This is the fifth in a series of interviews with vendors combating the problem of ad fraud. Other companies participating in this series include White Ops, Moat, Telemetry, Sizmek, comScore, Dstillery and Asia RTB. Read previous interviews with Forensiq, Integral Ad Science, PubChecker and Videology.
DoubleVerify is not new to the ad fraud game.
When the company first opened its doors in 2009, it was, as its name denotes, primarily focused on the verification process that ensures ads appear where they’re intended to appear in compliance with regulations. DoubleVerify collaborated with the IAB back in 2012 to develop the “Guidelines for the Conduct of Ad Verification,” now considered the industry standard.
Today, DoubleVerify aims to be a full-service solution, said Chief Operating Officer Matt McLaughlin.
“We’re not just attacking one element of the fraud problem,” McLaughlin said. “There are a number of different types of online ad fraud that crop up wherever bad actors can make money in new and creative ways by generating impressions.”
Hidden ad fraud (impressions concealed behind other content or ads displayed in a tiny 1×1 pixel iFrame), laundering fraud (discussed below) and nonhuman bot traffic are par for the course.
Bots in particular are becoming more clever, and therefore insidious, making it harder to tell the humans from the robots.
“We apply hundreds of data points against billions of impressions and combine that with unique external information to identify an individual nonhuman browser,” McLaughlin said. “We take a deterministic approach to bot identification where we look at hundreds of data points per user/browser in a variety of environments so that we’re absolutely certain that the activity that we’re flagging is nonhuman and that we’re catching as many as possible.”
McLaughlin went into more detail with AdExchanger.
MCLAUGHLIN: DoubleVerify authenticates the quality of digital media for the largest advertisers online. Our goal is to bring transparency and accountability to all online advertising through a combination of automation and human insight in order to drive performance.
How do you guys accomplish that?
Our fraud solution is not a standalone. The best example of that is viewability. If I’m an advertiser, but I have two different solutions, one for viewability and a second provider for fraud, then there’s no way to determine the overlap between the two solutions to see if a true human viewed an ad. By integrating a broad range of metrics, we ensure a single view across all dimensions, including viewability, brand safety and the elimination of fraud, all into one metric. That’s valuable because you’re not trying to figure out what percentage of non-viewed ads were actually human and what percentage was fraud.
What is DoubleVerify’s take on the state of video viewability?
Video viewability has only come under the spotlight relatively recently. It was only approved as a metric [by the MRC] that’s reliable enough to be transacted on a short while ago. It’s an important standard, but it’s still in its early days. DoubleVerify is part of an industry consortium [OpenVV] where we adhere to a common open-source technology. It’s not just about the mechanics – it’s about whether an ad has been viewed. There is a high degree of fraud in video and as a result we need to measure it in conjunction with fraud to make sure the ad is truly viewable by humans in a fraud-free environment.
Do you do fraud reporting, prevention or both?
We offer pre-bid and serve-time blocking. This enables transparency into where the fraud is impacting campaigns at the impression level before an individual is actually served. DoubleVerify can block entire sites or simply just the fraudulent impressions so advertisers don’t need to cut off entire inventory sources.
Can you see into iFrames using BrandShield? [DoubleVerify’s BrandShield product claims to prevent and block ads from appearing next to inappropriate content.]
Yes. DoubleVerify has one of the highest ‘see-through’ rates in the industry at over 99%. BrandShield was built to see through the entire ad daisy chain and DoubleVerify’s blocking technology is built off of that foundation, enabling us to see the actual URL where an ad would be served, to determine in real time if the ad fits within a brand’s criteria and if it’s fraud-free and then make a decision on whether or not to serve the ad.
Are you self-serve?
We don’t leave brands and agencies to simply use the tool. We help customers interpret the results and provide actionable recommendations. We believe the most value comes out of experts who understand the space and who are familiar with the challenges brands and agencies face, helping them solve the impression quality issue and optimize their media buying.
Who do you work with most – brands, DSPs or media buyers?
