Long a dirty little secret of the digital media business, the topic of ad fraud has been thrust front and center in discussions among agency executives, advertisers and publishers over the last three years. Bot traffic, or nonhuman digital traffic, is at its highest ever, and recent projections from the Association of National Advertisers have more than $7 billion in advertising investment wasted. A variety of organizations and companies are putting serious efforts into stanching the losses, and the ANA will make the topic a centerpiece of its annual Masters of Media Conference next month in Hollywood, Fla. Serving on a panel discussion on March 3 will be Amy Bartle, director of media and digital marketing, La Quinta Inns & Suites; Ron Amram, vp of media, Heineken USA; and Michael Tiffany, co-founder and CEO of cyber security firm White Ops, who will present findings of a study on ad fraud. The following is an edited Q&A with the panelists, along with insights from Bill Duggan, group evp of the ANA.
Adweek: What part of the digital ecosystem suffers the worst fraud issues, and why?
Amram: This has evolved over time, but right now the most concerning fraud is in the digital video space, specifically regarding high-end content. It's a medium with the highest CPM, or cost, and the greatest expected impact. Everyone assumed that fraud was more prevalent in the dark underbelly of digital and not in higher-end video content.
Bartle: Whichever one provides the highest return and lowest scrutiny. That may sound trite, but the reality is that ad fraud is driven by extremely smart individuals who are quick to see opportunity in an unguarded space.
Tiffany: Fraud is not pushed into the digital ecosystem—it's pulled in, when demand exceeds supply. Fraud is worst wherever marketers grow their spending faster than the real supply. That's what links fraud hot spots as diverse as digital video—where dollars shifted to video much faster than the rate at which consumers increased their viewing of ad-supported video—and demographic targeting, where too many dollars are chasing audiences that are just not that reachable.
Duggan: The advertisers fund the digital ecosystem and get hurt the most. It's the advertisers' money that is being wasted by fraud.
The mobile space in particular seems to be rife with the potential for fraud. Does it give you pause in working on that platform?
Tiffany: Right now, the wildest fraud innovations are happening in mobile. There are a host of new fraud models from apps that purport to do one thing but really monetize via ad fraud. And the old models, especially the widespread practice of buying huge volumes of visitors in order to serve more ads to them, really matured in the desktop space, but are now available for mobile, too.
Bartle: The mobile information provided in this year's ANA/White Ops study certainly raises concerns. The mobile space overall has existing tracking and verification challenges, and adding the specter of bot fraud into the already murky space is daunting. Careful planning and inventory selection, along with third-party verification, will help advertisers move forward with cautious confidence.
Amram: Yes and no. Unfortunately, we are at that place now where you have to always pause and think about the implications of fraud, and how it may impact what you're doing in all digital media. That said, we continue to move forward because mobile is so important, and it has quickly proven to be highly effective with consumers today. Mobile works, so you can't avoid it or shy away from it.
The ANA expects more than $7 billion to be lost to ad fraud this year. How is your business specifically affected?
Amram: It makes digital media less effective, and it has held us back from making broader changes to our media mix. The ROI in digital is impacted by the proliferation of fraud. We've been lucky in that we've been able to contain it with the right tracking—but it is a concern.
Bartle: Fortunately, due to proactive policies and early adoption of anti-fraud policies and verification, we see our biggest impact, while low in cost, in the form of the monitoring expenses. This is essentially an insurance policy to keep our media dollars focused on human exposure and minimizes our impact from bot fraud. Our exposure is low due to our partner selection processes, using a third-party verification company and our mandatory contract term to pay for human traffic only. We are also moving forward to require disclosure of sourced traffic by our inventory sources.
Duggan: For the 49 advertisers who participated in the ANA/White Ops initiative, the average annual loss due to fraud was $10 million per participant. Every advertiser is affected as the ROI of their digital advertising is pounded by such losses.
What are the most effective tools in place to combat fraud? Are they working?
Bartle: Third-party verification, preferably one that is more fact-based than probabilistic. And consistent application of anti-fraud policies and best practices to your media buy.
Amram: I think it's two stages. First you have to do the blocking and tackling, which is tracking. Tracking your buys gets you to blacklisting and pre-bit screening. In order to really take this to the next level and allow us to reduce fraud as an issue, we need to work with vendors, isolate the issues, remove it from the ecosystem and work both sides of the fence. But all of this starts with transparency—the more you can work with vendors to remove [fraud], the better chance we have at cleaning the ecosystem at the root.
Tiffany: Vigilance and leverage. Any website, network or platform, large or small, could start buying bot visitors tomorrow under the rubric of a pay-per-click campaign, and then go from being very clean to being very botty, with no apparent changes to their webpages. How can you protect yourself against that? We found the buyers with the lowest fraud exposure were predominantly using a combination of independent monitoring technology and business processes that strongly incentivized their inventory sources to fill their orders with organic traffic only.
Duggan: Clearly there are technical tools. But I believe that the most effective tool is a commitment by advertisers to personally combat fraud. That's not just a role for agencies and publishers. We asked study participants the question, "Who should be responsible for combatting fraud?" Seventeen percent answered that the advertiser is responsible. Meanwhile, the agency (36 percent), the publisher (21 percent) and all parties (26 percent) had higher levels of response. But those study participants who feel that the advertiser is responsible all had lower levels of fraud. Coincidence? I think not. Advertisers, it's your money and you should steward your investments.
What still needs to be done to get the problem under control?
