The Enterprise Social Graph - dummies

By David F. Carr

Social collaboration is as much about people as it is about content, therefore understanding how a collaboration platform represents the connections between people is important. The “social graph” is a term popularized by Facebook for the network of connections between people in an online system, particularly the Friend connections between Facebook users. An enterprise social graph is the same sort of thing applied to organizations.

Making friends with reciprocal links

Facebook was built around reciprocal connections, meaning that the connection exists only if both parties agree. Someone sends you a friend connection, and you approve it or not. Connection requests on LinkedIn work similarly.

On Facebook, LinkedIn, and other social networks that use this model, users can share content that’s accessible to their contacts only. This mode of private communication among friends used to be the default on Facebook although the service has moved to make more content public by default (so keeping it private takes more effort on behalf of the user).

The model of requiring reciprocal connections is not commonly used in social collaboration, though, because the entire network is supposed to be private to the company. Content may also need to be contained within a particular workgroup. Other than that, content shared on the social network is generally meant to be viewable and searchable by anyone within the organization.

Following people and subscribing to their feeds

Following people is the Twitter model. To follow another person on a social network is to subscribe to updates from that person. You generally don’t need another user’s permission to follow their feed. Although Twitter users can configure their profiles for greater privacy, individually approving follower requests, doing so still doesn’t create a reciprocal relationship because it doesn’t automatically mean the person agreeing to be followed will in turn follow the person who issued the request.

When reciprocal connections are formed — that is, people decide to follow each other — that’s still a signal of a tighter affinity that the social software can use to grant additional privileges. On Twitter, only reciprocal connections can send each other direct messages, which are a form of private messaging.

The follower model is also implemented in other social networks, including Facebook, as a way of forming one-way connections that don’t require individual approval. This became important originally as a way for celebrities and brands to participate in social networking by creating Facebook pages that users could subscribe to by becoming a Fan rather than a Friend. More recently, Facebook has added the option for individuals to add a Follow button to their personal profiles so people who aren’t necessarily personal friends can still subscribe to their public posts.

Social collaboration systems generally employ the follow model for connections between people, although they can also analyze reciprocal connections to construct a better model of organizational dynamics and do a better job of recommending people and content.

Subscribing to feeds about documents, events, projects, or customer records

The follow model is a good match for social collaboration because it’s easy to generalize. You can follow people and also groups, projects, documents, event listings, or records in an enterprise application: any object represented in the enterprise social graph. Following an object just means following the stream of updates about its status — say, when a document is updated or an order ships — as well as any comments attached to that object.

Searching for expertise

One classic application of collaboration networks is locating expertise on any work-related activity or topic, whether research and development findings, management strategies, or company history.

  • Connections: When you view a profile in a social collaboration tool, you can also browse a listing of that person’s connections. Again, in a social collaboration context, think Twitter (where the list of a member’s connections is public) rather than Facebook or LinkedIn (where access to your list of friends or connections is limited to existing friends or connections).

    In other words, even if you can’t get a response from one representative of a product team (maybe because that person is on vacation), you can probably click through to the profile of another team member and make your request. You simply look at the relationships represented on the contacts list and in the social stream and choose the most appropriate person to contact.

  • Tags: Some of this comes from explicit tagging of profiles, based on extracts from human resources systems and other information entered manually by the user.

  • Posts: The social component comes from relationships that can be derived about interests and knowledge based on content postings and participation in online discussions. For instance, if someone consistently posts and participates in discussions about the industrial design of the company’s products, then that person looks like an expert — or at least, someone with strong opinions on the topic.

  • Content: Because the social graph includes both people and content, you can navigate from one to the other, finding more resources than if you were exploring a directory of people or a repository of content alone.

    Say you found a terrific document on a topic of intense, ongoing interest to you. Look up the author with a click on his name or profile photo. Maybe you and the author of the document work in research and development for different parts of the company but have been investigating related topics. Now, you can follow this person or send her a message, perhaps proposing that you two collaborate on future research. From her profile, you can also see some of the other content she has posted, including several other research papers and a blog post or two that are all relevant to your work.

Expertise location is important as a way of locating organizational knowledge that may not be explicitly documented anywhere.

It works the other way, too, where the personal connection comes first. Maybe you only looked at the other person’s profile because you noticed she had started following you online. You see from her recent activity on the network that she is someone you should follow. Meanwhile, you find your way to the wiki she maintains with several other collaborators on her team. In addition to finding useful content in the wiki, you follow the profile links to the other wiki authors, each of whom is also someone you will want to follow.