The next generation business strategies will be heavily dependent on the information generated by the social media demography of the netizens. The doubt is only about the capability of present business models to leverage the potentials of this data. Best of our efforts are in the direction of analyzing the 'Big Data' generated by the multitude of social media networks. We constantly try to devise an algorithm that can approximate the behavioral pattern of the data generation in these networks. But it is high time to realize that the approximations and content matching algorithms involves lot of wastage of computational logic and infrastructure.
Instead of depending on the predictive computational models, it is time to devise difference engines. Our prediction filters should be replaced by difference engines. The more difference we have in the model and the actual data, the more information we are receiving from the social media networks. For example, if we have a template for a user profile in Facebook based on his activities for the last one month, if we get totally surprising data on the next day, we should welcome it with all happiness and curiosity. This difference should be analyzed and integrated back to the template. This feedback mechanism will generate more and more non-linear data about this user.
A further interesting point for the difference engines would be to locate the reasons of the increasing complexity of social networks. The first question is, if at all the complexity is increasing. If it is the case, what is the reason, is it caused by the accumulation of data, or is is due to the increase of connections between the nodes in the social network graphs.
Wikipedia presents one defintion of social media as this:
"In general, social networks are self-organizing, emergent, and complex, such that a globally coherent pattern appears from the local interaction of the elements that make up the system."
But are social networks self-organizing. Is data produced and consumed in social networks, self organized?
Note:: This post is a draft in progress ...