Over the last few years we’ve seen the Internet community applying social network analytics to solve a growing number of consumer problems; from social search to influence analysis to social recommenders — the big boys, such as Facebook, Google, and Amazon, and a plethora of niche vendors, such as Klout, Peerindex, Topsy, Kred, etc. On the other hand we have seen a paltry adoption of social network analytics within the enterprise. What makes this particularly frustrating — hence the rant! — is that the enterprise is the perfect environment for such analysis.
Unlike the Internet which is an open domain, that is semantically ambiguous, and where identity resolution is a sometimes unsolvable problem; the enterprise is a closed, relatively unambiguous environment, with just oodles of context. Now none of this is rocket science, so why the rant?
Over the last year I have been looking at some alternatives to traditional enterprise search, specifically looking at social recommenders. The more I compare these approaches the more convincing social network analytics becomes. Let’s just face it enterprise search sucks! Ok, that may be a bit harsh but I’ve yet to meet anyone who honestly believes that enterprise search is better than Internet search. And that is me being polite, normally the response to the question is much more colorful. One of the major reasons it sucks is because of the lack of link analysis and the high volumes of traffic that combine so beautifully in the Internet scenario. However, all the reasons that make Enterprise Search Suck! make Enterprise Social Recommenders Rock!
- The diversity of content sources, structured data, and business processes, provide a massively rich set of data for providing highly targeted and business relevant recommendations.
- The disambiguation challenge, which is a significant problem on the Internet, is generally not an insurmountable issue in the enterprise; firstly people tend to have consistent userids across applications, and secondly terminology if not totally consistent is generally relatively well understood.
- Its not creepy, as we are just tracking boring work stuff and not what you had for breakfast, where you get your morning coffee, or what you spent on your credit card last month.
- The objective within the enterprise is to help you be more productive, and not to exploit you by bombarding you with products you don’t want or need.
If we could just start to use the disparate data that is languishing across our corporate databases, content repositories, and business applications, to feed our enterprise social semantic networks, then we could completely transform the face of knowledge acquisition and sharing. We are missing a huge opportunity!!!