The digitalization of television is dramatically expanding the possibilities for advertisers to address target groups on TV. Contextual targeting plays a major role here - but what parameters are we talking about in the Connected TV (CTV) sector? And what role does metadata play? In the following, we will show how effective targeting in CTV works through content analysis, to what extent capturing context improves the advertising experience and how generative AI can ensure even more successful campaigns. A guest article by Mario Neumann on ADZINE.

Connected TV is a real advertising Eldorado for advertisers - engaged target groups meet special interest content, digital devices allow granular targeting based on GDPR-compliant data. What's more, it is possible to measure the success of campaigns in real time.

The current cookie-less debate hardly affects CTV either, as this type of media works without cookies per se. Is that a disadvantage? Not at all, because firstly, third-party cookies are an increasingly obsolete targeting method anyway, and secondly, without cookies, the focus shifts to contextual targeting in order to reach target groups in a creative and effective way.

Contextual targeting for relevant advertising

Contextual targeting uses users' interests, which are determined based on the content they consume. This method not only makes advertising more relevant and attractive to viewers, but also increases the effectiveness of campaigns. Unlike traditional targeting campaigns, which often take demographic data such as age and gender into account, contextual targeting in CTV allows target groups to be addressed more precisely according to their interests.

Reach the right target audience with metadata in CTV

CTV campaigns can draw on extensive content information to address target groups in a targeted manner. This metadata not only includes information such as the program title, description and keywords of the program, but also data on viewer behavior, such as preferred genres and viewing time. This allows advertisers to tailor their messages to suit the current mood and interests of viewers.

CTV also uses technical metadata such as device type, platform and similar (e.g. operating system) as well as geographical location for geo-based targeting. This metadata can be supplemented by measurements such as streaming views, completion rates, frequency of use or dwell time. With such measurement data, advertisers can track and optimize the performance of campaigns in real time.

Content metadata is collected both manually and automatically. Automatic generation uses text recognition algorithms and machine learning to accurately categorize streaming content. Capturing and analyzing the content metadata then makes ad matching possible: one example would be the display of automotive advertising in car-related content.

Content meets context

However, even better targeting requires more than just content metadata. If advertisers want to use the full breadth and depth of the context, the existing metadata concept must be expanded: In addition to user behavior, emotional states of the viewer and objects in the content such as "beach" or "car" can also be captured in the metadata. The challenge in the first step is to automatically capture all relevant characteristics of the images and videos in which advertising is integrated.

The second step is to create target group segments. So-called "content similarity detectors" analyze the metadata in terms of content and context. On this basis, these tools find the corresponding audience segments that are linked to similar viewing behavior, interests and topic context - keyword "content similarity". The idea behind "content similarity" is therefore to identify similar content and create target group segments on this basis that offer advertisers guidance. Similarity detectors are based on the concept of the digital twin, as used in digital marketing.

For modern targeting, campaign professionals should record content and context. Metadata that maps both will therefore become the new gold standard in connected TV. This allows advertisers to create a seamless and natural advertising experience throughout.

Contextual advertising improves the viewer experience

Not only advertisers benefit from such optimized advertising campaigns, but also viewers. Contextual ads are perceived as less intrusive and are therefore more likely to be accepted. This improves the viewer experience on CTV. Since advertising takes place in the viewer's area of interest, they interact more with the advertising content than with non-contextual advertising. For this reason, contextual ads are more effective and achieve higher conversion rates than non-contextual ads.

In addition to I/O bookings, ads can also be displayed programmatically based on metadata. This enables marketers to achieve even more precise segmentation and real-time targeting via demand-side platforms (DSPs).

Outlook: AI adapts advertising media to content and context

Current tools show that artificial intelligence and contextual targeting go well together. At a new level, AI systems are able to understand content in text and images. This will allow advertisers to customize their advertising based on content categories that go beyond the usual classification. An important further development is the ability to create existing advertising material in such a way that it matches the context and content. The supreme discipline in this area is "predictive personalization" - a method in which AI uses data analysis and predictions to create personalized advertising that is tailored to the future needs and interests of viewers. As in many other areas, AI is paving the way for even more successful campaigns on CTV.

The article can also be found here on Adzine.