The Amount of Questionable Online Traffic Will Blow Your Mind The World Wide rip-off

Click Fraud Detection PPCSecure

A few weeks ago, Lindsay Buescher, senior manager, analytics at Carat, read an article on Adweek.com about a company called FreeStreams.com that was pumping up its traffic by enticing Web users into accidentally visiting via hidden links on sites that house pirated content. As it happened, one of her agency’s clients, Red Bull, was a FreeStreams advertiser. Buescher was determined to find out what was going on. Her team discovered Red Bull video ads were running on FreeStreams through two different networks, including ValueClick, a publicly traded company. (ValueClick says it has since stopped working with FreeStreams.)

That wasn’t something Buescher had run into much. She kept digging. After about three weeks, Carat was blacklisting 77 more sites for Red Bull beyond FreeStreams. Many of the sites didn’t actually sell pre-roll ads, which was what the client was supposedly paying for. Some were merely gaming sites with interstitials. Some were sites that didn’t even exist or were blocked by her company. Others ran video inventory continuously. Many were simply a case of “fraud,” she found.

In the end, Buescher’s efforts resulted in not only a lot of legwork but also a $150,000 refund for the client. That’s hardly enough to set the industry on a different course, but to Buescher it said something. “Red Bull is spending 90 percent of its online budget direct with publishers,” she says. “Imagine if they weren’t.”

It might not matter either way. That’s because the online ad industry is facing a swelling crisis, one defined by fake traffic, bogus publishers and invisible Web visitors, a trend first investigated by Adweek in an online story, “Meet the Most Suspect Publishers on the Web,” published March 19 of this year. Once thought contained to a handful of rogue players that had figured out how to exploit ad exchanges, bogus ad inventory, as it turns out, is rampant. In fact, according to numerous sources across the ecosystem, fake traffic is essentially systemic to online advertising—it’s part of how the business works. And a slew of top companies are involved in this—whether wittingly or not. “You see it with almost any partner you work with,” as Alan Silverberg, media platforms director at Moxie Interactive, puts it. “From AOL and Yahoo to Facebook, from pure-play partners and the network space to portals. We can’t stop it,” he says, referring to the preponderance of questionable traffic. Though for many publishers, it may be a question of whether they can’t stop it, or won’t. Adds Buescher: “It’s the whole business. We see this even with direct-to-publisher deals. It’s really the media planners’ fault. But when you start seeing partners breaking contracts ... no one has time to monitor 3,500 sites when they are just cranking, cranking, cranking out plans.”

Of course, there are plenty of those who think this is being blown out of proportion. Some in the online ad world see bots, fake traffic and the like as a manageable nuisance, hardly a crisis. But the ranks of the alarmed are growing by the day.

During a recent interview, online ad veteran Wenda Millard, president of Medialink, made the bold claim that a quarter of the online ad market is fraudulent. “What we have found is the devaluation of digital media is causing us to lose about 25 percent of the roughly $30 billion that is being spent,” she reported. “It’s stolen [ad revenue].” In defining fraud, Millard lumped together piracy, nonviewable ads, ads stacked on top of one another, inappropriate content and, of course, deliberate malicious behavior, in her analysis. “In most people’s wildest dreams, they wouldn’t imagine how much [questionable traffic] there is,” she says. “People should be very, very worried.”

Let’s start with individual publishers. A group of Web-traffic analysts that did not want to be named for fear of pointing a finger at potential clients has identified a dozen well-known sites that exhibit questionable traffic patterns, including Break, CollegeHumor, Complex, Crackle, Entrepreneur and Totalbeauty. One source of information the group examined was comScore’s cross-visiting reports, which revealed that each site on the list has a high percentage of visitors who also visit sites suspected of living on bot traffic.

