Marketing

The frighteningly overblown reports of online ad fraud

Image Credit: Cranach/Shutterstock

If you have read anything about online advertising in the past year, you may be afraid that the industry is about to collapse under the weight of fraud. Headlines scream that robots are the source of alarming amounts of ad impressions — and a story even ran here on VentureBeat recently outlining the various types of ad fraud.

But take a closer look, and you’ll see dozens of Chicken Littles claiming that the sky is falling without much evidence to back it up.

Fraud is an ongoing concern in this space, and one that advertisers do need to confront. The bigger issue is that the coverage to date has been hyperbolic, overstating the issue to the point of mass paranoia. Fraud is nowhere near as large as they would have you believe, and the current spotlight is a result of sensationalist coverage and a score of new or pivoting tech providers trying to cash in on that coverage. It’s time to kill the fraud frenzy.

The most recent can of gasoline to get tossed on the fraud fire came during the IAB Annual Leadership Conference in February, when Ziff Davis CEO Vivek Shah cited a comScore stat that 36 percent of all web traffic is fake. That remark came as part of a larger, detailed call to action, but “36 percent” became the sound bite that everyone latched on to, from Twitter to the Wall Street Journal to Business Insider.

Most reporters attributed the stat to Shah and/or the IAB, even though he mentions comScore, as you can see in the video below. An alarmingly small number of reporters sought out the source of that stat, which comes from an October, 2012 blog post by comScore’s Brian Pugh. Fewer seemed to care that their headlines were broadcasting a 16-month-old figure. Even worse, the comScore report indicates the problem could be significantly smaller – as low as four percent in the U.S., as Pugh writes.

If fraud is only four percent, then it’s not worth all the hand wringing. None of this is Shah’s fault. He used a stat to help illustrate the IAB’s current initiatives and contextualize a known issue. The problem is that news outlets clung to one small statement without verifying it, fanning the flames of fraud paranoia and creating a disturbing precedent for covering the issue.

Another example is the commonly cited “fact” that fraud costs advertisers $6 billion annually. This comes from White Ops, a company with a stated goal to “eradicate automated fraud,” and part of a wave of new or pivoting tech companies looking to capitalize on the storyline. Then there are reports that half of all traffic originates from robots. That’s according to Solve Media, a company that profits off of CAPTCHA-based advertising.

These stories show two companies that profit from fraud detection saying that fraud is a very, very large issue. Because without fraud, what would these companies do? They need to scare up some business, or else they’ll be stuck stocking the shelves at science fiction bookstores. But the damage is done, as some of these numbers are now even appearing in sales material and presentations across the ad tech industry, further feeding the frenzy.

The bias is obvious here, and I’m not trying to throw the reporters covering this issue under the bus. I’m merely pushing for some accountability across the spectrum. This includes the publishers and advertisers themselves who get caught up in the “sky is falling” mentality without looking at the evidence in front of them.

All parties are confused, and one major reason is that there are many definitions, and therefore no standardization in detection methodology. Half of the traffic that Solve saw during one quarter may have been robots, but that doesn’t mean it applies to every ad impression across the web.

On top of that, the dollar amounts being thrown around are assumptions based on assumptions, and it’s not hard to shoot holes in them. Media industry commentator Tom Foremski uses faulty logic to equate fake traffic to lost money on an apples-to-apples basis, so that when the industry hits $50 billion, 36 percent of the total spend ($18 billion) will be lost to fraud.

That’s clearly not the case. Verizon, the eighth-largest advertiser in the U.S., spent $1.2 billion on ads last year and found that $1 million went to fraud, according to the Wall Street Journal. We know that AT&T, the third-largest advertiser, spent $1.8 billion last year. If 1 percent of its total ad spend were lost to fraud – similar to Verizon – that would come to $1.8 million. That’s still less than $3 million total in losses between two of the top eight advertisers. The tail of online ad spending is simply not long enough to push losses into the billions.

