Contextual advertising can help brands reach and respect diverse audiences.
Contextual advertising is a digital version of print media, where all targeting is based on the content to reach a wide range of audiences.
The technology behind digital advertising is a barrier to diversity. Consider campaigns targeting LGBTQ+ audiences. Most of the supply chains default to brand safety tools based on keywords, with blocklists that are broad and exclude content related to sexuality. This breaks the links between brands, publishers and audiences, essential to a diverse advertising media ecosystem.
There’s also the problem of invasive behavioural advertising. Even if anonymized, attaching protected characteristics to advertising identifiers is not ethical or legal. This is because shared devices, household targeting and data leakage can expose sensitive personal information.
This risk is not present with contextual solutions, as the targeting begins from the content, not the individual. Users can experience ads aligned with their interests through AI algorithms with contextual learning. This is done without leveraging any identifiable data and without being “followed” around the internet.
What is the current law on advertising that targets protected characteristics?
In the UK and the EU, GDPR prohibits processing special category data, which includes race and ethnicity, sexual orientation, and political and religious affiliations, without explicit consent. The Digital Services Act reaffirmed the EU’s stance on this issue, saying: “Platforms cannot present behaviorally targeted ads… that are based on profiles based on special categories of data.”
The proposed America Privacy Rights Act will likely include similar provisions, given how closely the US has followed the EU regarding privacy laws.
Vendors specialising in behavioural targeting in ad tech will undoubtedly find ways to get around these restrictions. This targeting might not be able to use special category data unless explicit consent is given—and let’s face it, consent walls with confusing terms and conditions can be a poor indication anyway—but by patching various proxy characteristics together, the law can easily be skirted.
It was revealed that the UK government had been using to invasively target Black, Muslim and LBTQ+ people on Meta by using hip-hop, Hijabs and RuPaul’s Drag Race. Fines imposed by the ICO show that proxy targeting is indeed a violation of data protection laws.
It is dangerous to target people based on their protected characteristics. A gay person living in an intolerant household could be outed by household targeting. Or, sensitive personal information stolen or leaked from the advertising supply chains could be used to blackmail a person.
Does it make sense to risk so much just to improve the precision of targeted advertising?
Advertising that is targeted without violating privacy
Although surveillance advertising is not allowed when it comes to protected characteristics, brands, publishers, and consumers all lose out on the opportunity to reach diverse audiences through relevant advertising. Advertising revenues that are only available to the mainstream will stifle the micro-economies and communities that serve the underrepresented.
Brands create products that are specific to ethnicities or gender identities. For example, shampoo for people with type four hair and cosmetic products for the transgender community. These products would struggle to find an audience if they couldn’t reach a wider audience. Publishers focusing on minorities would have to limit their advertising revenue if they couldn’t sell the unique characteristics of their audience. Meanwhile, consumers would miss out on events, products, and content that are relevant to them.
How can we reach diverse audiences without compromising their privacy and exposing them to harm?
Digital advertising is constantly looking towards the future but it can be helpful to look back at successful models from the past to get some inspiration. Print advertising was a successful medium for publishers and advertisers decades before digital advertising and audio avertising .The idea was simple: advertisers would sell ads based on the promise to expose certain audiences to specific content.
It was unnecessary to know who read the pages, but quality publications flourished, including those serving minorities.
Digitally, contextual targeting offers the same results. The concept has existed in some form ever since the early days of the Internet but was held back due to the difficulty of correctly tagging the content and the popularity of cookie-based behavioural targeting.
The contextual targeting solutions of today are different. AI models can automatically classify content into context categories with incredible speed and scale, automating a once laborious task and aligning the context with audience taxonomies at any level of granularity that advertisers require.
Content is not the only factor determining audience alignment. Anonymous usage patterns within publisher properties and across networks can reveal the overlap of interests. This allows advertisers to target different audiences across websites based on their actual consumption patterns rather than stereotypical assumptions.
No personal data is collected or used to create these audiences. The content is used to model all targeting, just like in the print age. This allows diverse publishers to sell to their audiences and eclectic brands to target them.
AI, however, is not the answer to advertising’s privacy issues. Examine the commitment of AI vendors to data security and privacy and whether they have robust measures to protect sensitive information.
Ask for full transparency on how AI models are created. Vendors who are hesitant to disclose the data input, the model used, or the methods employed to address ethical or bias concerns should be a significant red flag. Be wary if complexity is being used to hide the true nature of AI.
The advertising ecosystem is at a crossroads. It can continue to play cat-and-mouse with regulators and harvest vast amounts of data to maintain invasive surveillance ads through dubious proxy services. It can also embrace technologies that do not require personal data and invest in future-proofing solutions that target diverse audiences safely and respectfully.
Also Read – Digital-Driven Audio Advertising is Gaining Momentum in the U.S.