Skip to content

How Social Network Analysis Solves Real World Problems

2011 August 3

I’m LinkedIn.  I’ve got friends on Facebook.  I tweet.  Yo, I got stooopid Klout!  Look at me!  I’m connected!
And so are you and lots of other things, like ecosystems, molecules, our bodies’ metabolisms, the list goes on. Quantum physicists believe that all matter and energy is linked in strange ways and create mind-numbing field equations to explain how.

We in the marketing world often fall into the trap of thinking about social networks purely in terms of social media and miss the big picture.  Network science is explaining large swaths of mysterious phenomena that have eluded us for decades, even centuries.  That’s good news for marketers. These new insights are about to revolutionize our field.

How Do Our Hearts Beat?

About once a second, millions of pacemaker cells in your heart must sync up so that the muscle can contract in an orderly fashion.  If they fail to do so even once, you have big problems.  When we hear of perfectly healthy people dropping dead of a heart attack, it’s usually the result of pacemaker cells failing to sync.

So how do they organize themselves?  Is there a “conductor” pacemaker cell?  Who leads, who follows?  It turns out that pacemaker cells exhibit a behavior called coupled oscillation, first discovered by Christiaan Huygens, that manifests itself in things as diverse as lasers, fireflies and crickets.

It was through studying the coupled oscillation phenomenon that Watt’s and Strogatz discovered the basic principles of network theory and published their landmark paper.  It turns out that things like pacemaker cells do not need a leader, but rather form small world networks and organize themselves.

Why are Some Broadway Plays Successful and Some Not?

Okay, so you have some extra money lying around and you want to invest in a Broadway play.  How would you predict which one was most likely to succeed?  Past performance of the director?  The production budget?  The marketing budget?

As I wrote in a previous post, a study done by Brian Uzzi and Jarrett Spiro determined that in actuality it is social networks.  They found that the most important factor was the structure of the links connecting people working on the play.

When people with no previous ties start working together, the results tend to be poor, they’re not familiar enough with each other to develop a strong working process.  Teams who have some familiarity with each other fare much better.

However, past a certain point they get to know each other too well and creativity, along with financial performance, suffers. Networks that are too loose don’t allow information to flow, but ones that are too tight can keep new ideas out and squelch innovation.

Who led the 9-11 Attacks?

Within a day of the 9-11 attacks, the leadership structure of the hijackers was published in major newspapers.  It’s not like Al Queda released a organization chart available, so how did they figure it out?

Well, the answer is, of course, secret and classified, but network scientist Valdis Krebs performed his own social network analysis by piecing together links between the hijackers from news reports and calculating three measures of influence:

Degree Centrality:  How many links each hijacker had to the rest of the network.

Betweenness Centrality:  Their location in the network relative to other members.

Closeness Centrality:  The average social distance between a particular member and all of the other members of the network.

This is what he came up with:

You’ll notice that Mohamed Atta led by a wide margin in all three measures, so it’s not a surprise that he is considered the leader of the network.  It’s also interesting that Zacarias Moussoui, the reputed “20th hijacker,” figures in so prominently.

For a more complete picture, check out Valdis Krebs’ full summary here.  It’s a fascinating read.

How Do You Stop a Flu Outbreak on a College Campus?

We did not evolve to live in close proximity to each other, but nevertheless find ourselves crammed together in public transportation, schools and workplaces.  The result is that once a new infection takes hold it spreads quickly through networks of people (or computers for that matter).

The key to preventing a runaway epidemic is early detection, but how do you identify individuals who are most likely to spread germs to other people?  Christakis and Fowler tackled exactly that question in a recent paper based on a study of 744 Harvard undergraduates.

They took an innovative approach.  Instead of thinking up traits that would mark people as “connectors,” they simply took an small random sample and then took another random sample of the first panel’s friends.

They found that the friends network caught the flu earlier than the initial sample.  In other words, the friends were more central to the network.  This is called the friendship paradox.  People’s friends tend to be more influential than they are.

Why?  Think about the TV show “The Sopranos.”  If you were to ask all the characters to name a friend, they would certainly be no less likely to mention Tony Soprano, who plays the most central role.  That simple quirk in the math shifts makes the friends network more influential.

How Did Justin Bieber Get So Popular?

As Daniel Rigney shows in The Matthew Effect, “the rich get richer” is more than just an old adage, but a measurable reality.  In areas as diverse as sports, business and scientific research, a small advantage tends to grow into a large one.

How then do we account for Justin Bieber?  A boy of modest means starts uploading videos to YouTube and becomes a sensation.  Was he just amazingly talented that fame and fortune were inevitable or was he just lucky?

As I explained in an earlier post, social network analysis can help us answer the Justin Bieber question as well.  Albert-László Barabási and his student Ginestra Biankconi used Bose-Einstein statistics to develop a fitness model of networks that explains how a propensity to attract links can eventually snowball into an all-out phenomenon.

Although the model is too technical for everyday use, the idea that an overnight sensation follows a distinct mathematical path brings sheds new light on a phenomenon that used to have an almost supernatural mystique.

How Do You Integrate a Marketing Campaign?

“I know that I waste half of my marketing money, I just don’t know which half,” is an oft-repeated quote attributed to many different people.  Like most popular aphorisms, it’s witty while retaining an essence of truth.

Over the years, we’ve gotten pretty good at measuring individual mediums, such as TV, but effective marketing mix models have eluded us.  However, it should be clear by now that we’ve been looking in the wrong places.

