1

Knowledge Management: Putting the Puzzle Together One Piece at a Time

Albert J. SimardNatural Resources CanadaCanadian Forest Service

(Draft ver. 1.2, July 6, 2005)GTEC Professional Development Forum

Ottawa, Ontario, October 5, 2005

Abstract

This paper describes approaches used by the Canadian Forest Service to begin transforming itselfinto a knowledge organization. It begins by outlining strategic planning for an enterpriseapproach involving an environmental scan, a Business Case, a KM Framework, and anImplementation Strategy. It then explains a shift to a strategy involving individual projects thatsolved specific business problems. The paper describes five KM process projects: preserving,sharing, integrating, and managing knowledge and a knowledge policy. This is followed by fiveexamples of content-based projects: automated monitoring, compiling and sharing data,disseminating knowledge, enhanced client service, and preservation and learning. The paperconcludes with strategic considerations of this approach

1. Introduction

The emerging knowledge economy combines unprecedented challenges and opportunities for allorganizations, including those that create its basic resource - knowledge. As challenge, scienceorganizations must come to understand the unique economic properties of knowledge and how tomanage it for organizational advantage. In particular, they must transform themselves fromtraditional organizations of intelligent people into intelligent organizations that can learn andadapt to rapidly changing environments. As opportunity, global connectivity enables enormouscreative synergies of massive networks coupled with low-cost knowledge dissemination andsharing at the speed of light.

During the past four centuries, scientists have become highly adept at discovering and creatingknowledge, to the general benefit of modern society. Traditionally, once knowledge has beencreated and added to the scientific literature, the work of science is seen as finished and it moveson to the next discovery. Progressive S&T organizations have demonstrated that shifting amodest fraction of the amount invested in creating knowledge into managing it as anorganization’s most valuable asset can substantially increase it’s value (Buckman, 2004). It hasalso been demonstrated that knowledge management increases the value of forestry knowledge,from creation through innovation (Van Horne, et. al. 2005). Yet, in many S&T organizations,managing knowledge continues to compete for resources, often as a lower-class island of “non-science” surrounded by a scientific ocean.

1Unpublished documents on file at the NRCan, Canadian Forest Service, Ottawa, Ontario.

2

The Natural Resources Canada, Canadian Forest Service (CFS) is a science-based policyorganization within the Government of Canada. It’s mandate is to conduct forestry research insupport of forest policies that promote the wise stewardship of Canada’s forests and thecompetitiveness of Canada’s forest sector. The information and knowledge revolutions providetechnology and processes that enable the CFS to manage the knowledge that it creates so as toleverage its value not only for internal use, but also for use by the forest sector and all Canadians. This paper tells the story of how the CFS responded to the challenges and opportunities that itwill face in the knowledge economy of the 21st century.

2. Strategic Planning

Eight years ago, the Canadian Forest Service recognized a need to better manage its data,information, and knowledge, support the synthesis of new knowledge, and improve access to itsknowledge assets. In May, 1997, the CFS established a steering committee to create aframework for developing and implementing a knowledge infrastructure. By 1998, this grouphad produced four documents describing the context, concepts, framework, and terminology of aproposed corporate knowledge initiative, as well as a presentation to increase awareness andpromote the idea. In 1999, the documents were consolidated into a single report that wasformally published (Simard, 2000.)

In October, 2000, a Knowledge Management Division was established under the functional direction of a corporate Director-level Steering Committee. During the next two years, thisgroup produced a series of consensus-based knowledge management planning documents1. These were presented to senior management and various audiences as they were completed.

• Business Case - to focus on a direction and level of investment in knowledgemanagement and to prioritize what is most important. (current status, business drivers,vision, opportunities, challenges, options, implementation, and recommendations).

• Governance Structure - to address simple but crucial institutional questions such as:place in the organizational hierarchy, authority and accountability, and roles andresponsibility. (principles, framework, supporting documents, roles and responsibilities,external linkages, conducting business, amendments)

• Implementation Strategy - to map out a path for reaching the destination laid out in thebusiness case within the context of the approved level of investment. (principles,approach, management framework, KM community linkages, launching the KMProgram, path forward).

• Framework - to organize KM Program components into a relational hierarchy: goals(management, integration, sharing, preservation), dimensions (people, organization,process, tools, content), scale (strategic, functional, project), and definitions.

• Evaluation Framework - to ensure that best practices are followed with respect toevaluating the KM Program throughout the full life cycle of development and

3

implementation. (context, process, criteria and indicators, next steps).• Communication Plan - to ensure that the KM Program is communicated, understood and

supported by all staff and groups in the CFS. (strategy, objectives, messages, targetaudiences, action plan, timing).

As a final step in the planning process, an external KM expert was contracted to review the plansthat had been developed. As with any scientific endeavor, this insured that the work was sound,met professional knowledge management standards, and had overlooked nothing important. Thisresulted in a few revisions to improve some of the KM plans. The revised package waspresented to senior management who decided that an enterprise KM solution would be postponeduntil such time as a corporate reorganization was completed.

During the planning process, a number of project-scale activities had been started with theunderstanding that they would be needed regardless of the ultimate direction of the KM Program. They had been funded with available resources, so no new funding was needed. Much of thework was being done by volunteers who saw the need, usefulness, and importance of variousprojects, so additional staff was also not required.

