Social Network Analysis: Going to Become Too Ubiquitous for Its Own Good

[Latest News Headlines]

ByJames Kobielus


Socialnetworks are the future of online life, whether we like it or not. Before the endof the coming decade, relationships with everyone –including family, friends, colleagues,employers, merchants, suppliers, and government agencies—will hinge on your accessto these parties, and they to you, through online communities of all shapes andsizes.


Socialnetworks are becoming much more pervasive than today’s mass-market communities—suchas Facebook, Twitter, and LinkedIn—would lead you to believe. Before long, manywill be embedded services in the full range of business and personal applications.In ten years’ time, today’s social networks will have evolved into a powerful,seamless worldwide infrastructure for collaboration, sharing, interaction, andtransactions. Many will be integral features of the mobile, broadband, and streamingmedia services that shape business and consumer life everywhere. Many will be secure,robust environments that span across federated public and private clouds.


Froman enterprise perspective, one of the most important feature of social networksis that they are literally the “buzz” that can spell the difference betweensuccess and failure in a reputation-driven online economy. Already, we see astrong interest in social network monitoring and marketing tools. Everybodywants to know whether, how, how often, and by whom they’re being mentioned inTwitter, Facebook, blogs, and the like. Even more critically, everybody wantsto influence those discussions in their favor and extract maximum revenue potentialfrom sales to people and companies that use those media. Anybody who has evertweeted the name of a large company and been immediately greeted by anautomated “we’re following you” message from that same company knows the power—andpotential nuisance factor—of this new medium.


Clearly,social networks are highly monetizable. At heart, they are continuously refreshingstreams of customer-generated intelligence on requirements, needs, sentiments, andexperiences. As this new way of doing business gains traction, companies will focusthe full power of advanced analytics on social networks, especially those wherecustomers live. Forrester sees growing enterprise adoption of social networkanalysis, which is an emerging discipline of predictive analytics that looksfor behavioral, attititudinal, and other affinities among individuals. Thoughit can be applied to a broad range of scientific and other topics that involveno online interactions, social network analysis thrives on the deepening poolsof information—structured and unstructured, user-generated and automated—that streamsfrom Facebook, Twitter, and other new Web 2.0 communities.


Whatexactly is social network analysis? Essentially, it involves discovering, mapping, and measuring relationships among people,groups, companies, and any other entities—including products, online content,and personal computers—with which they interact. It leverages such key data miningcapabilities as segmentation, clustering, and regression to identify, forexample, who are the leaders, followers, influencers, and outliers in socialgrouping. If viewed in the time dimension, social network analysis can, for example,reveal the cultural dynamics that spell the difference between as successfulmarketing campaign and a flop, or that drive one group of customers to renewtheir contracts and others to jump to the competition.


It doesn’t take much stretch of the imagination to see wherethis all might lead. As customers reveal their candid thoughts in real-time viaTwitter and other social networks, enterprises can conceivably cut back on structuredsurveys, focus groups, and other traditional approaches to gauging demand.Instead, companies can simply and automatically “listen” to social networksthrough complex event processing; process the unstructured text streams throughcontent analytics; aggregate all the intelligence into massive analytical datamarts; and drive their sales, marketing, and customer service through inlinesocial network analytics models. The “killer app” for all this becomes the real-time“next best offer” your contact center makes from this intelligence, or themarketing campaign you re-arrange on the fly to save it from near-failure.


But there’s danger in this otherwise promising scenario. Theproblem is that social network analysis—automatic, real-time, effective--will becometoo popular. Enterprises will rely more and more on high-powered social-network-analysismodels to divine market trends, segment the customer base, and tune their campaigns.As this approach gains adoption, companies will be tempted to pull backsignificantly from traditional outreach efforts—such as opinion surveys andfocus groups—that involve direct interaction with customers and otherstakeholders.


It’s not that statistical models are inherently invalid. It’sjust that they can become another layer of abstraction between you and yourcustomers, preventing you from truly making personal connections with them andthereby securing their ongoing patronage and loyalty. And those models maycrowd out the intuitional approaches you should also use to gauge the cultural zeitgeistwithin which your business creates value.


To the extent that social network analysis encourages proliferationof superficial customer-segmentation categories—on the order of “Yuppie” and “GenerationX”—it will invite backlash. One might refer to these sorts of magazine-ready segmentationlabels as “statgeists,” meaning that they may have some grounding in statisticalmodels of demographics and other metrics, but are little more than trendyconsultant-speak.


When you use these and similar verbal concoctions, you gag alittle bit, as do the real people to whom they ostensibly apply.



More from the Forrester Business Process & Applications Blog...