Anti-fraud heat boils over to China

When ‘the future of fighting fraud’ is discussed in China, the pun-heavy Chinese language sometimes turns the phrase into ‘the financial prospects of ad fraud’.

And prosperous the ad-fraud industry certainly is, no matter how murky. According to AdMaster data, as much as 46 percent of ad impressions on web verticals is bogus but is still paid for by the country’s advertisers, who spent an estimated total of US$27 billion (eMarketer data) on display and video ads last year.

In January, when Marc Pritchard, chief brand officer of Procter & Gamble, delivered his broadside at digital marketing, and called the US media supply chain “murky at best, fraudulent at worst”, China’s publishers, anticipating spillover to their own ‘walled-garden’ market, started to change their attitudes.

“Because of Marc Pritchard, when we have met with media publishers recently we have noticed they are now more pressured to enact change,” says Andy Fan, CEO of RTB Asia. However, he adds that the biggest media platforms, which are laden with layers of processes and beholden to existing guanxi (relationships), have been less expressive than medium-sized, second-tier media owners.

“The smaller media are more open to third-party verification and viewability measurements, and are even willing to be billed accordingly.”

In AdMaster’s latest white paper analysing the ad traffic of 1,000 advertisers in China for the whole year of 2016, invalid traffic accounts for 30.2 percent of the total, and becomes more obvious towards the end of the year.

Vertical platforms, compared to ad networks, web portals and video sites, suffer the most from abnormalities. And within those verticals, content that focuses on women, mothers and babies, and automobiles is the worst affected.

In China, industry associations such as the Digital Marketing Committee of China, the China Advertising Association of Commerce, the China Computer Federation, and the China Mobile Marketing Association (MMA) have been trying to combat ad fraud with the help of standards and guidelines established by the China Media Assessment Council (CMAC). The latter is China’s version of the Media Rating Council, the US entity responsible for auditing online giants Facebook and Google.

Progress has been slow, though MMA China has shown the most improvement, says Martin Zhang, founder and CEO at the digital ad verification company Adbug. “Why is it so hard to produce anti-fraud results in China? Because the system in China is not the same as the United States — a democratic market in which civil society and industry take initiative,” he says. “Some of these Chinese associations are nationally-appointed; they are not spontaneous and are waiting for something to ‘happen’ before they will do something.”

The CMAC was formed in August, as China’s first media-evaluation and certification organisation, and is still in the preliminary stages in terms of broad ad-tracking in order to determine evaluation standards. The MMA, meanwhile, has dipped its toes into defining the three main elements of ad fraud.

Fan advocates that ad fraud should be reinterpreted, and argues that a distinction should be drawn between intentional malicious traffic with fake volumes and placements, and unintentional, invalid traffic. 

“I think the word ‘fraud’ is a bit too exaggerated,” he says. “Ad fraud is an unavoidable waste of ad dollars to a certain extent.” 

Unintentional invalid traffic occurs because of issues with viewability, a metric that aims to only track ad impressions that have actually been seen by users. But an ad is considered ‘viewed’ if at least two seconds of a video ad or half of a display ad (50 percent of pixel count) is seen. 

Fan says his firm’s data finds “unviewed” ad wastage accounts for 51 percent of all ads.

From the perspective of Ker Loon Ang, general manager of digital media for Publicis Media Exchange China, viewability is an important metric not because of fraud issues but because it is an indicator that the webpage may be too long, or the app may be too complex, meaning people aren’t watching it. 

That wastage overlaps with instances of intentional fraud including ‘ghost clicks’ and ‘stealth redirections’. The situation is growing more serious in China with the rise of what MMA China describes as “sophisticated invalid traffic” due to the prevalence of ‘phone farms’ — warehouses filled with low-end smartphones, which can rig rankings and manipulate apps. Fan estimates that between 2 million and 6 million people could be working in ad fraud in China.

“They can mobilise real personal computers and mobile devices for seemingly genuine user-engagement like online registrations, and even participate in product sampling and car test-drives,” states Mandy Hou, head of ecommerce and programmatic at Performics China.

Other examples of intentional fraud such as stealth redirection — also known as URL hiding or masking — are rampant in China, says Adbug’s Zhang, during a demonstration to Campaign Asia-Pacific. When a consumer clicks on an ad in a website link, he shows, the iFrame tag in the website loads a destination URL which belongs to a different ad. There may be a tracking code on, say, Ad Placement A, but only 20 percent of such impressions may be real. Furthermore, some so-called ‘tracking-and-monitoring tools’ even release internet crawlers to repeatedly load a webpage containing the ad, inflating delivery numbers.

Clients and their agencies are frequently not aware of the problem, says Zhang. “That’s because they are not requesting reports from publishers with placement deliverables at the level of specific URLs, not just on the level of the website domain,” he explains. 

One significant aspect of fraud detection is third-party verification, which has met with resistance from Chinese publishers. Other countries have found competent solutions, but these are built on the premise of high-level cooperation across the entire digital-advertising ecosystem, points out Fan, which is not in line with “Chinese characteristics” — being closed to outsiders and thus difficult to verify.

China’s third-party verification space developed too late, says Adbug’s Zhang, comparing it to the US, where ad-tech firms developed together with the rest of the advertising industry. The verifiers that developed in China are relatively weak when squared off with media behemoths, whose sweetheart deals based on relationships sometimes trump those based on technical competency.

Data bank: Rows of low-end smartphones form an ad-fraud engine.

In the future, emergent ad channels such as connected TV and OTT will be the new breeding ground for ad fraud, AdMaster’s white paper posits. There is incomplete data for comprehensive checks and measurements in this space, because these platforms only accept S2S (server-to-server) as the data transfer protocol, allowing fraudulent ads to be masked, since they are commonly generated directly from ad servers.

The speed of change in China will be anything but fast, concedes RTB’s Fan. Media publishers are still not ready to transact on newer metrics, and most media-buying decisions, even programmatic ones, are made and contracted one year in advance. “You will hear a lot of discussions for the rest of 2017, but we won’t see new technical measurements and inventory pricing mechanisms until the first few quarters of next year.”