While big data has already been harnessed effectively by some pay TV operators and OTT service providers that have one-to-one relationships with their subscribers, both commercial and public-financed broadcasters have struggled to develop effective analytics strategies.

That is changing, in part through concerted efforts by major broadcasting bodies such as the EBU with its Big Data initiative launched in July 2015. But almost inevitably it has been a few trail blazing broadcasters that have lit the stage for others to follow, with Brazil’s TV Globo in the spotlight at IBC this year. As the largest commercial TV network in South America and second-largest in annual revenue worldwide just behind the US ABC Television Network, Globo saw the potential pay back from early investment to incorporate big data analytics in core operational and business development areas. 

The network's paper describing the implementation of big data in three of these areas has won the IBC Best Conference Paper Award this year. Entitled ‘Big data for data journalism, enhanced business analytics and video recommendation at Globo’, it was awarded for having delivered proven benefits in different sectors at a time when many broadcasters are still barely at the implementation stage, according to Paul Entwistle, who leads IBC’s peer review team judging papers entered for the award.

“A lot of people are talking about the potential of big data applications in the media industry, but TV Globo has actually done it, and is reaping real benefits from it,” said Entwistle. “What makes the paper particularly interesting is that the authors talk about three very different applications: using public data to drive the news agenda; using business analytics to understand audiences; and using machine learning to recommend content to consumers."

One key point common to all three of these sectors was the use of the cloud for processing the data, both for scale economies and to perform the most effective analysis. The paper discussed how the data is hosted and processed in the Amazon Web Services cloud, which has helped provide the operational advantages and access to new skills required to manage the analytics.

TV Globo has also been willing to work with partners to acquire the required analytics skills and technologies. One such partner is comScore for its Digital Analytix package which ingests, validates and processes unstructured data. Another partner is Los Angeles-based Parrot Analytics, whose role is to help Globo map new trends and identify content production opportunities worldwide. As the world’s largest producer of telenovelas and producer of 6000 hours of content each year in Brazil, the broadcaster is keen to harness that elsewhere, both in and outside Portuguese speaking areas.

Globo was interested in Parrot Analytics’ proprietary technology measuring cross-platform global demand for TV programming, as well as research on viewership. It measures demand across multiple languages and the firm claims it to be the only product capable of providing a global view of audience demand in any country at any given time, on all major platforms.

BBC Worldwide is another client of Parrot Analytics, also seeking to exploit its content overseas and reduce dependence on license fee income in the UK. The BBC is in a similar position to some other public service broadcasting members of the EBU, which have identified big data as a central theme essential for developing one-to-one relationships with viewers, where they lag behind pay TV operators. This was a major motivator for the EBU’s Big Data initiative, which is also aimed at boosting content creation via data-driven editorial production and programming, like TV Globo.

Recommendation will play a key role here by enticing viewers to sign up voluntarily for logging on to broadcasters’ online portals with their credentials, even though they can access the service free of charge anonymously in their home markets. Broadcasters will have to offer effective recommendations that persuade viewers it is worthwhile undergoing the extra step of logging on, offering compelling suggestions.

The EBU’s initiative is focused on four key areas recognizing the importance of recommendation and audience measurement. These are improvement of audience measurement techniques, implementation of recommendation systems in line with PSM (Public Service Media) values, development of new content strategies tailored to audience behavior and preferences, and development of common EBU principles to meet data protection and privacy legal requirements. The latter is essential to win the confidence of users and make them feel comfortable signing on.

Again, like TV Globo, the EBU is highlighting the importance of the cloud in big data analytics. This is reflected in it’s a forthcoming seminar at its Geneva headquarters on November 2017 called Cloud Intelligence, covering application of Machine Learning to data-driven tools for personalization and automatic editing with AI.