Here at YapMap we have built a search technology that does a better job of searching social content. As we were going through the process of doing this, we realized that social networks (where the social content resides) can be broken down into two broad categories: Communities of Relationship (CoR) and Communities of Interest (CoI).
So why is the distinction between CoR's and CoI's important for YapMap? These user and network dynamics are very different. If you are building a search engine for social content (as we are) then you need to understand these dynamics and how they will affect a user's search behavior.
In a Community of Relationship, the focus of the social interaction is the personal relationship. Whether it is family, friend, colleague, or acquaintance, the reason I am connecting with that person is because of our personal relationship. For example, it is unlikely that I would be interested in what my sister is doing (raising kids, gardening, teaching crafts at an elementary school, celebrating a birthday) unless unless we had a personal relationship. In short, I care about what you do because I know you. Facebook, MySpace, and Twitter (in the case that you are following people because of a relationship) are some examples of Communities of Relationship.
In a Community of Interest, the focus of the social interaction is the interest itself. Whether it is motorcycles, cars, politics, knitting, arthritis etc.., the reason I am connecting with this person is because we share an interest. No personal relationship is required - I don't need to know anything about you other than we both have the same interest. In short, I care what you do as long as I have an interest in what you do. Discussion Forums, Question and Answer sites (e.g Quora and Yahoo Answers), blogs/articles with comments, and Twitter (in the case that you are following someone because they tweet about a shared interest) are some examples of Communities of Interest.
The difference can also be explained using the concept of a social object. A social object is anything that we use as a reason to socialize with someone else. Some good reading on social objects from Jyri Zengestrom and Hugh MacLeod. In the case of a CoR, the primary social object is the relationship itself ("Hey, we used to be friends in High School so friend me") while in a CoI, the social object is the object itself ("Hey, anyone know the stock gearing for a Suzuki DR650 motorcycle?").
In the first case, you might have nothing else in common with that person but the shared connection of your high school friendship, while in the second case, you might not have anything else in common with the person but you both own a Suzuki DR650. One social object is a relationship, one social object is a thing.
It is interesting to note that much of the industry of social gaming (e.g. Zynga's Maffia Wars and Farmville on Facebook) has emerged because people in CoR's need something to talk about. Yes, that's right, Jr. High friendships from 25 years ago run so deep that an entire industry of made-up social objects has emerged to facilitate them.
The difference between CoR's and CoI's fall into three broad categories: Identity, Content, and Network.
In a CoR you typically know something about the person outside of the online world. You have some relational context in which to place them and this enables you to place an appropriate ammount of trust in the interaction. On many CoI's the real name and identity of the other person is not known and it is likely that you know nothing about them other than what they have shared on the CoI. Trust signals must be derived directly from the interactions on the CoI.
An interesting evolution in CoI's is the move towards a real identity. Sites such as Stackoverflow and Quora are good examples of CoI's that require a real identity.
When you know someone as you would in a CoR, you usually have some idea of their interests and their level of expertise in these interests (e.g. my friend is a gourmet cook, my brother knows how to fix a jeep). But on a CoI, the lack of a prior relationship does not allow for this information. Initially, one has no idea of the credibility of the other members of the network. This can be seen when during the course of a CoI conversation when a number of different, and mutually-exclusive, "right" answers will be suggested for a solution to a problem. Which answer is really right? Which participant in the conversation really is an expert? In a CoR your reputation precedes you and so there is some context for the interaction. In a CoI, this context does not exist and reputation signaling becomes an active part of the interaction.
Content on most CoR's as the online version of small-talk. Sharing a recent picture, complaining about the weather, making a comment about a sports team, sharing a funny video, are all great at making people feel connected to each other. This content (with the exception of pictures and videos, which can be archival) usually has a short shelf-life and is often not useful outside of the context of the relationship (e.g. "my mom makes the best apple pie") or the time of the sharing (e.g. "We are meeting at xyz after work for drinks tonight, join us!").
Content on CoI's are the online version of a conversation, usually between a group of people. Controversial topics like politics and religion that are not usually broached on a CoR can be discussed on a CoI. It is rare to see comment streams of over twenty on a CoR, but discussions on a CoI can run into ten's of thousands of replies. Because CoI's are having conversations about things, and things are not dependent on a relational context or an event with a time, this content is often very long lived and can be actively mined for many years (e.g. "what is the stock wheel size on a 2004 Honda Civic?").
The need to find an exact individual to connect with drives the CoR to a "winner take all" uber-network (Facebook). Each individual is scarce, and so they all need to be in the same place to facilitate the connecting. But because CoI's form up around interests, their is a natural tendency for fragmentation along the lines of the interest. Everyone does not need to be in the same place. This is why there are over 2500 community forums about different types of automobiles, and over 4000 about online gaming.
Although there are reported to be over 500 million people on Facebook, each person's network is relatively small, only 120 connections on average. When it comes right down to it, most of us just don't have that many close connections (see Dunbar number for more on this). So Facebook can be thought of as a bunch of interconnected small networks, while some popular CoI's can be single networks with tens of millions of members (one popular gaming community boasts over 25 million members).
At YapMap we are tying to create a better way of searching social content and in doing so make this type of content more accessible and available. Given the differences in the types of social networks we realized that the problem of building a search tool for a CoR was much different than the problem of building a search tool for a CoI. Given the differences listed in the above analysis, we decided to focus our efforts of creating the best search experience for interest-based social content, the type of content found in a CoI. This is where our passion lies and where we feel we can make the largest contribution.