Trying to recognize actionable intelligence from the countless conversations taking place in the social web can be a bit like trying to find a needle in the proverbial haystack, particularly if your customers use multiple languages. Every day, the social web grows in size and influence with consumers. Facebook and Twitter have millions of users, which means millions of opportunities for companies to promote their products and gauge market trends. But trying to recognize actionable intelligence from the countless conversations taking place in the social web can be a bit like trying to find a needle in the proverbial haystack, particularly if your customers use multiple languages. English makes up only about 40 percent of global social media postings, so companies that ignore this language diversity are at greater risk of being caught off-guard by trends or developments that could impact their business.
Of course, there is a plethora of commercial search and analytic technologies that claim to be able to help marketers monitor social media, but these traditional analytics solutions are often ill-equipped to understand the unstructured information that lives on social networks.
Their keyword-driven approaches often fail to take account of the subtle nuances and meaning in a lot of user-generated content, and will struggle to understand slang expressions or emotions such as sarcasm. Add in the challenge of multilingual content and the shortcomings of using traditional search and monitoring tools become readily apparent.
Indeed, I suspect that these difficulties explain why Facebook, despite its huge resources, has not made social search a priority development area – or at least has not until now.
While this market is currently still in its infancy, I predict we will see a lot of interesting developments and tie-ups in the area of applying natural language technologies to social media analysis.
Facebook has yet to show its hand in this area, but once the dust settles on its IPO, I suspect Facebook’s “next big thing” will involve social search. While it presumably has no desire to compete with Google head on, a lot of advertisers would be very interested in gaining some insight into the millions of conversations taking place on Facebook.
On a much more modest scale, I was intrigued to read a recent announcement involving two US vendors, which gives a taste perhaps of what is to come.
NetBase, a supplier in enterprise social intelligence, has signed a deal with Basis Technology, which specializes in digital forensics and whose clients include the US Department of Defense and national intelligence services.
Readers of Tom Clancy novels are familiar with the hush-hush world of digital espionage in which operatives sit in darkened rooms monitoring the world’s communications channels for compromising messages. But perhaps this spy-book world is not too far from the truth as amongst NetBase’s application is one that handles “Arabic chat translation”.
NetBase is using Basis Technology’s linguistic analysis and entity extraction solutions to allow it to rapidly add multilingual capabilities to its monitoring product. NetBase argues that it is becoming increasingly important for its customers to access business and market insights hidden in non-English conversations.
In similar vein, Artificial Solutions already offers its Teneo Social Media solution to take some of the hard work out monitoring social media.
By searching through aggregated social media content, Teneo Social Media can understand the sentiment behind a comment made on a social media site and respond appropriately, whether by providing helpful information, a link to a website or a telephone number for a human contact.
While some people might balk at the idea of using “speech bots” in social media, research indicates that people respond positively to automated messages even on the social web — where person-to-person interactions are normally the rule — just as long as it is made clear that it is not a live agent and that the response is actionable advice.