I recently was selected to sit on a panel to discuss the topic of digital ad fraud. The discussion underscored for me the fact that non-human digital video traffic is a widespread concern in our industry. Some vendors have quoted that upwards of 50% of video ad impressions are served to non-human traffic (NHT). Based on Videology data, we do not see NHT penetration anywhere near that. However, it is obviously still a major concern for the industry. According to eMarketer, digital video ad revenue totaled $4.15B in 2013. Taking a conservative 10% estimate of malicious NHT penetration puts this at a $400M problem. Advertisers are demanding to “get what they pay for” and rightfully so.
So how do perpetrators of malicious NHT operate? Often times, they do so by creating malware. Malware is very different than a virus. A virus is meant to inflict harm on a computer and/or its owner - sometimes just for thrills, but also to accomplish illegal acts like identity theft. Malware is actually very much like a parasite – it requires a host PC to live and accomplish the goal of the perpetrator. Therefore, a key goal of malware is to avoid detection. Once embedded in the infected computer, it opens ghost browser sessions and watches video ads. It clicks on ads and navigates across the vast expanses of the internet, all while doing its best to avoid detection from its host. Typically, malware targets Internet Explorer (IE) users because IE is known to have certain security flaws and skews towards older demographic groups who are traditionally less tech savvy. In some cases, these fraudsters work to create fake user ‘profiles’ which they then turn over to their botnet. Actions they take can include creating free profiles on various news sites or putting merchandise in e-commerce shopping carts and then drop out just prior to purchase. These actions enable the cookies for the profiles to be filled with valuable behavioral data that returns higher payouts from audience targeted campaigns – for example, auto intenders are very valuable to auto advertisers.
Fraudsters have created a variety of methods to lurk in the shadows and reap huge profits. One key issue that some in our industry seem to miss is that it is not confined to the long tail – even premium publishers are frequented by NHT – this is a key tactic that botnets use to disguise themselves. Given the fact that malware can target the computers of older demographic groups, fraud also impacts audience verification measurement via Comscore vCE and Nielsen OCR.
Videology employs a global blacklist of domains which our analysis shows have a high proportion of potentially NHT. Every day our fraud detection systems reject an average of 1 billion ad requests from sites that do not meet our stringent quality requirements. However, we cannot solve this issue alone. This is why we’re members of the IAB’s Traffic of Good Intent Task Force, its Quality Assurance Guidelines Initiative, and also employ it’s Bots and Spiders “Dual Pass” methodology. We also work with leading third party measurement vendors including White Ops, Double Verify and Integral Ad Science. At Videology, a key tenant of our vision is to allow advertisers to fully utilize all screens on which consumers are viewing content. We are aware that fraud has the potential to be a roadblock in that vision and it drives our commitment to eliminating malicious NHT. It is not an easy obstacle to remove, but with diligence and the commitment of a united industry, we can and are significantly reducing risk for video advertisers.