Publishers generally view brand safety as a necessary means to protect their advertisers. Hard news sites and user-generated forums tend to toe the trickiest of lines when it comes to advertiser brand safety. No publisher wants to plant a sign saying “Danger: Beware of Unsafe Content” which inherently scares off a revenue source. But there may be a little bit of not seeing the forest for the trees in the sell-side’s aversion to brand safety technology integrations.
I’m not the first person to suggest that publishers might have something to gain from embracing brand safety. Usually, such arguments have been made on the grounds that brands will happily pay more to avoid reputational injury, so publishers who quantify their brand-safe inventory can sell ads at higher CPMs. Those increased CPMs, the thinking goes, will compensate publishers for the added costs of implementing brand safety technologies. But can they reasonably be increased enough, on the promise of brand safety alone, to offset the loss of inventory classified “unsafe”? Probably not. Publisher inventory is not valued on safety alone, however.
Among the many variables that go into the CPM calculus, contextual relevance is becoming increasingly important. Largely due to new data privacy-oriented regulations and browser updates, good old fashioned contextual targeting is making a comeback–and that poses a challenge when it comes to digital contexts. Contextual targeting requires aligning publisher content and brand messaging, but – as the many brand safety blunders have shown – all the publisher content out there is a largely unmapped territory. It’s pretty hard to target anything without a map.
If we can assume that brands are willing to pay more to avoid brand safety reputational risk, then it’s all but guaranteed they’ll pay more to reach receptive audiences in brand relevant contexts. And it happens that technologies used to classify content as brand-safe are the same as those used to classify content as “automotive” or “fashion”. Brand safety technology is really just contextual targeting technology applied defensively––contextual “anti-targeting”, so to speak.
Publishers want to classify their content, because contextual targeting-ready inventory will increasingly command premium CPMs, but they’re inclined to shrink from supply-side brand safety tech. There’s a bit of cognitive dissonance there. Publishers should start looking at brand safety technologies in a different light. The exigency of contextual ad safety for brands has inspired the development of advanced contextual analysis tools. Those tools – which can deliver intelligence about not just the superficial nature of the text, image, and video web content but also deep insights about sentiment, tone, etc – can actually serve a contextual targeting function.
While an argument can be made that it’s the demand side that benefits most of the high-quality contextual intelligence, there are challenges with regard to accessing that intelligence pre-bid that publishers are uniquely qualified to address. While brands and platforms can effectively integrate contextual analysis tools for brand safety and targeting pre-bid, most of what’s out there is limited to keyword and metadata search. The level of contextual intelligence that can really boost inventory value requires insights that come from image and video analysis using image recognition, as well as text analysis using natural language processing algorithms that can actually parse the meaning of what is being said.
The challenge of applying those kinds of contextual analysis tools in a pre-bid situation is technological: Processing power and download times––especially with video––prevent those tools from conducting the kind of full-scale, real-time content analysis that delivers the most actionable contextual insights. But publishers, who own the content and can analyze, categorize, and tag it with contextual insights prior to publication, are able to offer the kind of fully vetted inventory that is truly deserving of the premium CPMs associated with safe, contextually relevant inventory.
Sure, content classification tools and technologies will end up blocking some inventory in the name of brand safety, but they will also open up new inventory and increase overall inventory value. Plus, publishers who embrace them get the bonus of being seen as partners who care and are willing to do what it takes to give buyers safe and potent advertising opportunities.