One of the most valuable things about social networking is all the data that is gathered, especially about potential customers. The trouble from an IT perspective is that, until recently, it's been generally cost-prohibitive to first gather all that data and then analyze it. The end result is that companies such as Facebook, LinkedIn and Google wind up knowing more about what people think of your company than you do.
But Rod Smith, vice president of emerging technology for the IBM Software Group, notes that with the advent of various approaches to gathering Big Data, it's become a whole lot easier for companies to gather all the information and then analyze it themselves. That not only reduces dependency on the small number of companies that today dominate the social networking space, it allows companies to be smarter about how to respond to various events involving them, their customers, suppliers, partners or competitors.
The next challenge IT organizations will face in terms of making the information actionable, says Smith, is applying analytics to the trove of data they can now more easily collect. To accomplish that, Smith says IT organizations should start moving to organize that information around metadata constructs that will make it easier to identify trends and correlate disparate sources of information. In other words, the whole concept of sentiment analysis as applied to social networking is about to get a whole lot more sophisticated. For example, IBM is working with the Annenberg School of Communications and Journalism at the University of Southern California (USC) to apply analytics to social networks in order to determine how well a particular movie might do at the box office.
What will help make that more practical to accomplish is that with the advent of inexpensive memory and storage technologies, there's no need to spend days and weeks modeling data that normally would be obsolete by the time the IT department came up with the answer. Instead, all the queries can be handled in real time against all the raw data, which obviously in time will require the ability to routinely work with petabytes of data. Longer term, as the system learns, Smith notes that the system will eventually learn how to help shape the actual queries.
We're rapidly approaching a point where the need to make a decision based on instinct and experience without any data to collaborate that decision is becoming obsolete. That doesn't mean that gut decisions won't continue to play a role in making those decisions, but it does mean that a lot of the guesswork that goes into running the business today is about to get sharply reduced.