Knowledge Assets: Leveraging their Value at Raytheon: A Case Study, Joint Meeting of KMPro New England and Boston KM Forum, June 30, 2004, Bentley College

Over forty attended the presentation by three Raytheon managers/knowledge management champions responsible for KM activities across the corporation as part of Raytheon's Six Sigma initiative. The audience heard a detailed presentation on the missions of the corporation, which led to a discussion of how knowledge-based IT activities are contributing to corporate cohesiveness.

The speakers were:
Christine Connors, Raytheon's Metadata Architect, part of Corporate I.T.'s Information and Knowledge Management Strategic Initiative. Christine leads the Knowledge Representation Team, winning the 2003 Raytheon Excellence in IT Award for Collaboration and Knowledge Management. Christine earned a B.S. in Theatre Arts Management from Ithaca College and an M.S. in Library and Information Science from Simmons College.

Keith Cromack is responsible to both the Raytheon Company CIO and the VP of Communications, developing and implementing Information and Knowledge Management and e-Communications programs across the company. He is a member of the Conference Board e-business council and a Raytheon Six Sigma Specialist. Keith speaks often on Electronic Communications and Knowledge Management and the use of technology as an enabler of change. Cromack received his bachelors' degree from Bentley College.

Roberta J. Preve, Manager of Raytheon's Global Headquarters Library and has been with Raytheon since 1996. She is a member of the Knowledge Representation Team and a Raytheon Six Sigma Specialist. Roberta earned a B.A. in Geology from the University of New Hampshire and an M.S. in Library and Information Science from Simmons College.

Key points from speakers and participants:

KMPro Breakfast Meeting Summaries

December 19, 2003 - IT Infrastructure and Applications for Managing Regulatory Information Organizations that were present are predominantly active in issues relating to biotech, pharmaceutical, and health sciences. These areas have explicit mandates for the types of information they must retain and deliver to the FDA and US Patent & Trademark office for product approval and for securing patents. Other types of institutions mentioned Sarbanes-Oxley as legislation that will drive changes relating to record-keeping but IT efforts to systematize capturing regulatory information, as it is disseminated, appear to be early stage or fractured among many groups. Attendees recommended two services that meet some of their needs for being notified of changes on Web sites that they must frequently consult:

There was a strong recommendation to get on the mailing list of any agency that has oversight in your industry. These will usually keep you up-to-date on new government regulations that affect your industry. Attendees received an annotated handout with links to various industry specific Web sites. Typical of those on the list are:

January 16, 2004 - The Role of Records Management in the Knowledge Organization One company described an active push to distinguish between regulatory information and archives records in their records management area. In this case regulatory information they need to track comes from the FAA, DoD, and IRS, and the archives are the information that they produce for their sponsors. Their focus is on identifying the material and creating location information with no particular concern about searching the content. Another company indicated that their Document Management Department manages all published papers while non-published materials are handled by the individual groups producing them. The SAA society of archivists was mentioned as a valuable professional resource for information about records management. Some records management software packages that were mentioned are: True Arc (now owned by EMC), Cuadra Associates, and Documentum (now owned by EMC). However, concern was expressed that implementing a new (and improved) software system for records management presents many headaches, among them the need to re-verifying all the content. The statement was made that PDF may not be retroactively displayable in new systems and the JPG format may be more sustainable over time.

Additional readings.

February 20 - Where is Enterprise Knowledge Being Captured?. Discussion quickly evolved to show great variation in interpretation of the topic. The Harvard Computing Group uses a tool they have developed, TaskMap, to capture process knowledge in their clients' organizations, resulting in concise documentation of current or proposed business processes. This was evidence of a consulting practice managing knowledge capture on behalf of a client. Other attendees from larger companies referred to various department level applications, primarily specialized databases written internally to capture project information, financial planning documents, metadata about reports and research activities. Smaller organizations still rely on file management systems, spreadsheets and computer directories to locate their (mostly electronic data). There was almost complete unanimity among those present that their organizations are only beginning to work on the problems associated with enterprise-wide management of knowledge. This is in spite of the recognition that dissemination of knowledge to employees gives organizations a market advantage by enabling better performance. On the flip side is the perception by many that sharing their knowledge may not be a personal advantage. Cultural issues are still a barrier to centralizing and codifying knowledge capturing activities.

