Blocklists are still failing advertisers and minority media

These last few years, buyers have become aware of the choking point depriving these titles of oxygen. Keyword blocklists exist to ensure ads don’t appear on dangerous pages. As one example, airlines don’t want to advertise on stories discussing ‘hijacking.’ But these lists could number up to 10,000 terms. It’s been a blunt and broad approach that has cordoned off huge swathes of safe media. Titles including Attitude and PinkNews faced having as much as 73% of their stories flagged as ‘brand unsafe,’ with terms such as ‘lesbian,’ ‘bisexual’ and ‘drag queens’ on the lists.

Some brand safety vendors claim that the blocklists can be more effective if the software goes beyond topline keywords and subsequently defines the context of the page. The question is, can algorithms effectively monetize articles about ‘gay people’ ‘hating’ government policy while dodging those encouraging government ‘hate’ against ‘gay people’? For many, that’s where inclusion lists come in, blocking or encouraging placements by domain, rather than page.

Two years ago, brands were negligently unaware of the huge word spaghetti shield starving them of new, diverse audiences in their home media – that’s still happening.

Speaking to The Drum, Justin Taylor, managing director of Teads UK, compared blocklists to sock drawers. “They’ve built up over time to the point they’ve become unusable. Based on outdated technology and overzealous boardrooms, they’re in desperate need of a clear-out.”

In slightly more evocative terms, Claire Atkin, co-founder of consultancy Check My Ads and a leading voice on the topic, claimed “a bad-faith interpretation of brand safety” meant minority media was being blocked “while white nationalists get a hall pass.”

The issue

Taylor admits that the Black Lives Matter movement did help spotlight the issue of representation in media but it never fixed the problem.

“We can still see lists of over 8,000 words from clients, blocking content from a hugely diverse spectrum of content even though the content that is truly unsafe for brands, such as terrorism or drug abuse, is already blocked by standard technology options. But an over-zealous brand that includes the word ‘pot’ as an additional item will lose far more culinary eyeballs than smokers.”

There remains bias in the composition of these lists. Taylor admits he still sees ‘hood’ on blocklists even today. “They’re restricting reach among far more fashion readers than anything else. And it’s also incredibly prejudiced.”

“This is crucial content that is often brand-safe and not only should be supported by ads, but ads seen on such content will be seen even more favorably by the reader for doing so.” These measures have helped make some business models “unsustainable,” which results in publishers “committing less time and attention to communities and topics that really matter.”

Inclusion

Inclusion lists do the polar opposite of the blocklists. Some advertisers only appear on the pages of pre-vetted publishers. This can benefit premium, brand-safe titles. It can also be used to ensure minority media gets included in plans and receives backing.

Taylor urges buyers to “look at practices, processes and plans that have developed over time and see where intrinsic biases lie.”

Ben Pheloung, general manager of Mantis at Reach, agrees. He says there have been “some great strides by brands and agencies to remove discriminatory keywords from their blocklist, but there are still a lot of words and topics that are being blocked that need to be re-evaluated.”

Lauren Douglass, senior vice-president global marketing at Channel Factory, helps moderate an inclusion list of 10,000 creators since June 2021. Douglass admits that the industry needs to collaborate on these lists because “at the moment, there are no set guidelines, standards or rules in place.”

“We are constantly vetting them, and we’re also creating opt-in lists so that people can identify themselves according to their unique descriptions. For example, if I am a Black transgender creator, I can describe myself as such. This avoids any potential for profiling people or adding them to groups they don’t belong to or relate to.”

Brands are still tiptoeing around subjects deemed political – all too often that means avoiding discussions around the rights of minority peoples. Douglass says: “Imagine someone is discussing BLM and the latest protests, or posting content from a protest. That creator could be negatively impacted by content moderation for months until they’ve created more content that pushes that topic down. It can be a tedious and incredibly frustrating situation.”

Meanwhile, Lee Mabey is managing partner and DEI lead of media at Dentsu UK. For the network, he operates Together – its proposition to boost clients with under-represented media audiences.

He says the main challenge is in identifying the best routes to market without seeming tokenistic. “You cannot activate without truly understanding where you fit in with marginalized groups.”

The group is rolling out numerous campaigns in the coming months to “demonstrate the value of investing time to identify what matters to underrepresented audiences.” Most of the brands it works with use inclusion lists rather than blocklists as a default position.

Custom creative

Do marketers need bespoke creative to operate on these titles? Mabey says it is possible to run broader campaigns, but adds that in the age of “dynamic creative optimization” these spots could be tailored easily enough.

Furthermore, there are difficulties in bringing broader campaigns into minority media. “Potential brand partners must respect their values. For example, queer publishers will not carry ads for travel brands promoting destinations where LGBTQ+ folk face persecution, and Black publishers will not work with beauty brands that do not cater to their audience.”

Work to be done

The European Association of Communications Agencies (EACA) has today launched a guide to help buyers improve their buying from a D&I perspective. It admits that keywords/blocklists do inflate campaign CPMs by as much as 70%. It encouraged slimmer lists populated with real brand safety concerns.

A strong start is to remove communities terms such as ‘jew’ or ‘gay’ and their plural permutations from the lists. These are not derogatory terms and in most vetted media do not present any real risk. They’ve either been included due to historic bias and/or tech’s inability to differentiate from hate speech content.

The report says that keyword blocking “should be” topped up with semantic analysis to maximize available safe pages where possible. Advertisers sticking with keywords should focus on populating lists with the “worst of the worst” first.

Pre-bid technology enables the targeting of positive content before the implementation of these blocklists, and it also blocks most at-risk categories too. Marketers should regularly review these lists and phase in (and out) terms as appropriate.

Finally, marketers should look to actively include LGBTQ+, 50+, gen Z, Black, Asian and Muslim titles – ones they’ve likely completely missed until recently.

As the adoption increases, marketers will next have to face the biases or shortcomings of inclusion lists. Many of these media titles, largely due to the aforementioned shortcomings in ad revenue, can be small. Some may be prone to the misinformation, hate or propaganda marketers say they’re keen to avoid. The EACA report urges marketers to have a diverse panel review any transgressions before cutting these titles off the buy and going back to square one.