The never-ending fight against advertising fraud

The Association of National Advertisers had good news for Cannes Lions audiences earlier this month. In 2023, media buys on predominantly fake advertising-focused (MFA) sites accounted for an extraordinary 15% of ad spend. A year later, that share had dropped to 4%. On a slightly murkier note, digital ad spend is expected to grow by 16% in 2024, 4% of a larger number. And 4% is still a lot.

Of course, tricking brands into buying programmatic inventory on sites with little or no engaging content is just one type of ad fraud, and while there are varying opinions about the scale of the problem, ad fraud still consumes large portions of media budgets.

In addition to the use of MFA sites, there are the following types of ad fraud (but not limited to):

• Click fraud: The use of bots to generate large numbers of clicks on PPC ads, making them appear to be reaching a large human audience.
• Domain spoofing: Imitating a high-value domain. Clicks and users may be real, but do not represent a high-value audience.
• Pixel fill: Multiple ads are grouped into a single pixel frame. When a user clicks the frame, they click on ads that are actually invisible to them.
• Ad stacking: Similar to pixel stuffing, except the ads have normal dimensions. Only the topmost ad is visible, although clicking on the topmost ad effectively clicks all of them.

A very big deal

If you think ad fraud is a marginal problem, take a look at some of the higher estimates of its magnitude. Analysts at Juniper Research reported that about $100 billion will be lost to ad fraud this year, or about 22 percent of media spending. Forecasts suggest the percentage will remain relatively stable (23 percent) through 2028, but with media spending increasing rapidly, that represents a loss of $172 billion. Even if Juniper overestimates the loss by as much as 50 percent, that’s still an extraordinary number.

To paraphrase John Wanamaker, “Half the money I spend on advertising is wasted. No, let’s say 75%.”

We received the Juniper report from Ann Tarasewicz, CEO of programmatic platform Axis. When we spoke to her, she had been temporarily relocated to Romania from her hometown of Odessa, Ukraine. “The current state of ad fraud is a serious challenge for the entire industry. Fraudsters are trying to be more advanced, organized and on the cutting edge.”

How can brands accept this situation? “Of course, brands are primarily interested in volume, but I would say if a publisher is involved in low-quality traffic, it can damage the brand’s reputation.”

Are there any people watching?

Back in January, MarTech author Greg Krehbiel cited another extraordinary statistic from Bob Hoffman’s book “Adscam”: 3% of digital advertising spend actually corresponds to ads seen by people. “No, let’s say 97%.”

“This is a very bad problem,” Krehbiel said. “Since I published that article in January, there was the Forbes story where they basically created an entire website that nobody saw and it served up all these ads.” Forbes “mistakenly” ran that website for several years.

Maybe brands are willing to overlook the waste of ad dollars as long as campaigns are generating lift? “It doesn’t matter how much fraud there is. If it’s 99% fraud, the 1% is OK,” Krehbiel said. “So what’s the problem? I think it all boils down to whether people are basing their decision on whether to continue a campaign on numbers from advertisers or whether they’re basing their decision on actual results they see on their site.”

What should I do?

Tarasewicz lists a number of options for combating ad fraud. “Unfortunately, there is no one-size-fits-all solution,” she says. “Brands should combine options to address this problem.” She praises collaborative approaches like that of the Trust Accountability Group (TAG). “It brings advertisers, publishers, law enforcement agencies – all market participants – together. By sharing their individual knowledge, they can understand how to better prevent fraud. The group was founded in 2015 and is quite effective.”

Specific fraud prevention methods include the use of artificial intelligence and fraud detection algorithms, as well as ad verification partnerships. In theory, this approach uses machine learning, data analytics, and pattern recognition to evaluate the validity of impressions, clicks, and conversions in the hopes of detecting and blocking fraudulent traffic in real time.

The AI ​​approach

The problem with AI is that it’s also available to fraudsters. “They’re trying to be on the cutting edge here,” Tarasewicz said. “Where brands update their mechanisms, fraudsters are trying to test and adapt. But even if they can adapt to one mechanism, it’s possible that a different fraud tracking system will show a different result. That’s why it’s best to use a combination of different algorithms.”

Today, advanced bots can not only generate impressions and clicks, but even place orders or fill out forms. “It’s really difficult to tell if it’s a human or a machine.”

AI appears to be effective when used by banks to detect financial fraud. “The difference in this case,” Krehbiel stressed, “is that all the transactions converge in one place. The problem with ad fraud is that they claim there was a transaction, and I have no idea if it even happened.”

The approach to ad verification

Partnering with ad verification companies and integrating their solutions into ad exchange platforms can help – if they work. Even one established verification partner recently had to apologize for errors. “The different ad verification companies use different mechanisms,” Tarasewicz said, “and we don’t know exactly what they use. My suggestion is to use a combination of tools; not just one, but at least two. Unfortunately, different tools tend to produce different results, especially with different ads and traffic types.”

Tarsasewicz knows of an ad verification solution used by her clients that is not suitable for scanning CTV traffic but specializes in checking in-app advertising. “However, other tools successfully scan CTV but show results with a 60-day delay. There are different situations across the market. It is recommended to check the reputation of the tools across all channels.”

“I don’t want to be too negative,” Krehbiel said, “but the concept of an ad verification network trying to figure out whether an ad is being shown to humans is problematic because the network itself is a bot. I’m sure that’s better than nothing.”

Krehbiel agreed with Tarasewicz that these tools probably won’t be effective on all channels. “No one can be an expert on Twitter, Facebook and web advertising. It’s going to depend on whether they’re focused on B2C or B2B, whether they’re targeting broad or niche consumer interests; all of those things will play a role.”

Dig deeper: Online advertising: The fun, fuzzy math

Despite Tarasewicz’s advice to use a combination of prevention solutions, would Krehbiel still be pessimistic about the situation? “Yes and no,” he said. “If you have several different signals pointing roughly in the same direction, that tells you something. If an ad network tells me my ad is being shown so and so many times, and the fraud detection folks say that’s pretty much true – and then another fraud detection folks say that’s pretty much true – it would be petty for me to contradict all of that.”

Of course, that raises the question of cost. “Not only are you advertising, you’re also detecting ad fraud. That changes the way you calculate the ROI of advertising. Let’s take a Pepsi ad at the Super Bowl. Your measure of success is whether Pepsi sales increased. That’s a far better way to decide whether an ad is making a difference, in my opinion, than these technological solutions.”

To be clear, advertising works. Everyone has had the experience of seeing an ad and making a purchase as a result. It’s not just a cry into the void. “We know for a fact that advertising is and can be effective,” Krehbiel confirmed. “The real question is, is it as effective as we’re told?”