How pacemaker cells in our heart sync isn’t much different than consumers boycotting Nike or a line going around the block for a surprise hit movie.  Learning how contagion spreads on a college campus is essentially the same as how brand sentiment moves through the marketplace.

In other words, specific placements aren’t nearly as important how marketing messages travel through networks of ordinary people.  Moreover, as network theorist Duncan Watts points out in his new book, Everything is Obvious, mobile and social media are giving us a whole new tool set that will revolutionize how we understand and practice marketing.

Taming Chaos

In 1900, Louis Bachelier published his dissertation on speculation, which foreshadowed Einstein’s famous Brownian motion paper that arrived just a few years later.  Although ignored for over 50 years, Bachelier’s work was rediscovered by Paul Samuelson and became the foundation of modern mathematical finance.

However, all was not well.  Benoit Mandelbrot pointed out that markets don’t work anything like physics.  In human endeavors variables aren’t independent entities, but interacting ones that feed back on each other.  This type of chaos was not captured in the bell-curve driven statistics that the financial models were based on.  He predicted disaster.

He was, of course, proved right by the recent financial crises, which occurred just before his death in 2010.  The stakes are much smaller in the marketing world, but the same logic still applies.  Our marketing investments are not solely dependent on our directed actions but greatly affected on how information flows through the consumer network.

It is this murky, mysterious word-of-mouth transmission that is just beginning to be uncovered by social network analysis.  As our social interactions are being increasingly encoded digitally and technology enables us to overcome computational hurdles, marketing, as we know it, will be utterly transformed.

– Greg

11 Responses leave one →
  1. August 4, 2011

    Very nice article with its prequel “stories of networks!”

    The concept of “friends’ friends network” seems obvious yet compellingly efficient in terms of observing or even detecting patterns of behavior which could be typical to the network.

    This made me think about a definition of friend – I nominate a friend, influential/well connected, which he might or not be as well as cherish or not similar friendly feelings towards me – and whether it could not be substituted by colleague or any sort of relationship-delineating factor, especially in societies (I think many in West and East as well) where one has more colleagues with whom one interacts both for work and socially. In case colleague is what’s nominated, the same Granovetter’s ideas apply, except here bonds between points might be doubly strong as colleague implies a colleague and also possibly a friend.

    In your section for stopping the flu above you quote the study of Christakis and Fowler. Christakis in his TED talk tells of three ways that massive-passive intervention can happen: passive, quasi-passive and active. The first two are mostly based on lateral information one receives from key hubs of the network. For the active, as perhaps in the best of cases for the other two, the early detection is what Christakis preaches, this being, according to him the key to solutions/improvement of networks. While I agree about the importance of early detection in prevention or tacklign various network issues, I wonder if there are other active massive-passive interventions that can go further than an early detection.

  2. August 4, 2011

    Good points. It’s all still very new, so I’m sure there’s a lot yet to be uncovered.

    – Greg

  3. August 4, 2011

    Ya know bro, I like to sit down and take a couple of hours to really think in depth about things taking place around me and more specifically, how I’m going to interact with them.

    In those moments of deep thought, I begin to think exactly like this. This was a very in depth post, based on some very real stuff.

    I too believe marketing is transforming. This word of mouth thing is murky, isn’t it? Love it though. Sweet post.

  4. Barry Wellman permalink
    August 7, 2011

    Strogatz & Watt do fine work, but they didn’t “discover[ed] the basic principles of network theory” as you unequivocally assert. There’s a long history there. Read Linton Freeman’s book on the subject.

  5. August 7, 2011


    I see your point. Euler developed graph theory way back in the 18th century and Erdos, Rappoport, Milgram, Garnovetter (among others) all made important contributions as well. However, you could say the same about Einstin and relativity or Darwin and natural selection or just about anything else. As Kuhn pointed out, primacy is really a matter of perspective as much as anything else.

    However, network science in the modern sense really comes from Watts and Strogatz paradigm changing 1998 paper along with the Barabasi work that followed shortly after (work that he himself says was inspired by Watts and Strogatz). So while I don’t mean to unequivocally assert anything, I think there’s good basis for my statement.

    Thanks for contributing.

    – Greg

  6. December 4, 2011

    good blog but stop writing like this:

    ” Benoit Mandelbrot pointed out that markets don’t work anything like physics.”

    No. He argued that markets don’t work anything like physics.

    Do you see the difference? The way you wrote sounds like he was pointing out a universal fact. Wrong. He made an argument.


    “He was, of course, proved right by the recent financial crises, which occurred just before his death in 2010.”

    Stop. You are making huge leaps by citing very little data. That statement needs considerable data to back it up that you don’t provide. Using words like “of course” and “proved right” shows your, for lack of a better word, ignorance.

    Instead, write like this:

    “The recent financial crises gave support to his argument because of blah blah blah”.

  7. December 4, 2011

    Thank you for sharing.

    – Greg

  8. December 9, 2011

    Fascinating post, I’m off to do more research including viewing the Ted Talk video you referenced. Thanks!

  9. December 9, 2011

    Have fun!

    – Greg

  10. March 23, 2016

    Dear Greg,

    Greetings from India!

    Your articles have helped me learn how to think about things happening around us, and this is especially helpful for me to impart these lessons to my young undergrads who are in the technical education field.

    Thank you.


    Prof. RK Dash

  11. March 23, 2016

    Thank you, Dr. Dash. That’s incredibly kind of you to say.

    – Greg

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS