Improved access to audience metrics has long been promoted as an advantage of digital services and as reported in Broadcast Tech (November/December 2014) big data on consumer habits could be the key to developing revenue streams to fund future programming.

TV audiences have traditionally been analysed via audience panels – viewers willing to have additional equipment installed to monitor their viewing preferences. Inevitably such audience research has a high cost and resulting small sample size. Now a global search is on for a single hybrid audience analysis system that can seamlessly meld data from linear and non-linear audiences.

The debate also calls into question what future data science skills will be required by the media industry

Granular data is now readily accessible for digital audiences with public broadcaster Channel 4 at the forefront of digital audience data collection in the UK – the broadcaster now has 11 million viewers registered for online services. But while this information provides detail from income level to postcode it represents only 20% of the Channe l4 audience leaving a substantial gap in detailed data for the 80% of linear viewing measured by monitoring just 5,000 households.

The head of the UK’s Broadcasters’ Audience Research Board, Justin Sampson said at IBC 2014: “The future of measurement relies on a hybrid measurement strategy that draws on big data. We have to gather behavioural data but also harness new data streams.” The organisation’s Project Dovetail aims to address this. Specialists companies already monitoring shifting viewing habits include TVbeat, GfK and Amazon Web Services in the UK and Nielsen, Rovi and the Coalition for Innovative Media Measurement in the US.

Audience data is inevitably of particular interest for advertising sell, but increased data does not automatically mean better insight. For public media organisations the key point is how people feel about content rather than just raw viewing statistics. A number of data companies are already focusing on analysing audience behaviours; StreamHub technologies combine digital viewing habits with social media comments, French company VigiGlobe uses algorithm-based methods of analysing social media conversations, while the Canadian-British company CrowdEmotion uses webcams to record facial movements, heartbeats and changes in skin tone to record audience behaviour.

As the creative industries play an increasingly important economic role globally then this entire debate also calls into question what future data science skills will be required by the media industry.