For example, Break, CollegeHumor, Complex, Gamespot and Crackle all have a high cross-traffic index with sites like Mommyhotspot, featuring parenting content, as well as Missoandfriends and Dreammining, a gaming hub aimed at young girls, plus a variety of sites frequently blacklisted by ad buyers. According to Augustine Fou, founder of the Marketing Science Consulting Group, Dreammining raises many questions. Using Alexa data, Fou found Dreammining’s top search term to be “mining of selena,” which exhibits zero traffic on Google, he points out. The site also has a high at-work audience that doesn’t match its demo, while the second- and third-largest domains driving traffic to the site are “possible click farms,” he says.

Fou also looked at Mommyhotspot, which had unidentifiable domains driving traffic and barely any inbound links, per his analysis.

The overall pattern Fou and others examined—sites like Dreammining and Mommyhotspot sharing audience with Break and Crackle—does not in itself directly implicate these branded sites in anything scandalous, though it might strike some as unexpected that a group of guys’ websites share traffic with several young girls’ sites, let alone girls’ sites possibly employing click farms.

But a deeper examination is revealing. According to an analysis conducted by a Web-traffic expert who did not want to be identified citing confidentiality agreements, all these sites, whether wittingly or not, have some level of “bogus traffic.” Research companies regularly discount traffic from Web publishers before reporting data like unique users or impressions because such traffic looks questionable. And even they don’t catch everything. In this case, an Internet security expert suspects these sites (Break, College Humor, etc.) might all be purchasing traffic from the same vendor, which is likely employing bots.

Adweek reached out to Ben Edelman, an associate professor of business administration at Harvard, who has done extensive research on Internet architecture, advertising, traffic patterns and fraud detection, and who has built a proprietary Web crawler he uses to detect curious activity. Edelman says via his proprietary tracking tools, he’s gathered evidence that Crackle, College Humor, Break and others regularly employ “invisible traffic”—i.e., these companies deliver their entire sites via iframes or tiny pixels on other sites that no one else can see.

Consider the guy-oriented Break (which last week was reported to be merging with Alloy Digital). Edelman found that Break.com loads invisibly through a variety of complex methods. In one instance, a company called Ptp22 redirects Break traffic through a variety of middlemen before it is loaded invisibly via iframe. (The list of go-betweens includes Marketwithmogul and TooShocking.)

Another company, Deplayer, conducts the same kind of operation. Adweek was unable to find contact information for the company.

Analytics firm SimilarWeb would seem to provide more supporting evidence. Some 25 percent of Break’s traffic comes via referrals, much of it from the adult category. The top referring site isn’t Facebook or Twitter but a firm called ClickSure, which sells Web traffic. In Edelman’s opinion, Break knows it’s buying such traffic. “But [this traffic doesn’t deliver actual] users who can see the site or any ads on the site. But there are indeed ads on the landing page. I saw ads for AT&T,” he says.

“We occasionally use the common practice of trade link sharing, which accounts for a small percentage of traffic,” says a Break representative. “We consistently work in the interests of our partners and have created high standards to vet traffic for quality issues. We have immediately suspended suspicious providers in question for further exploration.”

(UPDATE: Carat's Buescher says her agency has found that a campaign purchased through Break is running on three piracy sites: allmyvideos.net, divxstage.eu, and movreel.com. They admitted to it," she said. Adweek has reached out to Break to respond).

Edelman says he has been tracking Crackle for years. That site, which is owned by Sony and features a series starring Jerry Seinfeld, Comedians in Cars Getting Coffee, exhibits signs of invisible traffic, says Edelman. According to SimilarWeb, 18 percent of Crackle’s traffic comes from referrals. Among the site’s top referrers are Inttrax, Lnksdata, Redirect.ad-feeds and Contenko. Some of these companies are known for selling pop unders, adware and the like, per Edelman.

A source close to Crackle says the site cleaned up some of its traffic sources from years past but still occasionally encounters suspect traffic when marketing the site via various ad nets. The problem is described as “minor and the firm takes measures to address it.” Also, per sources, College Humor has noticed suspect traffic patterns from traffic purchased from outside vendors and has taken steps to eliminate it.

“Complex typically shares visitors with quality sites such as GQ.com and Vulture.com,” adds a Complex spokesperson. “In August we conducted a small, one-off traffic vendor test that delivered impressions that fell far below our standards, and we quickly terminated the relationship. We take seriously the matter of bogus traffic and we support any and all industry measures to address the problem.”