Additional readings.

March 19 - Beginning Knowledge Mapping Workshop. When the group was asked, what do you think of when you hear Knowledge Mapping, there were several answers. Among the ideas were: defining the location of knowledge, defining what it is by format or links to it, and the context (who has it/who needs it, currency, where it was captured, distribution). In distinguishing between mapping knowledge and information the group agreed that both explicit representations of knowledge PLUS tacit knowledge should included. This means including such diverse resources as: resumes, images, transactional information, queries in log files, e-mails, project information, formal reports, etc. The information context may manifest itself in databases, in citations leading to experts categorized by subject, word of mouth, social networks, or through a special library as a connector. Tools that are being used by attendees for mapping knowledge or tools that act as knowledge maps are:

Additional readings. Phil Murray brought copies of issue information for his latest publication, The Barrington Report on Advanced Knowledge Organization and Retrieval (BRAKOR). It contains many articles related to the topics of this meeting and others. See the KMconnection.com for access to BRAKOR.

April 23 - Identifying Communities of Practice. After the meeting, Allan Lewis summarized the general consensus among members as the goals (of CoPs) are maximizing the amount of organizational knowledge sharing among people with common interests, which leads to increased organizational efficiency (making it easier to find commonly needed tools, making it easier to find experts and minimizing duplication of effort). These benefits of well-functioning Communities of Practice will then lead to improved rapport (and morale). This should then cause even more knowledge sharing to occur. He included this information as well. The group came up with a list of features that they believe CoPs contribute or embody. Among them are: mentoring, common interests, education, pulling information together, loosely controlled access, high social aspect. From CoPs often evolve centers of excellence, consulting practices, formalized processes, procedures vs. having best practices evolve from trial and error. Some of the drivers of CoPs seem to be: regulatory issues, P&L, education and training needs, plugging an expertise gap, mergers, need to solve a particular problem, economies of scale relating to practicality of meeting, creating awareness, using technology across boundaries, overcoming licensing barriers, need for finding common language for sharing and searching information, efficiency, rapport and knowledge sharing.

Additional readings.

May 21 - Differentiating Taxonomies and Ontologies. At Rebecca's, 28 KMPro participants gathered to begin a dialogue about this topic that touches knowledge management, particularly in the content management area. As a precursor to the discussion, all attendees received a bibliography with links to articles on the topic. The discussion began with some definitions of Glossaries (alphabetic lists of terms with definitions, usually in a specialized discipline), Taxonomies (for our purposes a list of terms in a specialized domain, organized by Broader to Narrower categories, similar to a taxonomic structure of biological organisms but may pertain to any subject matter, currently used for organizing Web searchable content), Thesaurus (a specific ANSI standard (Z39.19) structure of terminology in a specialized discipline with attributes similar to taxonomy, previously described, but with possibilities for more scope, more depth, and more relationships, applied to indexing content by attributes, such as, subject, corporate name, etc.) Phil Murray contributed a document on Terminology of Knowledge Organization to the session. A copy can be found in the second issue of Brakor at the KMConnection Web site, as well as these definitions of faceted classification.

We touched on Ontologies without getting explicit definitions but with some agreement among the experts in the group that:
1. The term has been adopted by developers of semantic search technology from its origins in philosophy.
2. Semantic search is a more advanced form of automated search than matching strings, and more advanced that "fuzzy logic searching" which is based on rules.
3. The intent of semantic searching applications is to enable natural language queries to find content that explicitly answers the question being asked. It is a field in its very early stages.
4. Ontological structures can provide frameworks for extracting semantic meaning from specialized clusters of terms by providing contextual relationships that extend far beyond a flat hierarchy of Broader and Narrower terms that we see in taxonomies.
5. In the realm of search, ontology has become a loosely defined term for: a specialized domain of terminology and all possible relationships (e.g. "is a kind of", "is an instance of," "will cause a," "is needed for") that give contextual meaning to term combinations. The usefulness of the framework is dependent on the presence of all meaningful relationships for the information domain being covered. [An example of "ontology" is the UML (Universal Medical Language) developed largely through the sponsorship and work of the National Institutes of Health.]

