Social Graph: Is it fully exploited

With the exponential growth of facebook, twitter and to some extent linkedin , there is a tremendous interest in the social graph, social networking and related areas.

We see that these web properties are centered around four pillars

1. Relation between users (social graph)

• Personal friends in case of facebook and myspace
• professional relationship in case of linkedin
• interests in case of twitter

2. Data around that relation
There is a lot of communication which is open (in groups or to everyone on the internet) and very easy to access.

• wall, comments on facebook (two way communication)
  
• resume like info on linkedin (one way)
  
• short thoughts and happenings around on twitter (one way, but ability to forward with retweet)

which caters to various needs and urges of a person : to converse, just put down ones thoughts in public, get support from people, to show the list of your resume, contacts to others (sometimes, this is kind of a ego booster), etc

3. Presenting the data to the user with a simple user experience
With many users and lot of communication happening in the network, it is very easy to get overwhelmed with data and number of users. The sites have good algorithms to give relevant updates to the users, suggest the users what they might be interested in etc. In addition, these are delivered over a very simple UX which just works makes it very user friendly.

4. Sharing of data between relations
This is still evolving and we are seeing signs of it with atleast the graph sharing (such as facebook connect) and private deals between major sites to mine the data

In this post, we focus on the relations between users that are being extracted in today's social networking sites and explore possibility of other potential relations and overlay of relations which hold promise.

Amartya Sen, economist and humanitarian, writes in his book "Identity and violence"

• "I can be at the same time an Asian, an Indian citizen, a U.S. resident, a British academic, a Bengali with Bangladeshi ancestry, a graduate of two colleges in two different countries, an atheist with a Hindu background, a non-Brahmin, an economist, a researcher and teacher in philosophy, a Sanskritist, a married man, a feminist, a defender of gay rights, ..."

Similarly, when we talk about social graph and relation between people, there are multiple ways in which people are connected. We currently see, personal friends and professional friends as the main way of relating people due to facebook and linkedin. Then twitter phenomenon happened where the "follower" relation came up. A "follower" follows another user for the information that he thinks is useful (for gossip, for getting a buzz of whats happening etc)

We see that the following relations are actively pursued by many companies

• Personal friends and acquaintances : facebook, myspace, hi5, ibibo
• Professional friends and colleagues : linkedin, naymz, peerpower
• following interests and news: twitter, smsgupshup
  

There are many other relations that are not yet actively pursued. Some of the relations that we would have encountered and made use of in some context are

• Geography based (such as living in the same location)
• Parents of a given school
• Users of a particular iphone app
• People who watched the movie "avtaar"
• Customers of Sony DVD handycams
• Music lovers of a certain genre
  
• Startup founders who are looking for funding
• etc

I would like to categorize relationships as primary and secondary. The definition of primary and secondary depends on the context and also on the primary purpose of the web site/app. Not all of the above mentioned relations might qualify as primary relation (due to the size of the group or due to the temporary nature of the relation)

• In facebook primary relationship is personal friends, secondary is groups/fan pages
• In linkedin, primary relationsihp is colleagues. Secondary are people's interests, people of same company, people in same geography etc
• In games, primary relationship is players, secondary is personal friends
• In purchases, primary relationship is people who purchased the same/similar product, seondary is like minded friends (maybe of similar income group)

I believe that there is value in every relation that can be extracted to give value to the user as well as other interested parties. And then, further value can be added by overlaying that relation with other relations

• One might be interested in people who are my colleagues as well those whose children are in the same school. This will help me to request some feedback on school from like minded people (assuming my colleagues are like minded :-)). Here, one would need an overlay of colleagues + school parents
• Say, we want to get feedback on a camera, apart from searching forums, we would like to ask the people who have bought the camera and who are friends. Thus one would need an overlay of "people who bought camera from same dealer" with "friends"
• To get feedback on a hotel, in addition to the web forums, if we just know somebody from our friends who has recently visited the hotel, we can get information to which we attach more weightage (compared to anonymous reviewers)

Thus, even though friends list (facebook-like) is a huge graph which encompasses all the people on this planet, there are many other relations which are yet to be built, maintained and made use of.

In my next post, I will try to cover some relations and discuss qualitatively, their potential value. One of the ways of getting the list of relations is to see the number of community/forum sites that are around on the web. Each of them is a pointer to a relation between people. The shortcoming of the forum sites seems to be that they are not integrated into the users' other social graphs. Thus, it becomes difficult to implement the third pillar that I have mentioned above (Algorithmic updates and insights)

I found the chonological figure in this link interesting and a discussion on the ownership and federation of identity and social graph in this link