Empowering Social Influence With Mark Granovetter

Hearing the phrase "Social Networking" immediately brings to mind applications such as Facebook, MySpace, and LinkedIn. Depending on where in the world you're reading this post you may identify with a slightly different set of social networks - Friendster is still alive and now hugely popular in Asia, as well as Hi5, Bebo, Orkut, and Mixi (Japan's biggest social network). The diversity of social networks has a rich history that encompasses some of the best Use Cases of what to do, and what not to do when building successful online communities.

Each network began their approach to building their online communities from slightly different directions - specific regional or cultural communities (Mixi), common social (MySpace), collegiate social (Facebook), professional social (LinkedIn), sub-networks (Ning) etc. Over the years many of them have expanded to overlap each other to some degree. While the direction of their approaches may have been different, the one thing they all have in common is that their value lies in the application and understanding of social influence.

However before understanding social influence, trust is where most social networks begin their focus. After all, without trust you can't build large and vibrant communities - and without large communities you can't really get a clear picture of how social influence works.

Over the years social networks have learned valuable lessons through mistakes such as Facebook's Beacon advertising program, and MySpace's extra open development platform which initially let Javascript run amok in ways that MySpace never intended. Whether it was to counteract spammers or preserve control over a person's personal information, all sites have made enormous efforts to offer finer and finer granular control along with enough oversight to ensure a pleasurable and "safe" experience. And as these networks have grown to number 100s of millions of users, so too have their valuations.

But VCs and investors are all still asking the same questions. What justifies these valuations? How do you plan on generating revenue let alone a profit? A look at these tables about various social network valuations from TechCrunch will make this point even clearer:

Although there are a few understandable holes and caveats in how TechCrunch has done their analysis, it's still a great start at normalize the valuations in such a way as to calculate what a user is worth based on demographic/network type, and online ad spending towards these users. As Michael Arrington says at the end of the article, "Two very similar networks may monetize vastly differently based on methods of advertising and even the brute effort and passion of the employees. This model obviously doesn’t take that into account." So what are the various methods of advertising available to these networks?

Ultimately, each network monetizes by offering its advertisers 1) reach (hey! we have X-million users), along with 2) an in depth understanding of their user base (stripped of specific names).

You love being able to find friends who live in your area, went to your Alma Mater, High School, etc. - and advertisers love to be able to target their ads to people in specific demographics, interest groups, and locations. It seems to be a win-win! But in the end this is still very 1.0 in terms of advertising. Having greater control of ad targeting is great, but in order for corporations to make a significant impact to their bottom line, they must start coming up with better "social" ad units while evolving their approach to online advertising.

Enter Social Influence:

Facebook's Beacon was just such an attempt at being innovative. The only problem was that people didn't have control over how their personal information was being used. Even worse, unlike other referral-based ad programs (Adsense) there wasn't any referral revenue split with the person who's picture and information was being used to advertise. You would think that if Facebook did a generous Google-esque 70/30 ad revenue split, people might have been a little more accepting of their program.

In fact, the latest iteration of Facebook's Beacon seems to be charging in just this direction. However, Facebook should be very careful with how any referral payment system is developed before a "Beacon Fiasco Part II" occurs. First of all, Facebook users don't want their recommendations to come across as being in bad taste - basically whoring out their once objective recommendations for profit. But who knows, during these tough economic times a little extra scratch through your social network might go a long way in softening impressions of this program. The other aspect for Facebook to consider is the investment of infrastructure to handle referral payments and customer support. In the end, Facebook needs to ensure that users are willing to support this new model, and support it in large enough numbers to have the approach make sense. Spending millions to roll out an idea that people are still averse to, and that backfires again, would be PR mess that Facebook would like to avoid. Yet with the deep coffers of Facebook, I don't think they'll be hurting to try this new approach. Taking another chance to monetize social influence is also something that researchers and investors alike would be interested in.

Facebook's updated approach involves being able to share your purchase information with specific friends who are also interested in the same types of products. While blasting your purchase history across to all of your friends is what advertisers would still prefer - sharing a purchase you've made with "interested" friends makes more sense. Advertisers can always pay more to have all of your friends see the information (in hope of some additional peer influence) , but this is basically where Facebook is going with this. Assuming such highly targeted and co-opted ads makes for a better ad unit, the cost for a click-through or purchase, based on this type of ad unit, should also command above average rates. But will any of this actually work? What is the science behind the approach, and what is the best way to make use of it without being evil or Big Brother-ish?

Social Influence, and Social Networks have been closely studied by Mathematicians and Sociologists since the late 1940s. In fact, while we tend to think of "Social Networks" as a relatively new phenomenon surrounding online communities, the reality is that a variety of disciplines have been interested in understanding social interactions from a variety of perspectives throughout history (Sociologists, Anthropologists, Mathematicians, Archeologists etc.)

The cornerstones of social network studies have largely been based on, 1) formal theories organized in mathematical terms, and 2) systematic analysis of empirical data. By the late 1970s with the advent of discrete combinatorics (particularly graph theory), the study of social networks really began to take off (insna.org). One of the most influential researchers in this field is famed Stanford sociologist Dr Mark Granovetter, whose 1973 paper "The Strength of Weak Ties" is considered one of the most influential sociology papers ever written. It is also one of the most cited research papers in this field. We are also extraordinarily proud to have Dr Mark Granovetter working with the deeda Team as we continue to build our application.

