Tech Tracks Social Media Crowd Reactions to TV

Bluefin visualization of real-time social media response levels to TV content.
Bluefin visualization of real-time social media response levels to TV content. (Image credit: Bluefin Labs)

Deb Roy recorded 90,000 hours worth of home videos and many more hours of audio to see how his baby son learned to speak in the first three years of life. That spawned a business idea when the MIT professor realized that the computer program that analyzed his son's early words could also analyze words found in TV broadcasts and social media.

Now Roy is founder and CEO of a startup that aims to map audience responses to mass media content almost instantly. Taking a cue from the language learning experiment, Bluefin Labs' technology relies upon cognitive psychology and artificial intelligence research to see how people chatting on social networks and blogs react to the ads and TV content flashing before their eyes.

"Once you have the basic audience mapping, you can analyze the properties of the social networks, how people connected to each other are talking about the TV content, or understand the content — whether it drives more or less conversation, or positive conversation," Roy said.

The program first evolved by learning to match the words of sportscasters with Major League Baseball action being shown on TV. Understanding words with the video context allowed the program to tell the difference during a search for the word "home run" between different scenarios, such as when an announcer says, "They sure needed a home run" as opposed to, "It's a home run!"

That allowed the MIT researchers to create a searchable video index with funding from the National Science Foundation's small business innovation research program. But the technology's potential truly became clear when it began sifting through 3 billion comments per month from blogs, tweets and Facebook posts.

"We're not getting play-by-play commentary from one sports announcer, but from thousands of people," Roy told InnovationNewsDaily.

The technology's fine-tuning has reached the point where it can show whether audiences respond positively or negatively to each ad or sports play, as well as movie trailers and TV show episodes. It can also track how much of the audience comments on each moment of TV consumption, and show the links between each comment and the content that drives conversation.

Having such consumer information excites everyone from the creative talents to the TV networks and Hollywood studios, as well as the marketers who run ads against that content. Bluefin Labs has already attracted more than $8 million in venture capital funding, and has been running tests with Fortune 500 companies and agencies.

"[The technology] is in operation with 10 paying customers in the offices of multiple global brands," Roy said. "We completed the first pilot with one of the world's biggest advertisers, and [our contract] has been renewed."

The company first came out of stealth to announce its technology platform on Feb. 2, 2011. Roy co-founded Bluefin Labs with one of his Ph.D. students, Michael Fleischman.

This story was provided by InnovationNewsDaily, a sister site to LiveScience.

Jeremy Hsu
Jeremy has written for publications such as Popular Science, Scientific American Mind and Reader's Digest Asia. He obtained his masters degree in science journalism from New York University, and completed his undergraduate education in the history and sociology of science at the University of Pennsylvania.