We work across the entire industry. We believe in our mission to build a better industry. We work with both media buyers and sellers to authenticate their impression quality. We also work with the advertisers, we work with their agencies and we’re working more and more with large brands and a range of inventory platforms.
Are you seeing more direct brand relationships as they take their ad buying in-house?
We definitely see that trend occurring. As more marketers are using programmatic and there are more and better tools for advertisers to utilize and manage their own media buys, we do see a growing tendency for advertisers to bring the buying in-house. But they need to be educated about fraud.
Who is your competition?
If you look at other companies that are accredited across a broad range of services as we are, two companies have a similar accreditation footprint: comScore and Integral Ad Science. But I’d say we’re the better choice because of the technical superiority of our solution. We were the originators in the verification space. We founded and created the space and drove the creation of standards around verification. It was based on technology that we developed that provides the highest level of transparency and accountability in a broad range of measurements.
You’ve recently brokered partnerships with Tremor Video and SpotXchange. What do they entail?
We help our media partners by enabling them to focus on building a great platform for buying, selling and serving their customers without having to build expertise in the underlying technology necessary to authenticate the quality of the ad impression that they’re placing. We integrate our fraud avoidance solution and impression-quality management solution into their technology so they can access sophisticated solutions that allow them to have a metric they can use to manage quality through their platforms.
How has the ad fraud space changed in the last year – and what’s coming?
What we’re seeing and what we’ve seen in the last year is a continued evolution. Five years ago fraudsters were doing relatively simple things like faking impressions one-by-one in iFrames that were served but not seen. Today they’ve become more sophisticated, using automated nonhuman bots to generate phony ad impressions and employing sophisticated coding techniques to avoid detection.
What’s a particularly nefarious form of fraud you’ve seen recently?
What we know is that fraudsters are opportunistic and that they’ll migrate to wherever the best opportunity is. One example we see is impression laundering fraud. They take low-quality sites that have large amounts of real human traffic and resell those ad impressions through phony content sites. Piracy sites, for instance. There’s a high degree of interest in those sites because they give free access to streaming content or downloads even though the owner doesn’t have licensing rights.
Advertisers try to avoid those sites, but in the case of laundered ad impressions, the impressions are routed through seemingly legitimate content sites, such as travel or food. We uncovered an elaborate ring last year that was doing millions of dollars a month in ad fraud. A similar scheme is also being used in online video advertising. It’s more sophisticated than we’ve seen before. Fraudsters are double-routing ads through multiple laundering sites to further hide the origination source.
What about bots?
We’re seeing more and more in terms of emulating human activity. Nonhuman bots are getting more sophisticated at mimicking the behavior of someone after they click an ad. After clicking an ad the bot might then go to the advertiser’s site and do things like watch a video, build out a spec car, fill out a lead form. A lot of times they’ll even attempt to complete a purchase and send a conversion code back to the advertiser. Depending on the tool the advertiser has, they might clock that as a legitimate conversion and give credit to the bot site and do an optimized media buy for that site.
And what about programmatic?
Ad fraud is clearly prevalent throughout the ecosystem. There are two things that indicate higher degrees of fraud. The first thing is when you see an ad impression sold through an indirect sales channel like an ad network or a programmatic exchange. You’re up to three times more likely to have fraud there than in direct publisher buys. When long-tail inventory is aggregated from thousands or more individual suppliers, advertisers don’t have direct relationships, and it’s hard to hold the seller accountable. Multiple outlets sell and remain undetected.
There is also a transparency gap, where the URL that the DSP bids on differs from the actual URL where the ad ultimately appears. This happens on about a quarter of exchange impressions. Many customers are using our transparency solution to identify the gap which exists primarily in the exchange environment.
Can advertisers and agencies trust the inventory they’re buying today?
There is a Russian proverb that Ronald Reagan made famous: “Trust, but verify.“ Trust comes from transparency and accountability and advertisers and inventory providers continue to utilize independent third-party measurement tools like ours to gain that trust and make it grow. That’s what we mean by “building a better industry.“ This is about building transparency under which companies can trust one another, but also not live in a blinded environment.
