Are Direct Buys The Only Way To Keep The Industry Fraud Free?

NYTimesDCNThe New York Times has a solution to the bot fraud problem: “Buy quality.”

So declared Michael Zimbalist, SVP of advertising products and R&D at the Times, speaking at OpenX’s session about trust in the supply chain at Advertising Week in New York City.

“You get what you pay for,” Zimbalist said bluntly. “We need to collapse the intermediary space and do more transactions between known buyers and sellers. How much more transparent can you be than saying you’re buying from The New York Times and it’s a direct transaction?”

It’s a loaded question. In April 2014, Zimbalist told AdExchanger that the Times was “very cautiously” going to start “opening up much more of our inventory to the real-time marketplaces” while watching “very closely what sort of impact that may or may not have on our direct business.”

The Times recently participated in a research initiative conducted by White Ops and publisher trade org Digital Content Next (DCN) examining the impact of bot fraud on premium ad traffic. The commissioned study looked at ad impressions across 32 different DCN member sites. [Download the study here.]

If that sounds familiar, it’s because the Association of National Advertisers ran its own study with White Ops roughly a year ago in an effort to determine the economic impact of bot fraud on the bottom line. According to that research, around $6.3 billion in ad spend is likely to go down the tubes in 2015.

DCN purposely mimicked the ANA’s methodology, but whereas the ANA looked at data across the entire advertising ecosystem, DCN focused on well-known publishers like the Times.

The DCN results, released Monday, noted that bot traffic accounted for around 3% of premium publisher traffic on average, far lower than the 11% cited by the ANA in its study.

But what does “premium” actually mean? It can be a slippery term, but in DCN CEO Jason Kint’s view, the notion of what makes something premium is “based on the trust of consumers and advertisers.”

“All of our members have direct, trusted relationships with both consumers and advertisers, so it’s core to their businesses,” Kint told AdExchanger. “Premium gets thrown around way too often. I know trust [and] there is an obvious connection here to ad blocking, too.”

One would be remiss in not pointing out, however, that DCN is an interested party, as it represents the premium publisher community. In addition to The New York Times, DCN members include The Onion, The Washington Post, AP, Financial Times, Forbes, Gannett, Meredith, Atlantic Media, Business Insider, News Corp. and National Geographic.

It’s been nearly a year since the ANA study was completed and there has been some action taken, namely the the formation of the Trustworthy Accountability Group, a fraud-fighting industrywide coalition comprised of the ANA, the American Association of Advertising Agencies and the Interactive Advertising Bureau.

But it’s also been a year of, as Zimbalist put it, “talk, talk, talk” – and any progress the industry has made in the fight against fraud is unlikely to have a material impact on how much money will be funneled into the pockets of criminals next year.

According to Bob Liodice, president and CEO of the ANA, it’s likely that bot fraud will take an even larger bite out of ad budgets in 2016.

There’s a practical reason for that, Liodice said, namely the fact that the ANA publicly announced when it would be conducting its study, thereby unintentionally alerting the bad actors to cut down on the fraud for the length of the study – which they did – meaning that the $6.3 billion number is probably a little low.

The ANA is in the process of running a second, this time more stealthy, study with White Ops, the results of which Liodice said should be ready by the end of the year. Liodice also said that he expects that oft-cited $6.3 billion number to “trend down” in the longer term.

“It’s striking that almost a year after the original study that the ANA did, there is not a smidgen less [fraud] and maybe there’s even more coming up,” Zimbalist said. “That should be a real call to arms for all of us. What’s taking so long for people to wake up to this?”

Part of the issue is that it’s really hard to stay on top of fraud. It’s not a “one and done,” said Jason Fairchild, CRO and co-founder of OpenX, which he claims rejects between 35% and 45% of traffic outright as part of its internal quality controls.

“Bad actors are everywhere,” Fairchild said. “It’s an arms race and it requires constant investment to stay ahead of them.”

