Nowadays, recommendation systems are everywhere, ranging from e-commerce sites, social networks, all the way to most media streaming services. However, in most pay-TV platforms, these recommendation systems do not usually achieve a suitable level of effectiveness, even when a huge linear and on-demand content catalogue is available. This paper suggests some reasons for that issue, exemplifying some of the shortcomings found in a south European pay-TV system. Driven by the actual perceived limitations, an innovative personalisation system is proposed, and the current state of the prototype development is described. As far as the prototype is concerned, it is already possible to present some identified insights that will be validated in an ongoing field trial and subsequent user evaluation.
Read full abstract