The Strength of Weak Ties

In marketing, politics, or your personal social network, the weak ties are what enable reaching large populations and audiences that are not typically accessible via strong ties. After all, your closest friends usually travel within the same cliques, so the information that is available is fairly uniform and the ability to widely distribute new information is naturally limited. In his 1973 paper, Granovetter used the example of rumors to show the limit of these strong ties. Because of the close interrelation between strong ties, a rumor would quickly spread through a few friends and then die out. Most conversations would soon start with, "Did you hear about...", and be immediately followed up with, "Yeah, I heard..." - end of story. But weak ties are where the broadest amount of information distribution occurs - and where rumors can take on a life of their own. The Boston Globe recently did a facinating story about the spread of rumors, their importance, and how certain actions either help spread or counter rumors. While rumors don't have a positive connotation, they still act as a simple example for information diffusion.

Getting advice through LinkedIn Answers is great positive example of leveraging the strength of weak ties and information diffusion. While the people within your immediate first degree network are likely to be exposed to the same information and resources that you are, it's the weak 2nd and 3rd degree ties that are likely provide you with some new insights and beneficial information. To be completely accurate, one of Granovetter's original definitions of weak ties involved friends and contacts who were once strong ties, but have since drifted apart. Today, we typically refer to your friends' friends and other broader definitions of what constitutes a weak tie.

In his Harvard dissertation, Granovetter expanded on these ideas to show how people were more likely to find a new job through their weak ties rather than their strong ones. Additionally, Granovetter's research also covered how fads are created. His Threshold Models regarding the "tipping points" of fads were later made famous through Malcolm Gladwell's book, The Tipping Point.

When monetizing any social network or social application, it is therefore important to understand where the value and growth in your network lies. Understanding the strength of weak ties, encouraging weak tie exploration, and creating systems that help diffuse information through weak ties, are all aspects that every successful social application should take advantage of. Combining, linking, and encouraging the growth of sub-networks (smaller groups within the larger mass of a network), are also key components to increasing the overall value and growth of a network.

In the diagram to the left I've tried to quickly outline the mathematical component of our discussion of social influence. These representations show the concepts of node and point centrality as they apply to this subject. The three types of centrality (relationship between nodes) are defined through degree, control, and independence. Visualizing these nodes also helps explain why leveraging weak ties is so valuable.

1. In this example, You and Tom are connected to the most people and therefore have a high level of degree centrality - or to put it another way, you both have the most amount of friends/connected nodes.
   
   
2. "Me" represent the concept of control centrality, since I provide a bridge between the most nodes and amount of information. My ability to control this information flow is important to the overall network.
   
   
3. Tom also represents independence centrality - since he is the most closely connected to other people through multiple people. Tom is minimally influenced by any single person, and is therefore less exposed to control. He is not dependent on a few specific nodes for information and communication.

A person who controls the greatest flow of information within a network is more important than one who may have more friends. Understanding centrality can also indicate which members are the most useful or well connected and therefore offer the best information resources. (Singh)

Sites such as Spoke leverage these concepts by showing you the shortest path to beneficial connections you'd like to develop within their network. If Tom has the greatest connections to investors, and I'd like to get introduced to them, Spoke will display the best "path" through which to approach Tom.





Monetizing Social Influence - Profits VS Privacy

So far we've covered social networks, the theories and models under which they flourish, and we've also touched on how various networks attempt to generate revenue through applications of social influence. We've also come full circle to the crux of the problem - sky high valuations without sky high revenue. A better understanding of social influence might allow networks to create better ad units for advertisers while also providing a highly profitable revenue stream.

However, a related problem with most monetizing plans around social influence is the amount of transparency, control, and privacy that is taken away from people and given to marketing and advertising companies. While we want our favorite social applications to be successful, that doesn't mean that we want them to pan out our personal behavior and social influence to the highest bidder.

Even though companies typically state they only use "data stripped of personal information", trust is the first thing that begins to fracture around these new business models. As Facebook's Beacon taught us, the backlash from consumers and the PR nightmare just isn't worth breaking the trust that social networks are founded on. What is the optimal business model that protects privacy while producing profits for social applications and advertising agencies?

Social networks are full of exciting data that can help us better understand our real and virtual behavior and social interactions. We can use them to help us find better jobs or fill the knowledge gaps within our current groups. Academic research can make huge strides and test new models with the information that is present in these networks. Advertisers can use you to influence your peers so they buy similar products, or stay loyal to your favorite brands. Yet, along the way each layer is peeling away a bit more of your privacy. And your loss of privacy is being bought and sold to cover the cost of bandwidth bills, servers, and hired employees needed to continually build, fix, and expand services.

Empowerment

Dr Granovetter and I have approached the problem of finding the right balance between profits and privacy from the perspective of Empowerment. The ingredients found in the studies of social influence, strong and weak ties, network structure, centrality, and even social marketing - are not inherently dangerous, meant to strip people of their privacy, or trample on trust. But this is mainly true when approached from a specific direction - and that direction must be the empowerment of the people these theories and models have been based upon.

What everyone has gotten wrong up to this point is in approaching the monetizing question from the direction of the people who want to spend the ad dollars. And just like the same ingredients in your kitchen that can help you, can also harm you - mixing social influence based research from the direction of marketers ends up being poisonous. Advertisers can't assume that they (who are often pretty slow at the technology innovation part) can pull the wool over the eyes of these tech savy online communities.

Are you really not storing my personal information? If I opt out of the program, am I really out of it? For generations advertisers have assumed that what consumers don't know will help them make bigger profits. But today's customers are a different lot. If social networks understand the importance of trust, then advertisers need to also understand how vital it is to their online marketing future. If social networks understand that this trust is the foundation of their network's value, then advertisers must also understand that any attempt to monetize social influence must first be built upon the very same trust.

In the next part I'll explain how Empowerment should look from the perspective of advertisers, social networks and the millions of people who flock to use social applications every day. Empowerment will be the key to ensuring that our Profits and your Privacy can both be produced and protected.