Recommendations Platforms

Recommendations technology is a relatively new field that sets out to provide users with specific recommendations that that are appropriate to their usage habits. In relation to TV, this would mean that a show is recommended, based on a user’s viewing habits.


The driving force behind the pursuit of a perfect recommendation engine is the need to keep viewers satisfied in an increasingly competitive market. This increased competition means that there is a lot more programming available to viewers than there was, even 15 years ago. So helping viewers to find what they might want to watch amongst the huge morass is something which all providers are striving to do.


 Most recommendations are made either based on a single show or episode, a series of shows, or a full user profile.


Key features of a recommendations platform


    * User activity and preferences – the platform can gather anonymous information on user activity across all services. This is relayed back into the engine, with different levels of importance varying from clickstream information to the preferences within email reminders that are sent to the user

    * Content analysis – A content analysis engine usually provides a detailed contextual analysis on all shows, extracting relevant tags and comparing tags between different shows to produce relationships based entirely on the content of the shows metadata. This mechanism is entirely automated without the need for complicated metadata schemes and can produce powerful results

    * Editorial – Regardless of how powerful an automatic recommendations platform is most customers will, at some point, want to make their own manual recommendations. This may be to support a brand new show or series or to filter out a programme from a competitor. The platform generally includes a simple interface to add and remove and modify the automatic recommendations


One other way that technology providers try to make recommendations is by using social media. This is using the idea that people will watch shows that other people have suggested to them. This can involve friends, family, TV critics and newspapers. They differ from other recommendations in that they are not typically relevant to the user who is receiving the recommendation other than based on who has made the recommendation.

 
As the competition to attract viewers continues to heat up, more and more providers will have to find increasingly innovative ways to keep users from becoming disinterested and moving on.