BlogThe latest for Creators, Brands, and Fans
Brand safety challenges have existed since the beginning of advertising. The issues can range from poor placements of cruise ads in front of news about a norovirus outbreak on a cruise ship, to much more serious misalignment with terrorism, hate speech and inappropriate child-centric content. Brand safety has always been a top priority at Fullscreen and we understand that YouTube and other user-generated social environments pose unique challenges for advertisers. As a result, we’ve taken a “human plus machine” approach to ad placement designed to complement other leading tools and filters.
Brand safety concerns are real and have totally changed the dynamic of the industry in the last several years. The topic of brand safety was thrust into the limelight in March 2017 when some advertisers on YouTube discovered their video pre-roll and overlay ads were being placed on or near ISIS and terrorism videos. This sparked an immediate massive boycott of more than 250 major brands on YouTube. As the second largest search engine, this was a pretty big deal. Although YouTube immediately made some platform updates that brought some brands back, there are still a few major players that have yet to return. It became immediately clear that a major change was needed in the digital advertising space around transparency, technology, and guidelines — three core challenges to brand safety. Enter Fullscreen Shield — our proprietary solution to the lack of technology for brand safety.
Most brand safety controls, developed by Google or 3rd parties, up until now have been a “black box” — meaning the solutions in place provide little to no transparency into how they work and only skim the surface of the brand safety issue at scale. This leads to a guessing game where complete brand safety can’t be guaranteed, leaving brands exposed and vulnerable to poor placements “falling through the cracks”. Whether it be filters in place on the ad buying platforms or 3rd party verification, it’s tough to get detailed answers about exactly how these filters or tools work and if they are actually effective until it’s too late.
The biggest gap we see in existing brand safety technology is that no tools are looking “inside” the videos to see what’s in the image and audio. This is a crucial missing element needed to decipher the context necessary to determine whether a video will be safe or not for a brand to align with. Fullscreen Shield solves this dilemma by reviewing video images and audio and converting them into text which is then mapped to a corpus of known unsafe keywords.
Shield combines an analysis of the image and audio with metadata, comments, titles, descriptions and audience ratings into a weighted algorithm. This analysis both inside and outside of the video provides the most comprehensive snapshot to determine how brand safe a video actually is. The result of the process are ratings; modeled after the guidelines used for movie and TV ratings. This system flags videos that could be problematic for the brand or campaign by giving that content a mature rating. But we didn’t stop there. As some contextual elements such as slang, comedy, and nuance, can be difficult for a machine to decipher and therefore receive a low confidence score through Shield, we added in the human element. Videos that receive a low confidence rating are flagged for human review prior to launching an ad campaign. We believe this combination of machine combined with the human review is what yields the highest amount of brand safety.
See an overview of the flow here:
Additionally, Shield enables us to have the necessary conversation about subjectivity with brands. What might be safe for a brand like Red Bull is likely not safe for Disney. Plus, content that seems like Disney content may actually not be Disney content at all. Later in 2017 “Elsagate” further illustrated that analyzing video context can not be a purely automated approach. Conversations around subjectivity need to be a priority for all brands and media buyers. Without conversations around specific brand safety guidelines it’s nearly impossible for an advertiser to prevent an unsafe brand alignment.
As we’re in the business of driving results for brands, we have a pretty good handle on what works. YouTube is expected to net $4.35 billion in US video advertising in 2020, and it remains a key component to the marketing efforts of major brands. However at the core, we know that brands need to feel confident about advertising on YouTube. Shield’s sophistication and scalable transparency, provides the ability to guide and protect brands from the vulnerability of brand safety issues. It invokes confidence in brands so they can ensure that their message is being absorbed in the appropriate environment and drives the results they expect to see.
Shield has been used in beta by a major telecom partner and is now widely available to all Fullscreen clients. To learn more about Shield and how you can apply this tool to your advertising campaign, contact us at firstname.lastname@example.org.