Tiffany: This is a crime that works, by definition, by evading detection. When the smartest bad guys figure out how to fool you, they don't tell you you're beaten. What you see instead looks like victory: fraud numbers going down! Then you're only beating the dumb crooks with no sense of what you're missing. This is a game where losing can actually look like winning. So the top action item is to reject complacency. If the fraud operators making big money today were defeatable with data science, scale and performance optimization—offerings of approximately everyone in all of ad tech—this problem would have been solved years ago. There used to be widespread denial: a belief that fraud didn't happen, or if it did, it got optimized away. Now denial has yielded to acceptance: that fraud happens … to other people. The most remarkable feature of the fraud landscape is that, when a buyer or platform gets aggressive about eliminating fraud, the fraud operators don't give up; they find new victims.
Bartle: The MRC (Media Rating Council) needs to move forward and provide accreditation to fraud detection providers. Many inventory sources have cited the lack of MRC accreditation as a reason why they are not using fraud detection services. This is an excuse that needs to be eliminated as soon as reasonably possible.
Amram: We need to have a more industry-wide approach to the problem. More advertisers need to hold vendors accountable, and we need to have publishers acknowledge and accept more responsibility. There are many publishers who take this seriously, and because at points they are on an island, they get punished. Vendors who intentionally or unintentionally benefit from the volume generated by fraud negatively affect the vendors who are more responsible. So the advertisers need to do their homework.
Duggan: There are some simple recommendations in the ANA/White Ops report. Advertisers should (a) place legal language in contracts that state the commitment not to pay for fraudulent impressions; (b) select media partners who proactively reduce fraud, and (c) create open dialogues with providers about traffic sourcing and carefully select the providers with a commitment to providing valid impressions.
Further, the industry needs to increase its support of the Trustworthy Accountability Group. TAG is an industry program designed to eradicate digital advertising fraud, malware, ad-supported piracy and other deficiencies in the digital communications supply chain. TAG has developed an Anti-Fraud Working Group with a mission to improve trust, transparency and accountability by developing tools, standards and technologies that enable the elimination of fraud.
Automation and programmatic have given rise to much greater opportunity in digital advertising, but they also appear to have enabled even more fraud to occur. How should your company and the industry as a whole balance moving forward in programmatic?
Bartle: Most importantly, clients need to abandon low CPMs and CPCs as the driver of their inventory selection decisions and replace them with a focus on actual business KPIs, be it sales conversion, email sign up, loyalty program enrollment or a measured brand health metric impact. Using a ROAS [return on ad spend] metric rather than an expense metric is the better way to ensure value in your media selections. Micro conversion "successes" such as landing page visits driven by the lowest possible CPC or CPM are rife with fraud exposure unless they are backed up by a subsequent KPI impact.
Duggan: The foundation of optimizing one's programmatic media investment, including reducing bot fraud when using programmatic buys, is understanding the programmatic supply chain. Advertisers should ask about the role of each player in the process, know the partners of their partners, and then ask for inventory transparency to know where their programmatic advertising is running. You wouldn't "blindly" run your advertising in offline media such as television or print without knowing the specific networks or publications that carry your advertising. Why accept anything less in programmatic buying?
Does the media agency or the media outlet have more responsibility to fix the problem? And are you confident in their actions?
Bartle: I put the primary responsibility on demand-side platforms [DSPs] and publishers to fix this problem. In the case of DSPs, the lack of transparency into their inventory and traffic sources leaves an open door for fraud to enter and remain in the ecosystem. Anecdotally, the clients I know that use private exchanges find they work best when there is complete pricing, inventory and traffic source transparency among all parties.
Amram: Much of what causes fraud sits underneath the levers [that] sites use to build traffic. So in that sense, media outlets need to be more responsible. But agencies and advertisers need to shift their thinking to people and not impressions, to make sure that we're all talking the same language. You can't blame sellers if the buyers are not scrutinizing what they are buying and are not being clear about what they are paying for. But now we have brought the conversation forward, and we've reset expectations, so I think we need to see a more serious reaction from the sellers. At the end of the day, media properties will get away with whatever we let them get away with.
How does your company address the problem internally, in gauging the size of the problem and how you act on it?
Bartle: We are driven to maximize each media dollar we spend. To that goal, we chose to move on this issue before it became a problem for us. This is definitely an area where moving quickly provides the greatest return.
Amram: We work closely with White Ops and Integral Ad Science to track and blacklist. Our media agency, MediaVest, as well as our DSP (TubeMogul) work closely with both to screen our buys and our inventory.
Looking ahead five years, will this problem get better or worse, and why?
Amram: I think we all would've assumed that we would've seen some improvement by now. It's clear that it's evolving and changing. If we all take the right actions, I can see it improving. But I can see it getting worse as well. So right now I would say it's unclear.
Bartle: As an inherent optimist, I think this problem will diminish in the future provided a few trends continue. First and foremost, all clients need to be vigilant and stand strong in their demands for quality human traffic and not exempt their agency-run platforms from the same scrutiny they would give an unknown partner. Second, I'm confident the movement to have sourced traffic disclosures become standard will have had some success. Finally, it would be a great achievement if ad-fraud monitoring becomes as pervasive as the AdChoices movement did—and as quickly!
Duggan: It will get better. With greater awareness of the issue, more advertisers will become involved in the fight against fraud. They will take actions identified in our report to minimize fraud. And as advertisers demand greater accountability for their media investments, agencies and publishers will increase their role in this fight, too.
This story first appeared in the Feb. 22 issue of Adweek magazine. Click here to subscribe.