Ad Fraud and Its Impact on Local Media Spending

Digital ad fraud, by some measures, is an extremely attractive “business” boasting both high growth and high margins.  Ad fraud is when ads are sold but not shown to humans or not shown at all, but still billed to the advertiser with proceeds going to the fraudster not the publisher. Of course, it’s also unscrupulous and can be illegal. It is also likely to be impacting local media publishers differently than national digital ad platforms.

Chances are, local media inventory is far less prone to desktop bot-traffic and mobile emulators and therefore offering inventory that is both brand safe and with much less invalid traffic. But while ad fraudsters have been targeting large national budgets for years, as more local marketing dollars shift into digital, fraudsters and criminals will follow the money. Recent shows that fraudsters increasingly are using highly sophisticated methods to escape detection and steal publisher dollars.

Ad fraud is a form of cyber crime. But surprisingly it can actually be legal, although others acts committed by fraudsters may not be. We recently saw the case where the U.S. Department of Justice indicted some digital ad fraudsters from Russia and Kazakhstan, charging them with an illegal scheme worth more than $30 million. That is a tiny drop in a big bucket of ad fraud.

In a report eMarketer issued this month, they cited estimates that global ad fraud runs somewhere in the range of $6.5 billion to $19 billion per year. And the thing is that it’s very difficult to detect and defeat fraudsters. They are very smart, adaptive and use leading edge technology. For example, one recent fraud innovation that’s been uncovered is “declaration fraud.” This occurs as buyers pay for full-screen mobile based on information about the ad unit but instead are served a unit that might be less than 5 percent of the screen.

In fact, another report out this month suggests that overall about 18 percent of online ad traffic is fraudulent.  This study from CHEQ, a cybersecurity company, also estimated that U.S. desktop-based fraud accounts for 55% of online fraud levels 46% is mobile ad fraud. Of this ad fraud,  77 percent of fraudulent ad traffic is “sophisticated invalid traffic” (SIVT). Experience shows that SIVT can sometimes defeat industry initiatives with ad verification services, use of tags such as ads.txt employed to mitigate this fraud and other attempted remedies.

The question BIA Advisory Services has investigated is how ad fraud may be impacting local media any differently than national media? Working with one of the cybersecurity industry’s leading expert in ad fraud, Dr. Augustine Fou, we’ve released two studies on this topic. To help further industry education and awareness, we’re making these reports available for free. Click here for more information.

The basic problem is that advertisers are deceived into buying ads they think are being shown to humans inside trusted published sites and apps. However, fraudsters using fake sites and mobile emulators appear to ad buyers as legit and offer cheap inventory that is apparently exposing users to their advertising. However, that traffic is bot-traffic, not humans. Humans never see the ads but buyers pay for the impressions. Advertisers lose, publishers lose, and audiences lose.

As Dr. Fou says in BIA’s Dec 2018 report: “As the saying goes, ‘you get what you pay for. The cheaper ad inventory is the result of ad fraud, where cyber criminals use fake mobile devices to load fraudulent ad impressions to rip off marketers. The mobile emulators are so advanced, they can download and install apps, launch and interact with those apps, and pass fake GPS locations and other sensor details.”

The bottom line for advertisers is that if what seems to be premium publisher inventory looks too good to be true, it’s probably not true.

For local media publishers, it turns out that for your own websites and mobile apps, the ad inventory is likely to suffer much less ad fraud. It’s because the fraudsters are less likely to target these media platforms and significant parts of the inventory are sold direct versus via programmatic exchanges. With higher quality inventory, local media publishers should not be afraid to purse the ability to command pricing premiums since a much higher percent of their traffic is valid human traffic.

In fact, we’ve recently chatted with a TV sales executive who manages digital sales for a group of network-owned local TV stations. In their case, they only sell their own audience from station and network websites and mobile apps. Given this high quality and likely largely fraud free audience, they charge a $65 CPM and not only get it but regularly hit sell-out. They explicitly avoid trafficking in third-party audience extension solutions because they think the ad fraud and brand safety risks expose their buyers to too much downside.

This approach may not be the best for every local publisher, but it certainly has its merits.

Now, local media are less in the sights of ad fraudsters. That is changing and high-tech countermeasures by publishers and ad verification vendors are in a constant race to defeat highly innovative and sophisticated fraudsters.

In our reports, here’s what BIA offers as considerations for marketers concerned with how to best mitigate risks from ad fraud:

1. If you are targeting local audiences by buying geotargeted ads in programmatic exchanges, and the CPMs are so low they appear too good to be true, they probably are. Build rules to detect this and support better buying decisions.
2. Marketers can also reduce their risk of exposure to ad fraud by buying from local publishers that have actual human, and local, audiences. Of course, CPMs will be apparently higher but that’s because you’re paying for human traffic, not bot-traffic.
3. There should be an obvious correlation between showing your ads to humans and higher ROI versus buying cheaper inventory shown largely to bots and fake mobile devices. Expect to pay a premium CPMs that comprise a higher percentage of human versus bot and mobile emulator traffic, actual human traffic is a scarce commodity.
4. Even though fraud detection services may mark traffic as valid human traffic, it’s important to go deeper. Examine your own analytics for signs of strangeness. For example, bounce rates that are too high or low, zero time-on-site visitors, etc.

For local publishers, BIA offers these considerations:

1. Be careful in “buying traffic” for audience extension solutions to gross up on targetable impressions.
2. Local media can take pride in the audiences they deliver in a premium content environment and seek to monetize this with premium rates. This takes educating the buyer. Among other things, marketers who are buying geotargeted audiences from exchanges obtain a different product than local publishers selling their actual local audiences.
3. As we suggest marketers consider going beyond what ad fraud detection services can do, we reiterate this for publishers. Ad verification services do an important job, but can’t do it all. Publishers need to run a variety of their own analytics to corroborate the percentage of genuine human versus fraudulent traffic they are providing to advertisers.

You can find our complimentary ad fraud reports here. If you’re interested in learning more about what you can do to detect and mitigate risks and impacts of ad fraud, either as a marketer or publisher, we’d be happy to discuss strategies with you.