tyler Loechner , Staff Writer @mp_tyler, March 5, 2015

Sites that contain a hefty amount of fraudulent traffic, naturally, give advertisers little bang for the buck. But new research suggests that’s not just because of the fake traffic.

The actual humans that visit sites with higher levels of fraud traffic don’t perform well either, contends Fraudlogix, a fraud prevention firm for programmatic marketers.

Fraudlogix says it analyzed campaigns from the financial services, consumer electronics, apparel, home improvement and transportation verticals for a study comparing fraudulent traffic rates to actual human behavior. “The campaigns included in the analysis ran real-time, CPM display ads across 13 different exchanges, displaying on 144,992 unique websites,” the company claims.

Fraudlogix found that 98% of all conversions occurred on sites that had fewer than 9% of fraudulent traffic, but these sites only received 64% of the budget. Sites with fewer than 9% of fraudulent traffic accounted for 72% of all sites studied for the report.

“This research shows that in digital advertising, ‘all real traffic isn’t created equal,’” stated Hagai Shechter, CEO & founder, Fraudlogix. “It appears that sites with higher percentages of fraud have questionable practices when buying traffic most likely as a result of trying to buy traffic at bare minimum prices, leading to poor performance even when the traffic is real."

The numbers are eye-popping -- they essentially suggest that sites with over 9% of fraudulent traffic have no hope of leading to conversion -- but there’s a significant piece of data missing.

Fraudlogix only analyzed conversions for the report -- they didn’t analyze the number of consumers or visitors per site. In other words, if 98% of the consumers that were a part of this study visited the sites with under 9% of fraudulent traffic, then the conversion numbers would line up perfectly, and the claim that "all real traffic is created equal" would actually be validated by this report, not debunked. The report lends credence to the claim, but it's certainly not conclusive.

Even with that crucial piece of data missing, some conclusions can still be drawn. Namely this: Programmatic marketers are not properly accounting for fraud in their budgets.

The fact that marketers put 64% of their budget toward the sites with 0-9% fraud suggests budgets need to be redistributed. Why spend only two-thirds of your money on sites that drive nearly all of the conversion?

It indicates two things: A) marketers could have spent less money for the same results or B) marketers could have spent the same amount of money but seen far more conversions if budgets had been properly distributed.

The budgeting decisions make even less sense knowing that each campaign studied “had [the common goal of] aiming for a conversion of a sale or a lead,” per a Fraudlogix representative.