Bluefin Labs

Bluefin Labs was founded by Deb Roy and Michael Fleischman in 2008.

The seed for the company originated within the MIT Media Lab's Human Speechome Project (HSP). Headed by Roy, head of the Media Lab's Cognitive Machines group, the goal of HSP was to understand language acquisition by making a complete digital record of the first years of a child's home life and analyzing that record computationally. Roy and his wife installed video equipment throughout their home and captured much of their son's life from birth to age three - a total of 240,000 hours of audio and video. Buried in this massive corpus lay insights into how their son learned language at a level of detail never before captured. To sift through the data, Roy and his MIT team developed numerous deep machine learning algorithms designed to uncover the context and patterns within the audio and video data. Their results are explained in Deb Roy's TEDTalk 'The birth of a word.'

As a student in Roy's MIT group, Fleischman adapted the concepts of data driven designed for language acquisition and applied them to another massive video data set: broadcast video. Fleischman's PhD work garnered the attention of the National Science Foundation which awarded Roy and Fleischman a Small Business Innovation Research grant and effectively bootstrapped the launch of Bluefin.

Just as HSP created algorithms that can “ground” the meaning of a word within a larger context, Bluefin is using deep machine learning to ground the meaning of a given comment pulled out from a social media stream. By looking at the context of words expressed by individuals, Bluefin can use the meaning of these words to connect comments back to the events, people, places, products, and brands that caused those words to be expressed in the first place.

On July 32, 2012, Bluefin Labs announced that advertising industry veteran JP Maheu will join the company as CEO.

On February 6, 2013, Twitter announced that it was buying Bluefin Labs for approximately $100 million.

Technology

Bluefin Labs uses proprietary algorithms to match social media comments to televised content. The methodology of associating comments to TV is composed of three parts: temporal analysis (looking at when a comment is posted), semantic analysis (matching written comments to what's on TV), and machine learning (the ongoing process of training and re-training the matching algorithm). Bluefin also employs a proprietary algorithm for sentiment and gender analysis of the social media comments.

The resulting data set is a large network of cause-effect links that measures people's engagement with TV, effectively creating a viewer/programmer feedback loop. This network is designed to provide marketers and media companies with insights that could be used for more efficient ad targeting and more informed content and creative development.

Bluefin calls this mapping of social media commentary back to its televised stimulus the "TV Genome" – a highly complex endeavor that, like the Human Genome Project, requires immense computational resources along with deep-rooted techniques in cognitive science and machine learning. Bluefin has a view into more than 5 billion public-facing social media comments each month, tied to a continuously growing video fingerprint archive of more than one million distinct airings of TV shows and commercials. Bluefin currently ingests and performs video fingerprinting of 115 U.S. TV broadcast and cable networks.

Products

Bluefin Signals is a social TV analytics suite that measures and organizes the conversations about TV that take place in social media. It is a SaaS analytics platform and is available to Brands, Agencies, and TV Networks on a subscription basis.