Connection finders

From [[ CPsquare]], the community of practice on communities of practice.
Revision as of 09:47, 10 July 2009 by Smithjd (Talk | contribs) (Examples and Related tools)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
... part of the technology for communities project,
started off by the authors of [Digital Habitats], Etienne Wenger, Nancy White, and John D. Smith.


Connection finders analyze online activities of individuals to infer interest and expertise through keywords and frequency of use.

Uses in communities of practice

To some communities, this may feel like having "big brother" watching. When a community is large, very dispersed, or just gathering, however, it may be useful to the members themselves to get help finding others with similar interests. Many people fail to create personal profiles or describe all their interests, so connection finder software can ensure that all key interests are reflected in profiles, leading to more effective expertise location. Most connection finder systems permit the individual to review/edit any postings before the content becomes public. Seeing the network structure of the community change over time can also be interesting.Kaye Vivian 13:46, 6 July 2009 (UTC)


  • Together / apart, Synch / Asynch: Most useful when a community is widely distributed, physically or socially but is using some kind of technology to stay in touch that can provide data about interactions and relationships. The extent to which this tool brings people together is the performance metric.
  • Participation / Reification: Connection finders represent participation via some other tool and reify the their underlying social (or professional) relationships.
  • Individual / group: Helps us visualize individuals in relationship to a group and its structure.

Key features

  • Access posts The system looks at postings in the system forums. Less feeling of "big brother" because people know their posts are public.
  • Access e-mail The system scans all e-mails. This one can really feel like "big brother" because e-mail is usually perceived as private. Most software provides an editorial review step to enable the author to remove or modify text before it is posted on the system.
  • Subject line scan The system only considers email subject lines. This can result in flawed results unless users are well-trained in how to create and change subject lines to expose message contents.
  • Full-text scan The system looks at the full text of posts or e-mails. Most systems have a means to allow users to "opt out" of scanning for certain types of messages or personal content that should not be published to the community.
  • Survey questions Uses explicit questions or an SNA-type survey to derive connections. Requires work from participants, and is often only really useful if everyone participates.
  • Build profile The system builds a profile of contributors based on content scans.
  • Suggest connections The system suggests good potential others to network with. Good for large and loosely-knit communities where members may not realize who else they should connect with. Some platforms go further and connect multiple databases and directories to connect resources outside the system as well. Innovative serendipity can occur on searches in systems that are broadly connected to many resources.
  • Graphic representation The system produces a sociogram that represents the network structure of the community. Good for communities that want to reflect on their evolution or identify areas needing increased dialog or expertise.
  • See profile built The user can see what the system has inferred. Essential to give people access to what the system is inferring about them.
  • Correct profile The user can correct inferred information. People usually want to be able to correct inferences, using insights from alternate communications means that might not be included in an automated system, such as who has lunch.
  • Add information The user can add information directly into profile. Usually, the more control people have over their profile, the better.
  • Semantic search Entries in the system populate a search engine index with both document content and any semantic annotations added by analysts.

What information the system considers;

Connection finder software may reference any organizational content assets or structured information, including documents, images or recordings already available in a variety of locations. The newer systems also focus on scanning "unstructured" content, business processes, line-of-business solutions, and information exchanges between employees, such as blogs, emails and instant messages. By evaluating all the writings of an individual in the organization, a meaningful personal profile can be built automatically. By analyzing text using artificial intelligence, sources and documents that might not have been categorized together in a traditional taxonomy can be discovered.

The ability to systematically capture, create, manage, access, review, distribute, publish, store and preserve all business content-from ERP/CRM systems, databases, e-mails, documents, file systems and external information systems-and leverage this information to its full potential is the goal of connection finder software. According to KM World, "Structured information only represents 20% of a company's enterprise information assets. The remaining 80% of valuable information is in uncover (sic) form dispersed across the enterprise--in e-mail, document management systems, file systems, instant messaging systems, records management systems, knowledge management systems and other stored documents. This 80% of unstructured information is often overlooked as a key information source for tactical and strategic decision making." [1]

Business content should include best practices and personal expertise, and needs to be linked, referenced and integrated to give a complete picture. A single interface then allows users to access, search, report, collaborate, analyze, track and audit against personal profiles and other enterprise content. Kaye Vivian 14:09, 6 July 2009 (UTC)

What it does with information;

How much control users have over the inferred profile

There are a number of variables governing how much control a user has over their inferred profile, including:

  • Corporate policies
  • Software capabilities
  • Editing/authoring permissions
  • Profile elements to be modified
  • Evaluations/rankings by other users (if allowed)

Examples and Related tools

What about examples such as: