BBC R&D investigate mood-based content discovery

With the explosive growth in the amount of content available to viewers, especially through connected TVs, the importance of finding good ways of discovering and navigating that content has grown as well. A personalised experience has gone from being a “nice to have” feature to being vital for helping viewers find what they want to watch.

Personalisation is not a new topic, and there are already many companies offering products that support a personalised experience or technologies to provide that experience. Traditionally, personalised content recommendations are based around factors such as genre, keywords or aggregated viewing habits.

BBC Research and Development is studying this issue from a slightly different perspective, using the user’s mood to steer the recommendation process as well as more traditional factors. This may help users discover content in a way that’s more suited for the “lean back” TV experience, in the same way that streaming audio services such as Pandora do. The BBC use five separate “mood scales” for classifying a TV show, and while this may not seem very high this additional level of classification may be enough to offer new opportunities for recommendations. One of the most powerful features of a good recommendation system is serendipity – those times when it throws up something unexpected that happens to be exactly what you want to watch or listen to. This additional information may give recommendation services better ways of making these less-obvious connections between programmes.

The BBC are the first to admit that mood alone is not enough to drive recommendations, but that doesn’t mean it can’t be a valuable tool in the toolbox.