How machine learning technology boosts contextual targeting for advertisers

The value proposition of machine learning for advertisers

To appreciate the synergy of contextual targeting and machine learning, it’s helpful to put a lens on recent history. Advertisers moved away from contextual once audience tracking entered the scene, but the present-day resurgence can be attributed in part to consumers and a growing concern for how personal data is being used.  

“Consumers are more conscious of how their data is being stored, however, 74 percent still want to see ads that match page content,” says Bichoï Bastha, Chief Business Officer at Dailymotion. “Contextual targeting is the answer to this. By nature, the solution is the most consumer-friendly targeting method as it uses context as the proxy for the audience in which to display an ad.” 

Contextual targeting — or better yet, contextual intelligence — is not the same as 10 years ago. Contextual intelligence solutions are rooted in providing the best possible user experience, leading to the best possible advertising experience. 

Critical criteria for contextually relevant campaigns

As new privacy regulations come into effect and browsers limit tracking capabilities, contextual advertising powered by machine learning is table stakes for standardizing targeting efforts and ensuring campaign success. But not all vendors have the advanced capabilities to deploy an effective contextually relevant campaign for partners, one that delivers — or increases — message receptiveness, ensures brand suitability and is privacy compliant.

Vendors must offer a robust list of criteria to help partners achieve campaign goals from start to finish, including relevant categories to increase media performance, best-in-class ad formats to ensure a positive user experience, granular and niche topic targeting, a depth of relevant inventory, proprietary machine learning models for precise content categorization, at scale and optimized performance metrics to ensure brands hit baseline KPIs such as viewability, VTR and completion rate.

Machine learning powers brand-suitable environments 

Ensuring brand-suitable environments can be a highly nuanced endeavor when being considerate of relevant content that aligns with a brand. Machine learning optimizes this process — for example, analyzing a catalog of over 150 million videos from premium publishers, covering more than 200 IAB categories and 500,000 topics — to serve contextually relevant ads resulting in a valuable experience for consumers in less than 100 milliseconds. 

The process starts when a publisher uploads owned video content with metadata — titles, descriptions and keywords. Machine learning models scan the metadata and select which topic from Wikidata to assign to a video through semantic annotation. The annotated topics are surfaced on the page and contribute to SEO. Once a video is annotated with a topic, it is associated with IAB’s categories to be monetized. 

Contextual is even more effective when it is combined with other elements. Partners that offer multiple levers in addition to context-based association will maximize media performance, including people-based, highly engaged global audiences and performance-based audiences as well. These outcomes further ensure advertisers meet their KPIs. 

With so much data across the Internet, especially for advertisers and brands, machine learning is crucial to tailoring a contextual approach to deliver not only positive contextual associations and brand suitability but also to create connections that convert.