Social Models in Online Software: Part 1

by September 30, 2009

In my Impact of Social Models presentation at IDEA09 in Toronto, I presented an overview of the different ways social relationships are modeled in online software and examined if these distinct approaches resulted in different online behavior.

The Power of Context

In his book, The Tipping Point, Malcom Gladwell outlined his belief in “the power of context” which states that people’s actions are a product of the conditions and circumstances of the times and places where they occur. Put another way, people’s behavior is a function of social context and even small features of context can produce big differences in behavior.

So if people are prompted to act a certain way based on their context, what is the impact of the social models we build software around? How do people act based on the relationships they can set up with others? How are we shaping human behavior through the way we model social relationships in online software?

To get an answer to these questions and more, it helps to first look at the different ways we can model social relationships in software.

No Relationship

Perhaps the easiest relationships to model is software is no relationship. But online it is not entirely true that people have no relationship. After all, any two people online are both users of the Web and that might actually tell us something interesting about their relationship.

Specifically, we can infer some kind of relationship by detecting their physical location. This is getting easier all the time through GPS-enabled devices (outdoors only down to 10m), Wi-Fi beaconing (through gelocation API in HTML5 down to 50m), or cell tower triangulation (down to 100-400m). If people are frequently in close proximity, that could be an indicator of loose or close social relationships.

We could also align infer some kind of commonality based on people’s technology stack (browser, operating system, and settings) or even their Web browsing history. But none of these inferences are declared relationships. So they are probably less interesting from a social perspective.


Where we begin to see declared social relationships is in online communities. Though the term community has been applied to many diverse concepts, we only need to know that two people are both members of the same site and can interact with each other in order to define a social relationship between them. Specifically, two people can have a community (1 to many) relationship defined in online software if they:

  • Are both users on the same site
  • Can interact with each other: have the ability to messaging each other and/or collaborate
  • Leave visible traces of behavior
  • Can manage their identity (profile)
You might wonder, if these elements are in place, can there really be a genuine community? Just because people are participating on a site, are they doing it as a cohesive whole or as individuals or distinct groups?

A 2007 Stanford study on Yahoo! Answers looked at this question (and more). In their analysis of the connectedness of 700,000 askers and 500,000 answerers, they found a single connected component of 1.2 million nodes indicating that, in fact, most users were connected through some question and answer activity on the site. Yahoo! Answers user base interacts with each other as one large community. Perhaps even more interestingly, this held true in diverse categories including geographic sets like local business.

So we know a community is actually connected through their activity and we also know community activity typically breaks down. As outlined by Bradley Horowitz during his time at Yahoo!, an online community follows the 1% creator, 10% curator, and 100% consumer model. As an example, in an online discussion forum:

  • 1% of the user population might start a group (or a thread within a group)
  • 10% of the user population might participate actively, and actually author content whether starting a thread or responding to a thread-in-progress
  • 100% of the user population benefits from the activities of the above groups
This break-down seems to hold up across many examples of online communities:
  • 1.8% of all users write more than 70% of all Wikipedia articles
  • .003% of digg’s users are responsible for 56% of the stories on the site’s home page
  • .o64% creator to consumer ration on YouTube
Coming Next…

We’ll look at the next set of social models including groups, symmetrical/2-way and symmetrical/1-way personal relationships.