ShowHeroes SemanticHero is our semantic targeting technology. It ensures brand safety and suitability for our advertising clients while publisher partners across the ShowHeroes premium network benefit from higher time on site and view through rates (VTRs). It does not need to rely on personal data or cookies and allows for privacy-friendly distribution and targeting of both, content and ads.
Semantic targeting leads the way to the future of online advertising placement: What used to be a promising technology, crippled by a massive lack of knowledge, computing power and bandwidth a decade ago, is making a big comeback in the 2020s, powered by modern natural language processing on the basis of data science and artificial intelligence (AI). Leveraging these improvements results in excellence at understanding semantics while providing realtime data processing capabilities.
Combined with programmatic buying, semantic targeting unleashes its full potential: Automated real-time buying of individual ad impressions ensures efficiency in targeting audiences and enables marketers to perfectly match their ads to web pages, articles, and even video content.
Thanks to the knowledge extracted from analyzing images, audio and video on the publisher’s websites our AI-driven semantic matching technology – the SemanticHero – is able to select the best-fitting content video from the platform’s video library whenever a request is issued through a ShowHeroes content unit. Not only is the choice of the content video based on semantic matching, also a corresponding ad will be picked by our adserver.
Contextually relevant videos generate up to 100% more engagement in terms of view through rates than manually placed clips. At the same time, semantic targeting comes with privacy protection and brand safety already built in, as it requires zero user tracking.
How to Harvest the Semantic Data?
SemanticHero scans websites in the ShowHeroes network by request and categorizes content and meta data according to IAB categories, keywords and further indicators to the page’s context.
With these machine learning approaches, SemanticHero leverages the power of statistical language analysis to provide human-like understanding of page content resulting in the highest level of automatically determined brand safety in the digital advertising ecosystem.
How to Implement Semantic Targeting for the Cookieless Future
SemanticHero’s website index stores an abundance of more than 40 million websites in 12 languages, collected from our extensive publisher network. Based on the analysis of the web page content, semantic concepts are generated. The sheer volume of extracted concepts allows for the computational approximation of their probability of occurrence per website.
This allows for selecting matching content and predicting the users’ interests when browsing a page. All of this does not involve any dropping, reading, matching or syncing of cookies. Content recommendation and ad targeting can be solely based on page content and external knowledge.