The Data Mining of Social Networking

The data mining of social media activity is now commonplace in business intelligence circles. Over the courseof the last 18 months, this stream of BI has experienced a startling rate of growth and reached lofty levels ofsophistication. Just a few short years ago, consumer-oriented businesses were stuck in the world of static“focus groups” and paper-based surveys. But not even the most forward-looking of these organizations couldhave dreamed of the present-day scenario, where newly forged nuggets of data about consumer behavior andpreferences wait to be mined by state-of- the-art BI computing infrastructure. Even the U.S Government, whichis usually late to any technology trend, is looking to get in on the action as well. The Obama administration hasexpressed a keen interest in performing data mining against content posted on the White House’s Facebook,YouTube, and Twitter pages.Anything and everything on the Internet is fair game for extreme data mining practices. Once something ispushed out to the World Wide Web, it will forever be fodder for a business intelligence or data miningapplication somewhere in the cyberspace universe. But despite our best judgment about the pictures weupload to Facebook or the posts we make on MySpace and YouTube, in the heat of the moment our prudenceoften fails us.

This is because social networking is, to a large part, about vanity: We want all of our friends andcolleagues to see how cool and fulfilling our life is: what we had for dinner; what fancy places we have visited;how our dogs made a mess on the living room carpet. Vanity (and electronically enabled narcissism) willforever take the form of impulsiveness, which is not a good character trait when posting under one’s own name,especially when one’s post may wind up being in the public domain for a long time.For many social media sites, the Terms of Service (TOS) are explicitly clear and to the point: If you postcontent to the site you essentially grant the site permission to use the content for any purpose they deemappropriate. While each site is different in their irrevocable and perpetual right to reproduce the informationfound in your posts, it is wise to err on the side of caution.

No matter how private you deem the content, privacycontrols usually only go so far - the demarcation between private and public information remains fuzzy at best.In some cases, once you submit content it may instantaneously become the intellectual property of socialnetworking site, even if you delete or purge the submission in its entirety. Speaking from personal experience,the array of privacy options on my favorite social media sites is quite robust; nevertheless, these privacycontrols never fail to confuse me every time I tinker with them, leaving me unsure about how my sensitiveinformation and opinions will be used.

Users of the world’s most popular social networking site remain unaware that Facebook recently launched anextremely controversial Instant Personalization feature that essentially “transfers” a registered member’s profiledata from Facebook to external third-party websites. What is alarming is that Facebook lends out this datawithout any sort of opt-in permission on the part of the member. When the member visits one of Facebook’spartner sites (such as Yelp), the site will automatically reveal (at the bottom of the page) which of the member’sFacebook friends have reviewed a similar business in their geographical area. I was startled the first time Istumbled on this. And when the initial shock wore off, I was more than a little offended and less trustful ofFacebook. Not only was my information going to third-party affiliate sites, my list of friends and publicallyavailable information about those friends was tagging along for the ride.

After some detective work, Idiscovered that there is a way to opt-out of this personalization feature, thank goodness! But what irks me isthat Facebook does precious little to make their members aware of this feature (and how to turn it off) in thefirst place.Taking this functionality further, a user of Facebook may have their posts show up in advertisements that arecustomized and targeted to their friends as part of a “Sponsored Story.” This story feature gives advertisers theauthority to republish and repackage a user’s post that references the advertiser’s products or services. Forexample: If a friend “checks into” a store or becomes a fan of a brand’s Facebook page, this information canappear on the right side of a user’s Facebook pages as a virtual billboard, for which an advertiser will, no doubt,pay handsomely. (Thankfully, the Sponsored Stories do not show up in a user’s main news feed; however thereis currently there is no way for users to turn off this feature and keep their tastes and preferencescompartmentalized and the viewership of these more controlled).The purpose of these social media cautionary tales is not to single out and beat up on Facebook; however, itsunparalleled financial impact (1.5 billion dollars in advertising revenue) and size make it the most relevantexample. (As of this writing there are over 500 million registered users of Facebook, giving it a population largerthan almost every country in the world.)

In the social networking world and in the Web 3.0 paradigm in general,innovation often comes at a cost to privacy. This bodes well for BI companies that are focused on data miningthe information on Facebook, Twitter, Linked In, My Space, etc, although it is confounding for those of us thatuse social networks on a regular basis and live by the mantra: “What happens in the Network, stays in theNetwork.”While the staying power of business intelligence and consumer data prospecting has been tested year in andyear out for decades now (BI’s importance and utility increases with each passing year); the long term viabilityof some of the world’s most popular social media sites is not a sure thing.

As more and more social networkingsites come online the competition for registered (and active) users will escalate. Not only will social networksneed to do more to procure new users and retain current ones, they will have to create new and novel ways tomonetize their websites. For better or worse, supplying content created by a social network’s users to outsidewebsites, advertisers, and affiliates for data mining purposes is a sure fire way to generate revenue for theforeseeable future. It will remain an important component of the social networking business model, as datamining methodologies progress far beyond traditional demographic profiling into interpolation and statisticalmodeling based on swarms and cluster groups.

So powerful and lucrative is social media data mining thatgovernments around the globe have begun to carefully scrutinize the need for regulation in this space,especially with respect to protecting the privacy of their citizen’s that post data to these networks. While someregulation is probably needed, what concerns me is when the legislators of the free world start demandingsocial networks involuntarily hand over their user-generated content in order to better enable centralgovernments to carry out their own “citizen intelligence” and data mining programs. That day may be closerthan any of us care to realize.