The BBC’s radical new data plan takes aim at Netflix

Advertisement

Young people are watching much less of the BBC – less than an hour a day, according to data from Ofcom

Advertisement
. They prefer streaming content, and Netflix is used weekly by 66 per cent of 15-24 year-olds in the UK. Just 28 per cent of them use BBC iPlayer.

The BBC has pledged to take action, and better, more secure, personalisation might be the key. The broadcaster has been placing greater emphasis on data-driven personalisation for its online services since 2017, when it began requiring users to be logged in to use iPlayer. More recently it has been pushing people to login to the BBC website to personalise the news they see.

Now the BBC has built an experimental system

Advertisement
that allows BBC, Netflix and Spotify data to be combined to present people with personalised music, podcast and gig recommendations based on data they control. It’s hoped that, in the future, similar technologies could encourage younger people to engage with more BBC content.

Building new technology to drive views is a priority in online media. Netflix spent about $1.8 billion (£1.3bn) on tech development last year. The BBC has just £236 million allocated across all of its online services, including tech development for iPlayer. And while the national broadcaster isn’t subject to quite the same competitive commercial pressures as Netflix or Amazon Prime Video, its recommendation engine has a more complicated job. 

As well as presenting quality entertainment, it’s in the BBC’s remit to expose its audience to a broad range of content. “If personalisation or recommendations are making people’s access to any content difficult,” says Tom Harrington, a senior research analyst at Enders Analysis, ”and therefore [limiting their exposure to] new ideas, programmes that are socially important … the BBC is failing at one of its core jobs.”

But the BBC has bigger ambitions. A crack team of BBC researchers has been working on a project that would allow you to save and import everything online services know about your viewing and listening habits, and use that data to create customised recommendations of series to watch, music to listen to and gigs to go to.

Specifically, they’ve been building technologies on top of Solid, an open-source Personal Data Store (PDS) developed by Inrupt, to provide a model of how you can have personalised services based on shared data that you control. The work has been led by BBC Research and Development’s personal data products lead Eleni Sharp and principal engineer Bill Thompson.

Solid is built around personal ‘pods’, controlled by a user, who grants permission to external apps and services to read and write data to them. Pods can be set up on a specialist hosting service, or self-hosted by more tech-savvy users. The BBC has called its project My PDS. 

Sharp says people in focus groups repeatedly said that they were “in too deep with the data, that they didn’t know what to do, they felt like it was out of control.” “Holding people’s personal data in order to be able to shape your services carries regulatory risk,” Thompson adds. “The security model that we are implementing around the demo of My PDS, we think, reduces that level of risk.”

The main benefit users identified from the My PDS project was that it gave the ability to see and organise the data services held on them. To the team’s surprise, the user groups immediately understood the benefits of the data store model.

My PDS pulls in viewing data from Netflix, your listening data from Spotify, and iPlayer and Sounds history from the BBC and. It gives users control over how they share and use very specific sets of personal data, but the original copies of that data are still held by Netflix, Spotify and the BBC. The experimental app invites users to create a Solid pod or sign into an existing one, and grant My PDS permission to read and write data to your pod. Using existing APIs, the team built connectors to pull in data from Spotify and BBC online accounts, and provide an option to manually import a Netflix viewing history.

Advertisement