Through programmatic, advertisers inadvertently feed the spread of false news. They should take steps to reverse the trend.
“Facebook has a moral duty to make authenticity a priority,” said Jon Snow, a Channel Four reporter. News Channel Channel Four attacked the social network denouncing its reluctance to quash with the fake news during the television festival in Edinburgh.
Citing a fake news bulletin that the pope reportedly declared his support for Donald Trump for the presidential race, this post hired a million users when it was shared on Facebook.
However Facebook is not the only player who must act on these fictitious content. Through programmatic, advertisers inadvertently feed the spread of false news. They should take steps to turn the tide, by stopping the foundation of this false information at source, and thus ensure that their brand reputation is not compromised by deceptive content.
How is the digital ecosystem releasing false information and how can advertisers ensure that they are not associated with inappropriate content?
Facebook and other social networks contribute to the appearance of fake news, because these platforms allow a rapid distribution of the content to a large audience and at lower cost with minimal regulation.
The more controversial the content, the more attention it receives. After the presidential elections, which took place in the United States, Buzzfeed revealed that the fake news on Facebook have considerably surpassed the commitment of the real information.
When social media audiences mass click on links to sites containing corrupt content, the programmatic ads delivered on those sites generate revenue for the publisher which, with the exception of occasional political motivations, is the reason main of the spread of fake news.
Facebook claims that it is already acting, including blocking pages that repeatedly share fake news on the platform, in addition to blocking ads that are linked to fake news and that display logos of editors next to the shared links.
Advertisers also have a role to play in preventing their ads from being associated with illegal content. Indeed, it may mean that the advertiser finances the creation of fake content, but it also gives the impression that he believes and approves the story.
A BrightRoll survey reveals that 96% of advertisers are worried about fake news in the programmatic advertising environment, but this concern is not big enough for them to abandon the programmatic and thus the revenue it brings in terms of reach, efficiency and targeting. When brands choose programmatic advertising, they inevitably sacrifice some control over the ad slot and need to put in place additional measures to prevent these ads from appearing alongside undesirable content.
Emerging solutions to counter fake news
As the problems of advertising hijacking become commonplace, there is an abundance of brand safety experts, all of whom promise to protect brands from bad associations with dubious content.
Unfortunately, the majority of these providers rely on outdated and ineffective techniques such as keyword filtering, which does not guarantee brand safety. For example, a false news about the actor Scott Baio and reporting that he was killed in a plane crash, included keywords such as “airplane”, “golf” and “Mar-A-Lago” this which attracted tourism and travel brands, and was associated with a hoax.
Similarly, techniques such as protection by the domain name (blacklisting) are only effective if the solution is sufficiently comprehensive to eliminate all sites at risk. Some “solutions” of blacklisting are based on the assumption that all fake news exist on poor quality websites such as gossip forums, so blocking these types of sites will keep the brands in a secure environment but that’s not necessarily the case. Sometimes, fake news are relayed in good faith by renowned news sites. For example, the news that the founder of Corona Beer had offered 2 million euros to each resident of his home village was relayed by The Independent and The Metro – among others – before it turned out to be false . In this case, it’s a fairly harmless story, but the situation is a good illustration of the fact that fake news can also be relayed on premium sites.
Machine learning at the service of brand safety
Fortunately, the industry is now seeing the emergence of more advanced tools that use machine learning and semantic targeting to solve this problem.
There is of course a skepticism about the use of algorithms to identify fake news. There are so many different definitions of the term fake news; From content that is entirely manufactured to real news relayed in a biased way for a specific purpose, how can machines distinguish true from fiction?
Machines have a huge advantage over humans to spot fake news. Because of the total volume of content, they can analyze much more than any human can read during his life. Machines can identify patterns and groups of words associated with fake news.
These solutions use machine learning algorithms to interpret the content of the page, determining its true meaning and context. They can identify which text elements are relevant and analyze the relationships between words and sentences using a semantic algorithm.
Semantic classifications are coupled with keyword identification models to ensure greater accuracy and thus allow integration of standards provided by the IAB. This content crawl takes place at a pre-bid level, so that the context of the content is understood before any programmatic bidding is initiated.
The use of these tools by the brands makes it possible to target or avoid any specific category. While brands want to stay up-to-date and buy impressions on hot topics, most also want to avoid highly controversial content, whether fake or authentic – to avoid dismissing too much of their audience. Brands can also create custom categories to avoid sites that can generate fake news.
Tackling the fake news issue therefore requires a joint effort from several groups, including advertisers, governments and platforms such as Facebook. For now, advertisers can counter questionable content by implementing pre-bid machine learning technologies to understand the content of a page. This will enable them to take advantage of the many advantages that programmatic offers without risk of achieving brand brand safety by also avoiding the global epidemic of fake news.