ADOTAS – Entering DoubleVerify’s midtown NYC offices, a giant poster listing the verification companies various offerings finally blares at the bottom, “DoubleVerify: Friend to the Industry.”
Plenty of publishers and platform companies are likely to disagree with that statement. Among the verification companies, DoubleVerify receives the lion’s share of bile, with network executives complaining about the company employing strong-arm tactics and overselling the capabilities of its technology. However, the proliferation of verification tools across the demand side and into the platforms suggest plenty of media buyers out there consider verification a buddy.
“We’re seeing verification on every impression,” says CEO and founder Oren Netzer. “We see the opportunity two or three years down the road to be the third-party for campaign measurement and accountability in the industry — a Nielsen for online, if you will.”
The acquisitions of two ad verification companies — AdXpose and Adometry — by companies with other core competencies has encouraged whether verification can stand alone as a service. However, DV, as well as competitor AdSafe Media, see verification as moving well beyond its roots in brand protection.
Arguably the biggest player in the field with a reported 40 billion verified impressions a month, DV is adding capabilities over the next few years to make the company’s solutions the standard and centralized tools for ensuring online accountability and measuring campaign performance.
To help make that vision a reality, DV recently scored $33 million in a Series C round led by later-stage investors JMI Equity and Institutional Venture Partners. Just like last year’s $10 million Series B round, DoubleVerify wasn’t actively looking for cash, but intense investor interest in the company signaled to Netzer and crew it was time to get grander plans rolling through a funding injection.
Netzer’s vision for verification’s future is about helping clients produce better ROI through lowering IO noncompliance rates — this equals less wasted impressions and more spend funneled toward target audiences, he argues. (DV also provides audience verification services.) Of DV’s 200 Fortune 500 advertising clients that use verification across the board, IO noncompliance rates with existing customers is down to a few percent.
“We’ve seen campaigns where noncompliance has fallen from 40% to 2%-3%,” he says. “Imagine 37% of the budget of the campaign that was wasted before heading back to the advertiser. Measure that against the cost of using verification, and there’s your ROI.”
In addition to measuring noncompliance, DV also tracks clicks and conversions — and all its data and metrics shared by clients suggests verification improves all three metrics.
On the other hand, the initial percentage of IO noncompliance for new clients keeps increasing, Netzer says — 31% currently vs. 20% a year before. “Because more and more advertisers are verifying, the ones who aren’t are stuck with the bad inventory no one wants,” he says.
Easiest definition of noncompliance: ads delivered to the wrong destinations. This covers brand protection — e.g., keeping family company advertising off of porn sites or social forums devoted to hate groups — which was the initial selling point for many verification services. In addition, DV still has a big business in battling fraudulent impressions with its Fraud Detection Lab. Fraudulent impressions include tricks like invisible ads (iFrames sized 0 X 0), stacked ads (iFrames again!), ad clutter (hundreds of ads on a page), etc.
DV does play the old game of white list/black list sites, but these lists are provided by advertisers and DV verifies according to company orders. A great deal of verification services are based on the 80 categories and four tiers DV uses to classify the 1 million sites it crawls daily.
Premium: Highly desirable content that is safe for all brands. Premium content, such as technology or lifestyle, may appeal to preferred demographics or command a high CPM.
Acceptable: Content that may be appropriate but may not appeal to conservative brands. This content may include matchmaking, user-generated or political material.
Restricted: High-severity content that most brands would consider inappropriate (some “softcore” pornography falls under this category).
Unacceptable: Extremely problematic or negative content such as pornography, fraud and illegal downloads.
Of course, DV has a process for dealing with complaints that a site has been unfairly squeezed into one of these tiers — there all always “false positives,” but the amount is minimal, Netzer claims.
“Many times they’re wrong, but sometimes they’re right and we fix the classification,” he says.
One of the biggest issues is mixed-content/media sites. Netzer cites BoxingScene.com — it may seem like a site about boxing, a great deal of the site’s traffic is driven by the user-generated content section — which happens to be full of uploaded hardcore porn. (Just to be clear, two men beating the hell out of each other is not considered porn to most people.) Netzer guesses BoxingScene doesn’t care as the UCG segment drives so much traffic, but advertisers certainly do.
