February 6, 2014

Why IBM Watson Knows You Better Than You Think

Tiffany Trader

Last week IBM launched a new global consulting practice to help clients take advantage of data-driven insights and technology platforms to drive deeper personalization. The new unit, called IBM Interactive Experience, fuses the design capabilities and user experience experts from IBM Interactive – the company’s digital agency – with the innovation and data expertise of more than 100 IBM researchers from its Customer Experience Lab.

As part of the announcement, IBM revealed a set of data-analytics tools based on the linguistic capabilities of the Jeopardy-winning Watson platform. Called Life Event Detection and Psycholinguistic Analytics, these tools employ algorithms that combine the psychology of language with social media data to shed light on an individual’s preferences and personality traits.

The Life Event Detection tool is so called because it can analyze specific “life events” like a marriage or divorce and then extrapolate that to buying behavior, while Psycholinguistic Analytics zeros in on more personal qualities. The tools can predict which services or products a customer will be most interested in and the best way to get a message across.

In this video from IBM Research, Research Staff Member Eben Habor explains how this Watson-derived technology helps business understand their customers in new ways using terabytes of public social media data. “Business have always wanted to serve their customers as individuals and social media analytics provides the opportunity to do this at scale,” observes Haber.

 

“We can generate millions of profiles from billions of tweets, identifying tweets such as purchase intent, demographics and interests,” he says. “Not only that we can dive deeper to understand more about who this person really is, identifying traits such as personality, needs and values, and even schedules.”

“We can infer these traits thanks to decades of psycholinguistic research which have found correlations between traits such as these and an individual’s language and activity patterns,” he adds. “This is data that marketers have never had access to before and it could be extremely valuable to them because research has also shown that traits such as personality are highly-influential over brand preferences and buying decisions. This could permit much more refined marketing than traditional approaches based on demographics and purchase history alone.”

The technology can do all of this without having official access to customers’ Twitter accounts. There are enough clues left on publicly visible social media accounts for the technology to identify the real-life person with a high degree of confidence. Those who are averse to having their social media data mined in this way can change their social media privacy settings. IBM says it is also working to institute a more global opt-out feature.

Related Items:

Watson as a Service Becomes a Reality

New Initiative Aims to Expand Watson’s Reach 

Rometty: The Third Era and How to Win the Future