Brand safety to suitability: driving video advertising forward in 2021

The rising demand for digital video is now a firmly established pandemic-driven trend. Various reports show viewing has increased for an overwhelming majority of consumers (96%) since the outbreak began, while leading social giants such as YouTube and Facebook saw substantially expanded UK market reach in 2020. Across this year, connected TV audience estimates alone have been revised up to 40.9 million. So, it’s therefore no surprise that supply is on the rise.

But alongside the opportunities created by higher volumes of video content for advertising and monetisation, there also comes challenges. Advertisers, media owners and their data partners are under rising pressure to process more inventory, but they’re also expected to do so without compromising accuracy, relevance, or brand reputation. Using traditional safeguards and text-based analysis, that’s proving far from simple.

With online video set to fuel two-thirds of overall ad spend alongside social media in 2021, maximising benefits on all sides will require better contextual understanding.
 

The safety to suitability shift must go further 

COVID-19 may have highlighted the problems with conventional brand safety guarantees, but issues aren’t new: in 2019, studies revealed directing ad budgets away from generic keywords had already cost global news media $3 billion. Recently, greater appreciation of the potential to limit reach and revenue has sparked a transition from pure avoidance to suitability-focused targeting - yet plenty of room for error remains, especially when it comes to video.

The main flaw of keyword blocking techniques is, of course, fuelling false positives and negatives that veto quality inventory while failing to protect against less obvious risks for specific brands. But simply prioritising safety doesn’t solve these difficulties. For example, many solutions aimed at finding appropriate content miss crucial contextual signals and incorrectly classify video due to continued reliance on text as the only means of semantic analysis - such as transcripts, comments, and metadata. There is also a persistent tendency to categorise videos by repurposing broad pre-existing platform labels and use this data to manually build static inclusion lists.

This ongoing imprecision means that despite encouraging progress, the nature of content is still frequently misread, and ads are often misplaced. Taking the suitability shift further calls for greater evolution towards more nuanced assessment on a video-by-video basis.
 

Bringing back a refined human touch

Typically, the argument in favour of blanket evaluation led by smart technology is based on the need for scale and speed. Agility and granular analysis, however, aren’t mutually exclusive. In fact, research has illustrated that leveraging machine learning and deep human cognition together can produce better results than standard video on multiple fronts, including higher brand favourability, recommendation intent and trust for major platforms such as YouTube.

What makes the vital difference is tapping into human intelligence to feed algorithms with higher quality data. When tools are trained to make informed decisions using human insight - instead of looking for blanket terms and trying to glean details from minimal text - they can drastically boost accuracy while maintaining real-time efficacy. For example, human-generated classifications of what is suitable for each brand can be blended with wide-ranging content reviews to create a rich pool of data that helps algorithms learn and spot contextual patterns: enabling them to consistently find and target the right kind of video.

Obviously, closer collaboration between humans and machines is only one aspect of the journey to seamless, suitability control. As well as taking advantage of new capabilities in refined content analysis, the industry must collectively strive for better unity on suitability standards that will allow buyers, sellers and partners alike to work from the same page.
 

Establishing more accessible frameworks

Several major forces have set the ball rolling on making digital a more responsible, accountable and suitability-orientated space. GroupM has recently launched a responsible investment strategy to ensure spend is allocated to ethical media and away from harmful content. From the video provider perspective, YouTube has become the first platform to receive MRC accreditation based on the Global Alliance for Responsible Media (GARM) brand safety floor.

There is, however, a need for brands to play their part. In particular, they can accelerate change by being selective in data partner choices and opting for those who are joining the mission to establish a common language for content analysis by embracing initiatives such as GARM’s universal suitability categories. Escalating the movement to implement shared definitions will be essential to set suitability as the industry-wide norm and transform current attitudes to how risk is perceived and managed for every format - video included. 

A regular feature on advertising agendas before the pandemic, video isn’t set to lose its star status anytime soon. Equally, however, the video landscape and brand requirements are only becoming more complex. To maintain momentum and reap optimal rewards, brands must enhance the sustainability drive by stepping up their game; utilising a mix of human cognition and intelligent assessment to delve into real context and pushing adoption of universal frameworks.