Just as counterfeit handbags sold on Canal Street come at a deep discount, fraudulent ad impressions cost a fraction of the real thing.

Between April-June 2018, cybersecurity company Confiant analyzed more than 50 billion ad impressions across thousands of websites and found that, on average, the CPMs that publishers received for impressions that intermediaries later misrepresented were 54% lower than the CPMs they received for impressions that were correctly represented in programmatic auctions. These discrepancies are driven by outlier impressions with extremely low or high values, according to Confiant CPO Jack Cohen Martin. Because the price distributions for ad sellers hoping to avoid fraud follow similar patterns, raising price floors isn’t a panacea for ad buyers hoping to avoid fraud.

To make their impressions attractive to ad buyers, ad fraudsters often sell their impressions at a discount. Ad buyers looking for cheap inventory purchase these impressions thinking they got a good deal, only to find out that what they bought is mislabeled inventory of poor quality.

"It's true that if you look at low CPMs you will see a higher percentage of ad fraud occurring, but the fraud you are seeing is unsophisticated and only a small proportion of what is out there,” said Barbara O'Malley, technical product manager at ad agency Infectious Media. “Raising floors will block the fraudsters who execute fraud poorly, but done well they will do their utmost to appear as quality traffic, and there is no one approach to take that will address it.”

Ad buyers must tailor their ad fraud fighting approaches to specific situations, according to O’Malley. If the goal is stop bot traffic, then advertisers should block the IP addresses of where they see the fraud coming from. If domain spoofing is the problem, then the advertiser should blacklist the suspicious sellers of inventory it comes across, she said.

Confiant found that 1.5% of impressions defrauded publishers and involved video ads being inserted into display ad slots. Because video ads have higher rates than display, fraudsters in these instances aim to profit off the difference between the video and display CPMs. Another 0.5% of impressions were infected by malware. An example of this is where an ad forces website redirects.

To avoid fraud, brand advertisers need to stop treating scale as the only metric that matters, according to Warren Zenna, founder of Zenna Consulting Group. If scale is absolutely critical to a campaign, then advertisers should consider TV and radio because they are less prone to fraud than online banner ads. Ad agencies must also be realistic with their brand clients about the tradeoffs of the cheap reach that can be found online, Zenna said.

“Fraudsters know that you want cheap impressions, so they create them for you out of nothing,” he said.

Ad fraud remains one of the biggest issues in digital marketing. In a January 2018 survey of more than 1,100 marketers worldwide conducted by Rakuten Marketing, 27% of the respondents surveyed said that ad fraud will pose one of the greatest threats to marketing this year.