BBC: big data helps us understand the audience

Big data is important to the BBC in helping to understand audience behaviour and sentiment, BBC CTO John Linwood recently told Computing.

"There's a piece of work going on now in understanding the relationship of a member of the audience and the BBC," explained Linwood. "We have two interests: understanding audience behaviour, and how they're interacting with us."

Big data is also about driving smarter decision-making in internal systems and understanding key cost drivers, he added.

"We're starting to use big data to help us make smarter decisions," Linwood said. "The challenge is that historically the BBC has not had this richness of data, so very often the inbuilt systems at the BBC don't take account of the information we now have available.

"If you look at social media, the sorts of web-based traffic information available these days is relatively new for the BBC to deal with."

The task for the BBC now is for the marketing department to understand what they can do with the information. Then the challenge for the technology department is how to mine the data and make it useful.

But this big data challenge is one for the whole organsation, Linwood said.

"You want to have capabilities that are shared across the whole organisation as much as possible. Big data is as important to finance as to marketing or even programme makers," he said.

"Imagine a talk show where we can see in real-time the sentiment in the audience reaction to what's going on. The director could ask the presenter to extend the discussion with a particular interviewee due to positive audience sentiment – or the reverse situation.

"Making real-time programme decisions is the sort of thing we're intrigued by in terms of what information we could be mining."

He admitted that measuring audience sentiment, for example across social media, is tricky.

"When something happens on a BBC programme, someone on Twitter immediately starts talking," he said. "But it's not immediately obvious, especially if you're doing it with technology, to know whether it's positive or negative. So are there ways to understand sentiment in near real-time? Or if there's misinformation out there, how quickly can you react?

"You can't have armies of people monitoring Twitter all the time, so you need ways to automate that," Linwood concluded.

Computing's Big Data Summit will be held on 20 March 2013 at the Tower Bridge Hotel in London. The event is free for end user delegates.

There are still some speaking and sponsorship opportunities available. For more details, visit the website.