We also touched upon automated methods for producing subject specific taxonomy lists by using the most highly relevant documents on a subject as "learning or training" documents from which a computer algorithm will extract terminology. That extracting terminology can form the basis of terms that will be controlled vocabulary, available for automatic categorization of a body of content. Often automatic extraction of terms is enhanced by human taxonomists who "correct" the terminology, add cross-references, and normalize synonymous terms to one entry with cross references to that entry from the others.

Finally, some discussion was devoted to the concept of faceted classification of documents. Faceted classification schemas assert rules for making sure that content is assigned subject terms associated with all attributes of the material. It focuses on rules to be sure that content is fully described from all points of view, or at least that the searching audience will be able to find it from any important perspective. (e.g. when indexing materials in materials science, facets to consider might be product names, chemical composition, electrical properties, strength properties, etc.).

At the end of the session, all participants were anxious to resume discussion at a future session, possibly half or full day workshop.

Additional readings.

June 18 - Examples of Taxonomies & Ontologies. Due to the high turnout of this meeting (over 35) and the location at Genzyme where we had Internet connectivity, we limited the roundtable discussion and focused on electronic examples of taxonomies and ontologies. The additional readings list at the following link are same as for the May meeting but also include definitions of terminology to help those needing to get grounded in the subject. Also in this document are links to graphic examples of taxonomies and ontologies that appear on the WWW. In addition to this overview, members of the group contributed other working definitions (cmsglossary.com) and examples from Norman Daoust of Daoust Associates on the EntityClass table for elements of the HL7 Reference Information Model, and from Deanna Briggs of MIT. Lincoln Laboratory a section of a taxonomy under development for subject categorizing their internal publications.

Additional readings.

July 16 - Keeping Content Fresh . Back to Rebecca's in Burlington, we had a great discussion captured in these notes from Lisa O'Donnell of Genzyme. The topic came from recent breakfast meeting where one of the participants asked about methods and processes for removing stale content on a Web portal. Large corporations engage professional content managers full-time to attend to the activity. Are there specific tools to help smaller organizations keep track of what needs to get weeded or replaced? What type of professional competency is appropriate for managing content? How can we judge what content is valuable, used and valued? Here is the discussion:

We began with a working definition of content from the leader as, all information made available to users through some type of portal interface. It may include static pages as well as transactional data and tools. Ken Bruss defined 'fresh' as accurate and timely and stressed the need for a quality control mechanism.

Ethel (Salonen) described her experience with the Plumtree portal at Millennium Pharmaceuticals. Their approach to keeping content up-to-date is to have portal administrators assigned to each community of practise. The onus for keeping content fresh is primarily on those administrators, though they are aided by automated reminders to review the workspace content after a preset time interval.

Distilling the lively conversation into 3 main subtopics:
Portal Wish List-what do people want?
- automated notification of updates to topics of interest specific to the user
- just-in-time delivery of information to users as it is needed
- Purpose-driven information-current, authoritative, reliable, usable and timely
- An automated system that does it all-monitors sources, classifies information, sorts by relevance, updates old content and delivers to the user

Barriers to maintaining freshness
- Social legacy or the organizational culture, e.g. chains of command, hoarding vs. sharing behavior, lack of communication
- Authority vs. experience: those who know and those who think they know
- Lack of standards
- Siloed thinking, vocabulary and organization

Keys to success
- robust classification schema, ontology
- reliable gatekeepers, most often a person or group who monitor the system
- well-defined processes that are widely known and practiced

There were several participants from the healthcare, biotech, and pharmaceutical industries. As a regulated industry, the need for such systems that can cull and deliver appropriate content is growing. Both Millennium and Genzyme are in the process of implementing the Plumtree portal, but it alone is not the silver bullet. These industries often maintain multiple tools to handle the documentation, data collection, statistical analysis and communications that are all required content. We need to recognize that any system needs a lot of human support not only in the planning and implementation phases, but also for ongoing maintenance. Librarians bring a skill set that can efficiently classify and sort content. Subject matter experts are most often tapped as gatekeepers, usually as an adjunct to their "real" jobs.

Additional readings.

August 20 - Defining the "Content" in "Content Management". This meeting was a true roundtable with each person weighing on their own or their organization's definition of content. Here is the list with some context as to the opinion giver: