Contextual targeting is making an AI-powered comeback

Data privacy: a new era for audience targeting

If there’s one word that has been top-of-mind for all players in adtech over the last two years, it would be ‘privacy.’ Companies all over the globe have begun to crack down on privacy measures, making it harder for advertisers to target their audiences.

Apple in particular has introduced several measures to reduce ad targeting and measurement capabilities: hindering access to location data coming from iOS apps, introducing AppTrackingTransparency – which prompts users for their opt-in before sharing their data with apps – and announcing plans to obfuscate IP addresses with its new Private Relay iOS 15 feature. However, the most significant change will surely be Google’s plans to remove third-party cookies from Chrome in 2023, as this will completely alter the way marketers effectively target at scale.

While this privacy-focused sea change has only recently hit its apex, the concept of advertising getting ‘creepy’ has been a long time coming. On the one hand, idealists – or anti-advertisers – either created ad blockers or exploited data with exasperating clickbait arbitrage schemes. The realists, on the other hand, attempted to create better ads through paywalls, native ads, platforms and more to facilitate the creation of better content. Substack, for example, has created a space for writers to send newsletters directly to their readers with a digital paywall in place to access content. The platform has given journalists free reign to create content without the traditional editorial oversight, as well as a financial incentive via their own base of paying subscribers.

Advertisers must roll with the adapting ecosystem and live with the fact that consumers are uncomfortable with the idea of them collecting and using their personal data. However, without some measure of targeting, advertising simply cannot fulfill its purpose to connect with and reach its intended audience.

But don’t despair; this doesn’t mean advertisers need to go back to burning money buying ads based on broad contextual categories and then crossing their fingers hoping for conversions. The future of targeting and advertising will have to rely on a more intelligent version of what worked before the internet. What’s needed now is a high-performing combination of contextual data, machine learning and AI.

AI and emotion-based targeting

Prior to the many data points and targeting options that exist today, early targeting methods consisted of tedious manual work. Advertisers would look through the catalog of available websites, carefully choosing those they believed their target audience would visit. There were no elaborate cookie trails. With the rise of social networks in the early 2000s, display ads became increasingly targeted and more personalized – not only to a given user’s demographics but also to their interests, ultimately becoming what we now know as behavioral targeting. Today, with new privacy policies and laws altering the digital space once and for all, many experts believe the future of advertising will be anchored in contextual targeting.

The truth is that we no longer need humans to scan through the contents of a webpage to determine its audience. Equipped with natural language processing, AI can be applied to understand the overall sentiment of a page instead of analyzing simple keywords. Furthermore, with image recognition technology, we can have the full view of the imagery and video content included in a page in mere seconds. AI is always learning and improving its decision-making process, and is more often than not backed by real people. The future of AI and targeting will not only lend itself to a more intelligent contextual platform, but also, somewhat ironically, a more human, empathetic and emotional experience for users.

On top of this, the amount of readily-available contextual data that advertisers can piece together without infringing on personal data is incredible. Today, contextual data can include the time of day, local weather, local current events and trends, the publication and the article content. Adding machine learning to the mix would enable a continuous learning cycle, allowing advertisers to figure out which combination of data is creating conversions. This would then provide a dataset that could mirror performance previously achieved with user data.

The future of contextual targeting for audio and video ads

While understanding the actual context of content has long been a blind spot to mediums such as audio and video, the advances of AI have made this task as simple as gaining the context for a text article.

AI can now analyze audio content instantaneously by transcribing words in real time. There’s even a way for AI to clone the voices of our favorite celebrities, a concept known as audio deepfakes. Not only that, but by combining AI and machine-learning practices, we can determine and categorize tones of voice. All of this contextual data opens up a wide range of possibilities for audience targeting – a notable use case is modifying ads in real-time to match the context of a podcast.

And then there’s video. Livestreams have become our favorite way to engage on social media platforms, as is evident in the rise of TikTok and Instagram. By adding AI capabilities into the mix, we could completely customize the video-streaming experience with live built-in product placements based on the context of the video itself. Therefore, we will move beyond straightforward video categories such as health, food and travel by adding more complex data such as sentiment, imagery and sound to allow advertisers to target their audiences effectively.

The advertising world’s journey with third-party user data may be coming to a close, but modern contextual data coupled with AI can enable targeting that is equally if not more effective. If advertisers want to succeed in a privacy-first, responsible digital advertising ecosystem, they’ll need help from AI to fully understand the content with which their customers are engaging and in turn create better advertising experiences.

Rob Fan is chief technology officer at Sharethrough.