Can we protect advertisers' distribution contexts while maintaining revenue?

Should brand safety come at the expense of publisher revenues?

The emergence of anxiety-provoking content justifies the desire of advertisers to ensure that their ads are broadcast in controlled environments. Advertising performance is also key. The appreciation of advertising is higher when campaigns are broadcast in quality contexts. It is easy to understand the interest for brands to want to protect their distribution frameworks and seek to ensure that their content is not associated with negative news. However, too restrictive use of keyword lists by advertisers certainly protects sensitive distribution contexts but has an impact on the advertising revenue of publishers. Because in a context where caution is the key word, advertisers will choose keywords to ensure that their ads appear in environments that suit their brand. This can generate very large block lists and thus reduce the monetization of inventories if too many editorial contexts are excluded from distribution frameworks. To establish increasingly large keyword blocking lists, which act more as a foil than as a targeting technique, advertisers are potentially depriving themselves of quality inventories. Publishers are beginning to take a stand, like Vice Media Group, which in 2019 called on advertisers to review their brand safety strategy after showing that the majority of blocklists were aimed at content devoted to diversity and inclusion: “Faced with this observation, VICE made the decision to no longer block the list of 25 words and expressions related to diversity, including words such as Muslim, transgender, refugee, or even interracial”, Recall the press release.

From brand safety to brand suitability

Although publishers must provide a “brand safe” environment for advertisers, the essence of a news site is to cover news. If these are unfavorable, some content will be abandoned by advertisers. Rather than opting for an avoidance strategy by blocking certain keywords, advertisers must challenge the relevance of a distribution context and seek the adequacy of their advertising message with the editorial content. Indeed, a keyword used out of context has little meaning. These efforts can be supported by the use of semantic analysis and targeting solutions that are necessary to move from a keyword logic to a brand suitability logic. They offer publishers the opportunity to value the “good” locations and therefore to make the best use of their inventories while having control over them.The emergence of these new technologies, most of which have been made possible with the democratization of artificial intelligence, offers many possibilities whose limits we will always seek to push.

New models to implement

Should we then question the words in the keyword lists and look for more contextual solutions? The guarantee for an advertiser to broadcast an advertising message in an environment that is 100% controlled and in line with its values is complex but is answered in the use of hyper-contextualized solutions based on semantics. This targeting approach is currently the most successful and nuanced. Indeed, semantic intelligence is capable of fully understanding the editorial context of a web page and the associated feelings, positive or negative. The Machine Learning algorithms behind this technology make it possible to process an exponential volume of data and to offer hyper-contextualized targeting. By combining semantic data with the automation of buying and selling processes, it is now possible to optimize the distribution context by analyzing and filtering content and associated themes in advance. The brand suitability approach is fine enough to not penalize either advertisers or publishers and is a solid alternative to exclusion lists. Preserving distribution contexts while maintaining the economic balance of news sites is a complex issue that concerns all actors in the advertising industry and not only advertisers and publishers. Concrete and easily activatable solutions are beginning to emerge and semantic intelligence is one of them.