3 Emerging Trends in Social Network Analysis


Courtesy of Krishnan Parasuramari and IMEDIAConnection,

A (relatively unsuccessful) Facebook boycott brought about by privacy concerns put the social network in national headlines earlier this year. Just last month, Harrisburg University banned Facebook and Twitter for a week. And this month, Hollywood released a movie on the social network.

Americans are spending nearly a quarter of their time on social networking sites, up from 15.8 percent just a year ago. Many concerned citizens are wondering exactly how dependent we are as a society on social networks, while businesses are busy trying to figure out how to take advantage of the trend, and most recently, how to measure, collect, and react to conversations on social media.

Get connected. Want to meet up with the companies that are leading social media into the future? Check out the exhibit hall at ad:tech New York, Nov. 3-4. Learn more.

Changing the world one tweet at a time
Speaking at the recent ExactTarget Connections event, Sir Richard Branson indicated that his organization strongly believes that social media conversations can change the world one moment at a time, one person at a time. For example, one of his Virgin America passengers, who had been waiting a while for his in-flight food order, decided to voice his concern through social media. Using the Wi-Fi capability on board, he decided to tweet his problem and provided specific details, including his flight and seat numbers.

Virgin’s ground crew, which had been monitoring the company’s Twitter feeds, noticed the concern from the food-deprived passenger and immediately relayed the message to the air crew, who promptly made sure that the needs of the passenger were taken care of. Is this the nature of our conversations of the future? What happened to the good old means of pressing the flight attendant call button when you need something?

The social media bias
In their attempt to be extremely proactive in playing in the social media space and “participating in customer conversations,” are organizations creating a bias toward those who have exclusive access to the megaphone? Who exactly is the person behind the persistent angry (or occasionally happy) tweets, blog comments, or Facebook updates?

Even though the intent of participating and listening through social media might be noble, often times organizations tend to view this medium as a disconnected channel and create another information silo. Rather than view their customers holistically as individuals who engage with their brand, they tend to view them as a visitor on their websites, or a friend on their Facebook pages, or a follower on their Twitter accounts, or a user on their mobile apps.

Until we are able to connect our social network measurement and analytics strategy to our existing CRM or marketing analytics strategy, we will not have the kind of unified view of our customers and influencers that will ensure we’re efficiently reaching — and responding to — the right people, with or without the megaphone.

Identifying the right set of strategies, tools, and processes that pertain to monitoring, measurement, and analytics of social networks can be confusing and even intimidating. However, as the market matures, we see the emergence of three common social network measurement objectives and techniques:

1. Measuring brand health via social media analytics
Brand health helps an organization gauge overall brand awareness, impact of a specific marketing campaign, competitive advantage, share of voice, and customer satisfaction. Recent studies have shown that 51 percent of active Twitter users follow companies, brands, or products on social networks. Most organizations have a fan page on Facebook. Both of these provide a platform for consumers to voice their opinion about brands.

James Kobielus defines the measurement of information sourced from social media such as Twitter and Facebook as social media analytics. The typical social media metrics that are indicative of brand health include a combination of engagement, sentiment, and activity metrics.

The types of tools that allow one to perform social media analytics range from real-time social media monitoring tools such as PostRank to “listening platforms” that enable the shaping of a social media marketing strategy. Alterian, Converseon, Nielsen, Radian6, Cymfony, and Visible Technologies are some of the leading vendors providing this capability.

2. Improving targeting precision by enriching customer data with social information
Marketing organizations and service providers are constantly striving to improve their targeting precision. Having more information about their customers enables them to direct specific, relevant, and targeted messages at them. Social network information acts as another channel from which one can extract an individual’s information that has been voluntarily made public such as status updates, recommendations, connections, friends, interests, etc. Using data mining techniques on this data, one can predict an individual’s responsiveness to campaigns, risk profile, influence across their personal social network, etc.

In a discipline that is being termed as social media profiling, a new crop of companies have emerged that glean social network data from across the web and create composite user profiles that can be enriched with traditional customer data for marketing activities such as segmentation and targeting. Rapleaf, 33across, media6degrees, and XGraph are leaders in the social media profiling space.

Traditional customer data companies and marketing service providers partner with social media profiling companies to create a robust customer profile that includes demographic, geographic, psychographic, behavioral, and social information about their customer or prospect. For example, using social network information for targeting enabled Sprint to promote the launch of the Palm Pre smartphone and quadruple related online sales.

3. Understanding customer influence through social graph analysis
Within any given social network, individuals are either creating relationships with one another or engaging with content or a community.  Individuals are creating relationships with one another by making them their friend or following them or recommending them. Individuals engage with content by expressing their opinion about that content (like, dislike, rating, etc.). Mathematically, this behavior can be expressed as a graph where nodes represent an individual, and a tie (or edge) represents a relationship (or engagement). The resulting social graph can be analyzed to detect social behavior patterns, emerging trends, and influencers.

According to a recent report that studied travel and hospitality marketers, 28 percent of email marketers are analyzing the social influence of their subscribers and using that to target their email campaigns.

Telecommunications companies have been successfully using social graph analysis to predict churn. Their customer lifetime value models incorporate a subscriber’s ability to influence others within their network, and they use this information in retention marketing efforts.

Social graph analysis can be a complex computational problem to solve. Organizations either use pre-packaged influencer marketing tools such as Pursway, or build their own using special purpose graph databases and network visualization tools.

These emerging trends and techniques clearly indicate that there are multiple nuances to analyzing and gleaning insights from social networks. As the market matures and organizations have a clearer understanding of how they would want to leverage social network data, a convergence of these approaches is very likely. The traditional architecture of a customer master database feeding into an enterprise data warehouse with customer purchase transactions fueling operational and analytical reporting will be augmented with social network data.

The customer master data management system will have hooks to third-party data sources that provide clean, up-to-date sets of “social attributes” to enrich the customer profile. Likewise, enterprise data warehousing platforms will have hooks to social network events and streams that can be assimilated along with other customer activities. Reporting and analytics tools that facilitate conversion of social graphs to relational constructs coupled with scalable analysis on social graphs will enable use of this information in conjunction with other types of data.

Once we achieve these goals, social network analysis will cease to exist as a stand-alone discipline. It will just be part of every firm’s CRM or marketing analytics strategy.

Krishnan Parasuraman is chief architect, digital media, Netezza.

About these ads

Comments are closed.