The paper introduces radio personalisation, as part of the strategy to keep radio relevant, taking into account new available technologies and audiences changing their media consumption habits.

A novel model for proactive context-aware radio personalisation based on the hybrid content radio framework is discussed, and a prototype based on the model is described. Finally, a user evaluation of the personalised radio service is described, and the results are discussed, showing a path of evolution for future linear radio.


Radio has been a strong and popular medium for almost a century. In the last years, technology has created the foundations to quickly build and distribute a broader range of audio services, mainly recommendation-based music playlists, and new players are challenging traditional broadcast radio, acting as competitors or content distributors and possibly as gatekeepers.

On the other hand, traditional radio is slowly losing ground, especially among young audiences. Tailoring radio for each listener and mixing editorial scheduling with recommended content in a seamless way is both one of the possible strategies to keep radio relevant, and an evolution of the radio service itself to better fulfil the goals of public service.

After an overview of a few relevant projects in the area of radio personalisation, we propose a model for radio personalisation services, based on the hybrid content radio (HCR) framework and leveraging on context-aware and content-based recommendations. Recommendations are based on user behaviour and content semantics on the one side, and available listening time, content complexity and channel cost, on the other.

A prototype inspired by the model has been developed to validate the concept, following the framework presented in. Specifically, the purpose was to measure user propensity to switch between linear and on-demand content and user reaction towards automatic proactive recommendations. A first panel of more than 200 volunteers briefly tested the prototype and participated in a survey.

Then, a second panel of 50 volunteers was selected to evaluate interactive and proactive personalisation using a customised prototype installed on their smartphones for two weeks. Very positive feedbacks on content-search and skip functions emerges, while the feedbacks towards proactive personalisation suggest that further investigation may be required.

Download the full technical paper below