Beyond direct publishers with invisible traffic, what about the assessment from Moxie’s Silverberg that the Facebooks, Yahoos and AOLs of the world could be touched by the questionable traffic issue?

In Facebook’s case, the company has wrestled with scammers creating fake profiles and selling artificial likes. It’s a problem that’s mostly under control, according to a Facebook spokesperson: “As part of our ongoing site integrity efforts, we have recently updated our automated systems to remove ‘likes’ on pages that may have been gained by means that violate Facebook’s terms. On average, less than 1 percent of likes on any given page will be removed.”

Facebook is generally a closed environment. But it’s the wide-open, murky world of ad networks, ad exchanges and real-time bidding that experts say is rife with fraudulent activity. That’s because clients continue to look for huge quarterly CPM decreases and expect the Web to deliver cheap, efficient results. Meanwhile, agencies benefit from marking up ad inventory through their trading desks. Every vendor is plugged into every other vendor and platform, and sellers are incented to drive the most clicks at the cheapest rates possible. Thus, this ecosystem is practically begging for scammers.

“A certain set of conditions has led to a proliferation and explosion of this, and everyone will be touched by this problem to different degrees,” says Steve Sullivan, the Interactive Advertising Bureau’s vp, ad technology. “Add in retargeting and audience buying, it really represents a challenge in differentiating quality impressions vs. non quality.”

AOL and Yahoo both operate in the network exchange realm. And experts consider them among the good guys—the companies that have billed themselves as safe havens in this lawless space wouldn’t be vulnerable to bogus traffic. But exchange buyers say that even these companies can be vulnerable to questionable traffic. Take AOL’s Advertising.com, which has long billed itself as a premium exchange. However, ad buyers reveal that AOL campaigns sometimes feature chunks of inventory going to unidentifiable sites, reported back as “dummy publisher” or “house account.” AOL will also deliver major swaths of inventory to companies like Swaave, which will return a 0.0 percent clickthrough rate. CTRs are low these days, but not that low. A look at Advertising.com’s inventory mix includes properties like the live video chat platforms ooVoo and Tinychat, the Web-browsing disguiser AnchorFree, as well as companies like Integri and JCarterMarketing—essentially other ad nets. Meaning that buyers who come to Advertising.com for transparency may not find it, leaving it open to all sorts of deception. “We see these [kinds of] companies selling ads on 20 to 30 ad networks,” says Carat’s Buescher. “The percentage of inventory on our blacklists is astounding.”

“Whenever you buy from someone who won’t tell you where your ads are running, there is a real danger they are ripping you off,” says Zach Coelius, CEO of the ad tech firm Triggit.

According to one ad buyer, Ad.com’s supply currently includes questionable inventory and “copyright-infringing sites.” An executive from a major agency trading desk confirmed that Ad.com runs ads “on suspect sites and reports back traffic from unnamed publishers.”

AOL’s svp publisher services Dave Jacobs declines to discuss specific partners but says the company employs an in-house quality-review team that is constantly monitoring these issues. Jacobs adds that Advertising.com only works with other ad networks on a site-by-site basis—and never buys blindly. He was unable to discuss why some clients receive reports with traffic from the likes of “dummy publisher,” while maintaining that such a practice was not common.

“We think we are at the leading edge of protecting our partners,” he says. “We are very much focused on maintaining blacklists. With Ad.com, our focus is the direct-supply area. There may be instances where we can identify opportunities to deliver subset audiences outside our network. It’s not unusual for any company in our position to evaluate multiple sources of supply ... but with our network, we’ve been focused on telling a story that is about premium.”

Similarly, Yahoo’s data-driven targeting platform, Genome, has billed itself as a tool that “allows you to benefit from direct access to Yahoo premium inventory, publisher partners such as MSN and AOL, and comScore top 1,000 through a simple, streamlined transaction point.” Yet Genome campaigns can also include unnamed sites or properties like the file-sharing site MediaFire and the ad networks eHealthcareSolutions and Blackboxmedia. Meaning Genome buyers may be flying blind, while the platform is vulnerable to abuse.