If publishers want to keep the lights on, they’ve got to operate well-lit environments – and that takes constant vigilance and an acute awareness of where they’re sourcing their traffic from.

The Gray Lady isn’t inherently premium – that takes work, Zimbalist said.

“You can’t rest on your laurels,” he said. “Trust is earned every single day, one transaction at a time. Premium brands are premium brands because they do all this – they monitor. [It’s] an ongoing thing.”

But The New York Times has the resources and the wherewithal to stay continually on the lookout for fraud. What about small and mid-size publishers?

They’re in a bit of a pickle.

Small publishers generally can’t justify their own dedicated sales team, which leads them to outsource that function to agents who can help sell their inventory for them in return for a piece of the action, Kint said. These days, that means working with one – and in most cases, way more than one – of the many ad vendor companies that show up on your average Ghostery scan.

Technology “provides scale, which greatly helps small publishers,” but the problem arises “when those intermediaries are able to go beyond the inventory itself and pool the cookies/data across sites, which leads to the many issues in the ecosystem,” Kint said.

“Buying direct isn’t the requirement to solving this,” he said. “It’s just the low-hanging fruit to eliminate the shenanigans.”

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. It doesn’t eliminate it at all. It just provides a nice false sense of safety by hiding behind a publisher’s historic reputation.

    There are still bots that go to the NY Times and get served fraudulent impressions. The NY Times may choose to decide to simply not charge for these impressions, but that will often be whatever is determined fraud by whatever 3rd party measurement partner they choose. Due to very large measurement and methodology differences, simply because one vendor doesn’t declare an impression as fraud, it doesn’t mean that it is.

    Even then, premium publishers do plenty of shenanigans that are rather fraudulent. One major US news site (not the NY Times) gets nearly a quarter of its revenue from mugshot photo galleries. This is done by having ad impressions reload with every new picture. Users will only see these impressions if they decide to scroll a quarter-mile down the page with every new picture in the slideshow (below even the footer). This publisher knows that these ads are never seen or clicked, but that doesn’t stop them from selling it for $15+ CPMs.

    • I don’t think it’s that straightforward. Most of bot traffic exists because someone is paying to drive traffic to a certain site in order to increase the opportunity for that site to generate revenue. Sometimes, a publisher may buy traffic with the understanding it’s “real” even when it’s not. Premium publishers will be much less likely to buy traffic, since they can grow traffic organically around their content and brand name. The opposite goes for a less premium site, who would be more likely to buy traffic to increase their pageviews. I’m not claiming this to be the only reason bots exist, but a publisher’s “historic reputation” may be what allows them to avoid fake traffic.

      • Bots also regularly visit premium publishers, even those that don’t pay for any traffic whatsoever. It’s an easy way of making their cookie profiles look more like real people and avoid detection. It’s often their historic reputation that makes them a prime target for bots some. Many of the more savvy bots even spend the majority of time going to highly reputable sites rather than driving paid traffic.

        Unless these premium publishers, which are more victims of fraud than perpetrators, block these unwanted bots from loading webpages and impressions entirely, they are still going to be selling or delivering fraudulent impressions to their direct clients.

  2. Jason Kint


    Actually it very much does. Read the report. I’m very proud that our members’ brands and their ways of doing business lower their rates by up to 90%. And I counsel them that they’ve earned this trust through their reputations which are based on their track record and renewing that trust each and every day into the future. Much more work to be done.


    • I’m not sure you should be promoting that paper much as “proof” about your members levels of ad fraud as there is no discussion about methodology in the paper, and also there is no mention about caveats. It is simply marketing, and not research.

  3. the similarities between managing a security network and a publisher network are strikingly similar and both are a game of whack a mole… The answer lies in new ways to do programmatic – based on context & users initiating the action. Examples of approaches that mitigate the bot issue is programmatic against very niche topics. You can scale but frustrate bots. There are many others – we just have to tilt the problem a bit to see new solutions.

  4. Under careful investigation, there is very little correlation between direct buys and reducing exposure to ad fraud.