BBC Uses AI to Find New Audiences for ‘Sherlock’


Viewers in Britain and the United States have been clamoring for the return of the critically acclaimed BBC series “Sherlock,” which debuts Jan. 1 with a special episode set in Victorian times.

But where else in the world might the British broadcaster find viewers for the contemporary interpretation of Sir Arthur Conan Doyle’s fictional detective? To sniff out clues, BBC Worldwide has retained Parrot Analytics, a New Zealand firm that uses artificial intelligence and data science to evaluate global demand for TV shows.

“Parrot suggests very very strong global demand, including in Germany, China, India and Singapore,” BBC Worldwide Executive Vice President David Boyle. “What’s most interesting are the countries with the highest demand but where we haven’t seen it come through in previous deals.”

Since the advent of television, programmers have struggled with measuring the audiences they already have, let alone predicting where they might be in the future. The industry’s traditional approach to estimating audience size — TV ratings — doesn’t count viewing across multiple screens, distributors and markets around the world. Measurement firms have been scrambling to fill in the gaps. The dominant player, Nielsen, plans to introduce a new total audience measurement for the U.S. early next year that includes online and mobile viewing. It also has partnered with Twitter to develop a separate rating that reflects social media conversations about TV shows. Specialized research firms such as Fizziology plug into Twitter, Facebook, Tumblr, Instagram and blogs to give Hollywood studios insights into online conversations about movies.

Parrot takes a different — and, it argues, more comprehensive — approach to evaluating interest in TV shows in markets around the globe. It creates a measurement called a “demand rating” that reflects interest in a TV show as expressed across photo-sharing sites like Instagram, online video sites like YouTube, social media platforms like Facebook, file-sharing sites and fan and critic blogs.

“If I want to express my demand for a piece of content, say, ‘House of Cards,’ I can stream it on Netflix or I can watch clips on YouTube or [post comments] to microblogging sites like Reddit, where 200 million people discuss TV content,” said Parrot Chief Executive Wared Seger. “You look at all of this and essentially you now have a truly ubiquitous measure that tells you how much demand there is for a piece of content.”

Parrot’s technology, developed by a team of data scientists and entertainment executives pulled from Sony Pictures, MGM Studios, the MIT Media Lab and Pukeko Pictures, uses pattern identification and contextual techniques to synthesize petabytes of data from 249 countries into meaningful information. The technology weighs viewer sentiment, evaluating just how obsessed people are with a show (“Liking” “Orange Is the New Black” on Facebook is less of a sign of true fandom than blogging about it).

“Not all fans are equal,” Seger said. “Some will talk about it, advocate for it. Others will be passive consumers who drop off after the third episode. Our demand metric takes that into account.”

Click here to read more.

SOURCE: Re/code, Dawn Chmielewski