“Yahoo takes supply quality very seriously,” said the company in a statement. “We are committed to maintaining a healthy marketplace with our advertiser and publisher partners, and use technical and procedural safeguards to support that commitment. ... Genome uses a combination of internal tools and third-party services to help maintain the quality of the network. Our Genome network purchases inventory on specific sites based on quality and audience, and regularly provides advertisers with site lists before and after a campaign runs for increased transparency.”

“In the spectrum of networks out there from reputable to sketchy, I’d put AOL and Yahoo on the reputable side,” says Chris Paul, gm, svp at VivaKi. “The most common issue is when new sites join the networks without being fully vetted for advertisers’ content standards by the network administrators.”

Agency trading desks also deliver lots of inventory on networks into which buyers don’t have much insight. Often buyers will receive traffic reports listing buckets of inventory from something labeled “microsoftadvertisingexchange,” say insiders.

While the display market has seen dicey practices growing for a while now, the challenge of bad inventory is suddenly escalating in video, where CPMs can be 10 times greater than display. Take the BrightRoll Exchange. Besides housing loads of inventory from the aformentioned Freestreams, the company also delivers large volumes of inventory via sites like Fave.tv and Videoswag.tv. Plus, it delivers lots of ads via "opaque sources" or sites it doesn't report on, as well as a good amount of blacklisted inventory, per buyers. One buyer says his company blocks one in four impressions sold in the exchange.

BrightRoll CEO Tod Sacerdoti says that in the case of Fave.tv, that inventory wasn’t supposed to be available. The site was on a list of unapproved URLs from an Israeli company called HIRO. Sacerdoti notes that BrightRoll does provide an option for buyers called “tier 4,” adding, “We recommend people question it.”

But overall, Sacerdoti says, BrightRoll is an open platform that plugs into nearly everybody selling video, and is not something that the company can be expected to police. It’s not “BrightRoll’s inventory, after all. The BrightRoll exchange is the inventory in the entire industry,” he says. “Is there an inventory problem in the industry? Yes.”

Indeed, the ad exchange space is so fraught with danger that companies like the independent trading desk Digilant run massive reports every week tracking which companies are peddling the same ad inventory on different exchanges—with completely different labels. According to Digilant COO Nate Woodman, the situation is so ungovernable that the agency has found instances where it’s ended up buying impressions from itself. Digilant blocks one vendor, CPX Interactive, for this reason. Woodman says that companies can blacklist sites all they want, but that they are better off creating whitelists—i.e., lists of preapproved sites. “The problem there is, that will kill your performance,” he says. Why? Because when you’re just out for clicks, bot sites perform better.

And the bot guys are slick. “As soon as sites get on blacklists, there is little incentive to maintain them,” explains Kiril Tsemekhman, svp, chief data officer at Integral Ad Science. “Then a new site pops up.”

And so, the problems persist. John Snyder, CEO of the keyword-targeting firm Grapeshot, says he’s lost business because his company won’t sell bad inventory. “We’ll hear, ‘Your competitor got great clicks,’ but all on two sites and it was all fraud. But it’s these optimization algorithms that find those clicks.”

Says Woodman: “When we try to tighten things up, our measured performance goes down. There is an incentive among buyers to let the floodgates open. And publishers need more money, so they ignore.” So the bad traffic persists. “We need to fix this as an industry,” he adds. “Somebody needs to give a shit.”

The IAB seems to. Per Sullivan, the organization is working on devising a standard for publishers akin to the Good Housekeeping Seal. He’d like to see the biggest stakeholders get more aggressive about the problem, including brands and agencies. “If buyers came out and said, ‘I will only buy from certified vendors,’ that would change things,” he says.

Carat’s Buescher thinks bold steps are needed and urges more of her brethren to take big steps. “It’s a sad thing,” she says. “There is no single provider out there that can fix this. Until then, all brands should have a manager that handles brand safety.”

Chuck Reynolds
Contributor