Going even deeper into the issue, only ads sold on a CPM basis are subject to impression fraud, and only 32 percent of ads are sold this way, according to the IAB. Using the most recent figure of $43 billion in total online ad spend, that leaves a little more than $14 billion that is even susceptible to fraud. If that oft-quoted $6 billion figure is true, nearly 43 percent of CPM spending is wasted. And if $2 out of every $5 spent on a CPM basis were lost annually, it would seem that the FBI or another government body would take interest. You can be that if Google is losing more than a billion dollars to fraudsters, it’s going to flex some lobbying muscle.

Finally, we also have to take into account the number of players in the space. Think of all the companies clogging the LUMAscape for the $43 billion online ad industry. A $6 to $18 billion fraud problem would result in a booming anti-fraud industry, much larger than the handful of companies currently addressing the problem, many of which will admit that online fraud is only a piece of their business. None of it adds up.

If advertisers actually used the tools available to them, they’d be in position to weed out fraud even further, which would allow us to stop focusing on this conversation and move on. Fraud will persist, one way or another, but it does not have to be seen as a harbinger of the collapse of the online advertising industry. It’s something that needs to be dealt with proactively, and knowledgeably. Unfortunately, sensationalism and new-business hype are creating a lot of noise in the market. Advertisers right now are best served looking at their real numbers and viewing everything else with a great deal of skepticism.

Alex White is VP of Product Strategy at Sizmek.


VentureBeat is studying the state of marketing technology. Chime in, and we’ll share the data.
7 comments
todd chu
todd chu

What a silly, shill article.  Sizmek simply lacks the data to understand the ecosystem: their verification "technology" is a non-factor.  The company does a fine job with rich media and video serving but employs no hard-core security engineers.  I consider this a chump-vertorial

Tal Franji
Tal Franji

1. CPC is subject to click fraud. Anyone who does not acknowledge this does not know the industry

2. The financial interests for ignoring fraud are larger than the "fraud industry" - advertising industry is driven by agencies spending as much budget as the can.

- these makes the above analysis quite questionable. 

Digital Media
Digital Media

You must trolling with this article


- We know that AT&T, the third-largest advertiser, spent $1.8 billion last year.

On Display solely? Probably not.


- If 1 percent of its total ad spend were lost to fraud – similar to Verizon – that would come to $1.8 million.

Your math is *still* wrong after your update.


- Going even deeper into the issue, only ads sold on a CPM basis are subject to impression fraud, and only 32 percent of ads are sold this way, according to the IAB...

You don't think CPC and CPA are susceptible to fraud!?


- If advertisers actually used the tools available to them, they’d be in position to weed out fraud even further?

so fraud does exist? also weren't you bashing fraud detection / prevention companies a few paragraphs up?


I think you meant to post this on theonion.com/ad-tech

Digital Consigliere
Digital Consigliere

"The bigger issue is that the coverage to date has been hyperbolic, overstating the issue to the point of mass paranoia. Fraud is nowhere near as large as they would have you believe."  


This is not correct.  From actual data, fraudulent ad impressions could be as high as 100%.  There is no hyperbole in that. But there is also fraud that is in the single-digit percentages too. Some clients and their media agencies use the "good guy" technologies as well as good policies and practices to manage fraud very well. 


The "good guys" include:

Spider.io (owned by Google)

Pixalate

WhiteOps

Adometry (owned by Google)

Improvely

Solve Media

PageFair

Nielsen/IAB

Integral Ad Science

DoubleVerify

YouTube (Google View Through) - does not charge for fake views or skipped ads

Digital Consigliere
Digital Consigliere

"You can be that if Google is losing more than a billion dollars to fraudsters, it’s going to flex some lobbying muscle."  This is not correct.  


Google is not losing money due to fraud. In fact, many companies in the value chain are also making money due to fraud because even those fraudulent ad impressions generate ad revenue which is shared between the publishers' sites and the ad networks. The more ad impressions, the more money for everyone. 

Digital Consigliere
Digital Consigliere

"If advertisers actually used the tools available to them, they’d be in position to weed out fraud even further" is true. 

Joshua Ellinger
Joshua Ellinger

The math on the AT&T line is wrong 1.8B x 10% = 180M, not 1.8M.