Google has been accused of charging AdWords advertisers for YouTube clicks even when its own systems have correctly identified the viewer as a bot, according to new research. A group of European researchers behind the study carried out an experiment in which they uploaded videos to YouTube and then bought YouTube adverts targeting those videos. […]
Google has been accused of charging AdWords advertisers for YouTube clicks even when its own systems have correctly identified the viewer as a bot, according to new research.
A group of European researchers behind the study carried out an experiment in which they uploaded videos to YouTube and then bought YouTube adverts targeting those videos. They also created bots, software that runs automated tasks over the internet, to view the videos.
While the bots “viewed” two of the videos 150 times, YouTube’s public view counter only listed 25 of the views after apparently correctly identifying the rest as fake.
The paper – Understanding the detection of fake view fraud in Video Content Portals (PDF) – by a group of European researchers evaluated the performance of the fake view detection systems of five major online video portals.
“YouTube uses a seemingly permissive detection mechanism to discount fake monetised views,” wrote the researchers, who are from four institutions – UC3M, Imdea, NEC Labs Europe and Polito.
The report’s authors used software that runs automated tasks over the internet to view the videos.
It found that YouTube’s detection system “significantly outperforms all the others” but that it may still be susceptible to simple attacks.
“In practice, this means that views identified as fake and discounted from the public view-counter are still monetised,” it found.
According to the research, online advertising was worth $49bn (£32bn) in 2014 in the US alone.
“While YouTube is shown to strive to protect its users and clients, for example by reacting quickly when suspicious behavior is identified, we speculate that its setup seems to place an unnecessary burden of risk on clients,” said the research.
“For example, fake views can be discounted equally for public and monetised counters, but they are not,” it added.
It concluded that the detection mechanisms deployed by video hosting portals was complicated as it currently relies on the companies’ own systems.
“In this context, the development of independent tools able to monitor and audit the fidelity of these systems are missing today, and needed by both industry and regulators.”
A YouTube spokesperson commented: “We’re contacting the researchers to discuss their findings further. We take invalid traffic very seriously and have invested significantly in the technology and team that keep this out of our systems. The vast majority of invalid traffic is filtered from our systems before advertisers are ever charged.”
