Digital ad fraud is a real and growing threat. By 2025, global ad fraud will reach an estimated $50 billion, according to the World Federation of Advertisers. Advertisers, agencies and publishers obviously have to take steps to protect against ad fraud, otherwise they will waste a greater number of ad dollars on worthless impressions. The mistake they make, though, is failing to understand the potential downsides of these tools, and to test and compare different solutions.
For leading brands and agencies, it is basically a given that they will partner with a name-brand ad verification solution. (This is advertising, after all. Brand cachet matters.) The thing is, these solutions aren’t always effective, and in the worst cases, they can actually do more harm than good.
The Risk of Oversimplifying
Determining whether traffic is fraudulent is complicated. Often, ad verification companies focus on one or two key metrics, such as viewability or nonhuman traffic (NHT). While this makes it easier for advertisers to understand how the solution works, it also makes it easier for the bad guys, who are continually developing tactics to beat the system. Fraudsters are fully aware of how the industry measures traffic quality, and they know what they need to do to circumvent fraud prevention and detection tools.
Let’s pick on viewability for a moment, an important concept for reasons beyond fraud. Leading advertising organizations, i.e., ANA, 4A’s and IAB, worked together to define what counts as a viewable impression, the goal being to ensure advertisers aren’t paying for ads that were never seen. But who cares if your ad was viewed if it wasn’t a human who viewed it?
While more nascent fraudsters may build bots that show their hand with suspicious patterns, such as clicking too quickly, sophisticated criminals know how to design scams that appear legitimate and satisfy the industry’s viewability criteria.
The Problem with Percentages
Many ad verification solutions use probability statistics to rate the likelihood of a user being fraudulent. But what should an advertiser do if a user is 70 percent likely to be bad? How about 50 percent? There is no definitive answer—you might as well flip a coin. It is up to the advertiser to determine its level of risk tolerance.
When advertisers focus on super-clean, premium traffic, they throw out some good impressions with the bad and can end up spending more to drive the same results. Imagine you’re using a solution that scores traffic on a scale of one to 100. Traffic with a rating of “one” is definitely bad whereas traffic rated 100 is super high-quality. Now picture a triangle. At the top is the 100-rated traffic. There is not a lot of it. At the bottom of the triangle is the worthless traffic. There is plenty of that. No one wants the bottom-of-the-triangle, garbage traffic. Everyone wants what is up top, but there is only so much to go around, and the high demand drives up prices.
Focus on Performance
So, what is the solution? First, marketers need to understand the potential shortcomings of ad verification software. They are doing their best to measure the legitimacy of an impression, but no solution is perfect, and the bad guys are really good at what they do. That means marketers need to work harder. As much as they would love a “set and forget” solution, they have to be prepared to test out their strategies. Don’t rely on a tool’s reputation. Compare it to others.
To test ad verification solutions, measure traffic performance by monitoring something more verifiable than an impression: conversions. If the ad fraud solution is working, the traffic it sanctions should perform well, or at least in line with historic benchmarks. If fraudulent impressions are sneaking through the system, or if good traffic is being filtered out, that will affect the conversion rate.
Although surprising, it’s not a given that brands and agencies test out an ad verification tool before they purchase it. It does require more upfront work for marketers, but it’s worth it. Marketers will know they have a strategy they can trust—at least for a while. Ad fraud is fluid, so the defense needs to be, too.
Also read: The State of Ads.txt: Not a Big Clean-up Yet
For instance, P&G and JPMorgan Chase, two well-endowed brands, recently called for a review of their digital marketing efforts. P&G cut more than $100 million in largely ineffective digital ads without making much of an impact to their bottom line, and Chase vetted 400,000 websites where the company’s ads were running, only to determine a mere 5,000 were indeed ‘brand safe.’
Ad verification software is important and well-intentioned, but people have to understand how these tools work and not just choose one because they recognize its name. Then they can adopt better strategies and use anti-fraud tools the way they were intended – to help them spend their advertising dollars more effectively, and safely.