“We crawl all the pages of a site, and if we find enough pages with pornography on them, we’ll classify the site as porn,” he says. “You can’t blame the ad network for not knowing — it’s our job to find out.”
On the Platforms
See, DoubleVerify doesn’t believe verification is just an advertiser thing — the platform business is one of largest growing segments. Numerous DSPs, exchanges and networks are integrating DV’s real-time API and using Navigator, its sell-side suite of solutions.
It’s easy for ad networks and exchanges to get burned by unsavory publishers, which is why DV’s site classification system can be used for prospecting publishers as well as network hygiene. DV’s Illuminator reports give data on classification as well as the number of ad units on the page (and those above and below the fold). Networks can cross-reference DV’s database to get the lowdown on publishers that want to join and check to make sure the current roster passes muster.
In addition, the tag monitoring service is pretty literal — it allows networks to see all the sites where their ad tags are running, including the ones the network may not be aware of. For example, one network client boasted a list of 8,000 publishers, but when DV turned on tag monitoring, network tags appeared on more than 100,000 sites.
Tag-jacking — typically when a site owner slaps the ad network tag for an approved site onto other owned sites (think porn) — is rampant and easy to conceal with iFrames. So networks unintentionally serve ads on unintended (and perhaps inappropriate) sites and potentially lose advertisers because they’ve been bamboozled.
But the most exciting part is that verification is becoming preemptive. DV’s Brandshield Connect is increasingly being integrated into exchanges, DSPs and ad network ad servers to verify impressions before they are served.
“This way you can better manage your yield as an ad network by asking us in real-time how to deliver the impression,” says Netzer. “Marketers can bid on impressions on exchanges knowing they’ve past the DV criteria already in terms of their quality and brand safety.”
Dancing by Itself
Among the numerous charges levied against ad verification and its providers, the most damning may be that it’s another adjunct technology layer taking up space in the grotesquely cluttered display ecosystem. Worse, there are endless complaints about verification providers being middleman who simply bump up prices (dramatically according to some accounts).
But more important, the questions persists: why isn’t verification a technology that the media-buying platform — be it network, DSP or exchange — provide?
“Verification existed before we came along; it was always first party,” Netzer answers. “All the networks claimed they did some kind of verification internally until we came along. When we came along we found so many problems that the agencies saw first-party wasn’t working.”
Netzer cites two reasons for why a verification service should be a third party. First, ad networks are a principal in the transaction, making money off the inventory they’re signing off on as quality.
“It’s a huge conflict of interest because at the end of the day they have to deliver the amount of impressions they promised,” he says. “If they want to make a bigger margin, who’s going to say they aren’t going to use less-quality inventory.”
Second, there huge advantages to having verification as a core competency. Take the Tag Monitoring solution for an example — networks simply don’t have the technology to quickly identify nested iFrames or tag-jacking, he says.
“Some players seem to think verification is simply content classification, but that’s not what it is,” chimes in Jennifer Hyman, vice president of marketing. “It’s the technology behind it, the ability to actually see where the ad is being delivered and making sure its appropriate.”
However, the acquisitions of AdXpose and Adometry by comScore and ClickForensics — two companies whose core competencies are certainly not verification — cast further doubt about whether verification operations could stand alone. comScore aimed to integrate AdXpose into its analytics offerings while Adometry (the merged entity took that name) has its sites on attribution.
DV’s worthiness as a standalone business can be judged by the Series C haul, Netzer says, particularly considering one of the lead investors: JMI, the company that brought DoubleClick private before it was sold to Google.
“JMI knows the ad space better than anyone, and I think they’ve seen a lot of similarities between us and DoubleClick,” Netzer says.
The narrow view of verification as mainly a brand-safety function is going to disappear as its functionality keeps expanding, he suggests.
“Verification is becoming ubiquitous, it’s on every impression,” Netzer comments. “We believe it’s becoming the standard for measurement of campaigns, even from an impression standpoint.
“Verification is here to stay, and it’s thriving and expanding in terms of what it can do.”