Cable operators and programmers are looking to content recommendations solutions as a way to offer a more personalized TV experience. First and foremost, next-generation recommendations technologies must be capable of personalizing the primary TV experience.
The set-top box experience is critical and often more challenging than performing recommendations across the second screen.
This is because delivering second screen content is similar to VOD architecture and traditionally easier to implement in comparison to the primary TV experience, which involves dynamic and linear TV (including EPGs).
These next-generation recommendations systems rely on a combination of robust metadata and a variety of recommendations algorithms, including implicit and explicit data points, direct user feedback such as likes and dislikes, and consumer viewing behavior.
But also crucial to these recommendations and navigation technologies is understanding the consumption data and user patterns of linear real-time programming so that operators can maximize their content libraries and add value to existing services. Success hinges on being able to extract valuable and useful learnings from all the noise in the zapping data, and relying on a methodology and approach that is capable of scaling across the demands of real-world operator deployments. For example, a pay-TV provider with 10 million subscribers offering 80,000 programs would need interacting algorithmic technology capable of supporting 800 billion recommendations combinations. Incorporating these consumer-centric metrics creates an opportunity to understand a consumer's unique preferences, incorporating data such as explicit preferences, implicit preferences, moods, likes, dislikes, etc. Additionally, this kind of data can extend the use of important consumer learnings to any part of an operator’s organization, such as marketing, retentions, package configuration, acquisition, and channel marketing.
The paper will delve into the technical specifications and advantages of implementing personalized recommendations techniques that personalize cable TV services on the primary screen and extend that experience across the broader device ecosystem. It will compare and contrast competing approaches, and outline how incorporating consumption patterns and viewer usage patterns improve the overall accuracy of content recommendations. It will look at real-world implementations of how operators today are using these new data sets to truly understand and enrich the subscriber experience.