Twitter has acquired social TV analytics firm Bluefin Labs for an undisclosed fee, according to Business Insider. Bluefin labs grew out provides social analytics solutions to brand advertisers, agencies and TV networks. The company offers a technological, data and professional services linking social media commentary to the shows and ads watched on TV.
The acquisition will Twitter’s second screen offering as the company aims to position itself as an essential advertising and analytical bolt-on service for every TV advertising campaign. In December, Twitter joined forces with Nielsen to create the ‘Twitter TV Rating’, but the Bluefin Labs acquisition will provide them tech capable of looking at the entire social TV landscape (it’s going to be interesting to see if Twitter will be willing to provide data on the social performance of ads on platforms other than its own?).
‘Breaking the Semantic Barrier’
While Twitter has access to no shortage of data, it’s not easy to make sense of it. For the most part the data is qualitative, so context matters. It’s harder to sort through and categorise than, say, Facebook’s data, where the user provides their own data on their interests, their relationship status etc. On Twitter, for example, if someone is watching Top Gear and tweets, ‘Wow, I really do hate Jeremy Clarkson’s jumper’, you need technology that can properly pick that sentence apart and interpret it correctly.
A simplistic computerised system might be able to ingest data and pick out words in an attempt to link them to sentiment, but the result would in all likelihood lead to a muddled analysis e.g. when taken out of context, the word ‘wow’ could suggest they like the programme or Jeremy Clarkson; ‘hate’ could mean they disliked the programme or Jeremy Clarkson’s personality (has been known to happen); when in fact all the person was referring to was Jeremy Clarkson’s jumper. When the human brain would be well-equipped to understand the tweet, it’s far more difficult for a computer. Solving that problem and grounding the language in its true context is what Bluefin call ‘breaking the semantic barrier’.
The Bluefin Backstory
The Bluefin Labs origin story is a fascinating one. The company grew out of the MIT Media Lab and claims to have made the largest investment in social TV analytics innovation and R&D. The company’s CEO, Deb Roy, was director of the MIT Media Lab’s Cognitive Machines Group,which was originally working to understand human language development.
The group set up the ‘Human Speechome Project’, which aimed to make a complete digital record of the first few years of a child’s life and to analyse that recording to uncover the relationships between spoken language and context. Roy and his wife recorded ver 240,000 total hours of audio and video, capturing a huge chunk of their son’s life from birth to age three. The data was then analysed in the MIT lab using various deep machine learning algorithms that the team developed for the project.
When Michael Fleischman, Bluefin’s President and CTO, joined the Cognitive Machines Group, he adapted the analysis concepts originally designed for child language acquisition to broadcast video. The value of Fleischman’s research was recognised by the National Science Foundation, which awarded Bluefin a research grant enabling the company to launch.
Finally, here’s a video of a fascinating talk that CEO Deb Roy gave to Ted about the original Human Speechome Project: