The Evolution of Contextual and Semantic Targeting Reaches CTV

In the wake of dwindling third-party cookies, contextual and semantic targeting is becoming increasingly important. What may sound like a return to classic environment planning for some, represents a serious alternative to user-based targeting for many. The methods for deciphering context or semantics and thus reaching the appropriate target group differ significantly from those of the “old” advertising world. Additionally, the process is now even available for new channels such as connected TV.


We asked contextual experts to go into more detail on what the technological evolution consists of, how it comes into play with CTV and when targeting reaches its limits.


Our Global Director of Product, Data, and Machine Learning at ShowHeroes Group, Patrick Jähnichen provides his feedback on contextual targeting in CTV environments.

Recognizing and correctly classifying film titles is a challenge that the Showheroes Group has dedicated itself to. Patrick Jähnichen estimates that the importance of contextual targeting for CTV is still “quite low”; however, this is mainly due to the fact that there have not been any mature solutions for this use case so far.

There is a lot of interest. All of the CTV providers we talk to are quite aware of the challenges and very interested in the application of semantic targeting and our approach to it. Due to the main use of the actually outdated VAST standard in version 2.0, the possibilities are limited and creative approaches to solutions are in high demand."

The CTV segment must cope without textual information. “The integrations are based on the VAST tag and do not allow Javascript. The big challenge is to get reliable information and metadata with the current content.” “For most channels in linear TV streams, it is still quite easy: electronic program guides are readily available on the internet, and together with channel information and the current timestamp, you can retrieve information about the content that is currently running,” says Jähnichen.


This is more difficult with video-on-demand content, where direct integration with the publishers comes into play.


“They give us access to their video CMS and thus to the metadata of the content, which in turn, can be analyzed and stored by us. A combination of both approaches opens up any connected partners’ CTV inventory for semantic targeting in relation to the video that is currently running.”


There is still some room for improvement in semantic targeting in CTV.


“The next steps will most likely be utilizing subtitle data or transcripts, which will enable even more precise contextualization in the current stream,” predicts Jähnichen. “To put this in an example, not only would Lufthansa advertise in the context of travel documentaries, but perhaps also Jever beer in a favorite series after a conversation about a cozy barbecue evening.”



Read the original article published on Adzine (in German).