By using website you agree to our use of cookies as described in our cookie policy. Learn More

TDWI Upside - Where Data Means Business

In Social Media Analytics, Beware of "Social Contagion"

Marketing professor Andrew Stephen sees two primary problems with the way most companies handle social media analytics: a lack of data-driven assessment and social contagion.

When it comes to harnessing the power of social media analytics -- especially for fine-grained use in sales and marketing -- most companies are still operating in the dark ages.

Most Social Media Use Not Data-Driven

Andrew Stephen, L'Oréal Professor of Marketing and head of the marketing faculty with Oxford University's Saïd Business School, says most companies don't take a data-driven approach to social media analytics. He likens this to throwing darts at a dartboard -- blindfolded.

We can and must do better, Stephen told attendees at Teradata's 2016 Partners Conference, held in September in Atlanta. "Brands often don't use sophisticated analytics approaches to understand content performance or to improve content design and planning. Instead, they're ... throwing darts blindfolded and just seeing what happens."

"That's great if you see what you're doing as experimentation, but the point of an experiment is to test and learn. It's not clear this is happening in a data-driven way."

The Two-Fold Problem

Stephen sees two primary problems. The first is a critical lack of robust and data-driven methods for assessing and understanding the performance of owned social media content.

"In my experience, [the tools aren't] very robust in terms of how they assess that content performance [of social media analytics] and try to link it to relevant outcomes. They're collecting a lot of data, [but they're] not really using it smartly," he said.

The second problem is what he calls social contagion: when people connected with one another in a social network are unconsciously biased by the opinions or behaviors of their peers.

"Social contagion can make customers' social actions non-independent, interdependent, or biased. On the one hand, social contagion [can have an effect such that] everyone [in a network] ratchets up and gets better; on the other hand, it can [manifest as] bad ideas spurring more bad ideas."

Defining Social Contagion

Social contagion is a particularly compelling concept. Stephen used a real-world example to illustrate its effect: Starbucks asking its customers for promotional ideas via Facebook.

"Someone starts a thread [in response] saying: 'You should get a free drink at Starbucks on your birthday.' The next post, less than an hour later, someone says 'Let's have a [Starbucks] Birthday Club, or some other birthday[-themed] thing.' Someone else chimes in and says, '[Let's have] a Starbucks card that says Happy Birthday.' Then someone else says, 'What if we could evite our friends to Starbucks for our birthday?' At some point, an employee chimes in and says, 'At Starbucks, we say happy birthday every time we brew coffee.' This is social contagion taking place."

The initial post yielded an idea Starbucks could consider, analyze, and possibly act on. Social contagion crept into the ensuing posts as a kind of self-reinforcing redundancy. It's an echo chamber-like effect whereby subsequent contributors offer refinements on an initial idea -- or, just as likely, take a thread way off-topic.

"It's good on one hand, [but] on the other it's self-reinforcing," Stephen pointed out, explaining that the company would probably rather have more unique ideas. "Redundancy could also be a problem in terms of the creativity of these ideas, [the] innovation of these ideas."

Prevent Clustering to Prevent Contagion

This doesn't have to be a show-stopping problem. In an article published in the Harvard Business Review, Stephen and several collaborators presented a simple solution: limit clusters so people can't see what everyone else is saying. "People offer better ideas when they can't see what others suggest," he said.

"Connection in clusters might be bad for innovation. If I'm seeing the same people talking about the same stuff, we'll [all] end up taking about the same stuff. If I'm exposed to a bunch of ... influences all saying the same thing, I don't have different perspectives inspiring me to create my own [ideas]."

In response to a question from an attendee, Stephen offered up one more tantalizing nugget: there's an optimal size or density to a social media cluster. When there are too many followers in a cluster, they don't generate as much interaction or feedback. The trick is to figure out what that optimal density is.

"There is an optimal density, but remember, [as] you're going down in density, you get to a point where essentially people don't pay attention to a network because no one's connected to anyone," he said. "We [used a metric from] 0 to 1: 1 [is when] everyone's connected to everyone else, 0 is when nobody's connected to anybody. We went down to 0.1 and we still saw benefits to lowering [the cluster size]. Beyond that, it sort of bottoms out.

"The idea is to keep the network connected enough so people can use it for ideas [or] inspiration."

About the Author

Stephen Swoyer is a technology writer with 20 years of experience. His writing has focused on business intelligence, data warehousing, and analytics for almost 15 years. Swoyer has an abiding interest in tech, but he’s particularly intrigued by the thorny people and process problems technology vendors never, ever want to talk about. You can contact him at [email protected].

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, and Team memberships available.