Subsequent to the decision by senior management, it was decided that the KM projects wouldconstitute pieces of a “KM puzzle” that we would assemble one piece at a time. Unofficially,this was dubbed a “KM by stealth” approach. Each project would design a useful user-friendlytool that solved a specific, well-recognized business problem. Although the “picture” thatintegrates the KM puzzle into a program would be downplayed, the tools would be connected toeach other to enable users to easily move among them. After a critical mass of tools had beenimplemented and being used, promoting a KM program would be resumed.

This strategy is a hybrid of the two most commonly employed successful KM strategies usedtoday. An enterprise solution that is supported by senior management and properly integratedinto the organization is usually the most efficient and most successful in the long run, but itrequires the largest up-front investment, in the short run. At the other end of the spectrum, manygroups have successfully developed one project-scale element of KM. This reduces initialresource requirements and results in early success, but it does not yield the broad spectrum ofbenefits and return on investment of a full KM program, which was our long-term vision.

Two aspects of the KM framework guided the selection of projects. The first was that allprojects involved one or more KM goals to distinguish them from other activities, such asdatabase or information management. Second, all projects involved at least three of the five KMdimensions to span the breadth of KM.

The remainder of this paper briefly summarizes some of the KM projects that have beenundertaken under the “KM by stealth” strategy. They have been classed into two groups: KMprocesses and forestry content. Process projects focus on knowledge management solutions forbusiness problems related to running the organization. Content-based projects focus on

4

knowledge management solutions to create and disseminate forestry content.

3. Knowledge Management Processes

Process projects focus on solving business problems by accomplishing one or more KM goalsacross multiple content domains. They are presented in order of ascending goal hierarchy:preservation, sharing, integration, management, and policy. The Knowledge ManagementDivision led the development of consensus-based solutions for most of the KM process projects. All process sites, except Metafore, are only available internally through the Intranet. AsGovernment of Canada sites, all Web templates are available in English and French, although insome cases, content may be captured only in the language of the author.

3.1 Preservation: Briefing Note Database (BNDB)

• Problem: It was difficult to find briefing notes previously written by other authors. This

led to duplication, inconsistency, and slow response to requests for information. • Solution: Capture briefing notes in a database with multi-criteria search capability.• Status: Implemented (May, 2004); 800 briefing notes; quarterly usage reports.

Because briefing notes are “controlled” documents, two interfaces are required. Anadministrative interface allows designated individuals to enter and edit management-approvedbriefing notes into the database. Information includes: the briefing note document, which isstored in the language of the author; document details, such as reference numbers, dates, andlinks to other corporate systems; information about the originator, subject classification,background notes, and related information. Data entry takes about 5 minutes per briefing note,using “cut and paste” from electronic documents.

A Web-based user interface allows all employees to search the database. Several criteria can beused for searching, such as subject, key words, author, or date; a full-text search is also available. Search results are in the form of a list of all briefing notes that match the criteria. Clicking on anentry brings up the individual briefing note, which can be printed in letterhead format or copiedto the user’s word processor. The user can also view all supplemental information related to thebriefing note as well as contact the author through built-in links.

3.2 Sharing: Directory of Expertise and Skills (DOES)• Problem: It was difficult to find out what CFS employees know and rapidly bring in-

house expertise to bear on emerging issues.• Solution: Develop a process to capture and organize CFS knowledge and know-how and

develop a searchable database to easily and quickly find expertise.• Status: Pilot testing

Web-based access to contributor profiles is controlled through a password. Extensive help filesguide contributors through the submission process as well as searching. User profiles include:staff information, language(s), photo (optional), current projects, expertise, associations, CFS

5

publications, subject classification, Web links, and keywords.

Staff information is automatically downloaded from an organizational staff directory as aconvenience and to reduce errors. Each field has a context-specific pop-up help window. Different forms of data entry are adapted to the type of information: controlled list, free-formtext, browsing the user’s PC, and templates for photos, the Bookstore, and Web links.

Because employees complete their own profiles, submissions are reviewed and translated prior toposting in the database. We are currently automating notification of supervisors of “pending”submissions and requesting their approval.

All employees can search DOES by employee or by expertise. Searching by employee supplements information contained in the staff directory as well as provides links to CFSpublications and relevant URLs. However, the primary purpose of DOES is to enable searchingby expertise. Users can search one or all fields using a controlled list or free text. A searchproduces a list of all employees that match the selected criteria. The list can be iterativelyrefined. Clicking on a person’s name displays their profile and contact links.

3.3 Integration: Regional Libraries (Metafore)• Problem: Library users had to make six separate enquiries to access all CFS regional

libraries. Further, each library had it’s own system of managing their holdings.• Solution: Develop a process to map regional library catalogues to a single metadata

standard. Develop a search engine and Web portal to facilitate internal and externalaccess to all libraries through a single query.

• Status: Operational (2000); 2 million (internal), 1 million (external) records• http://catalogue2.nrcan.gc.ca/html/Metafore2/English_MultiLIS/Welcome3.html

This project began as a way to enable single-window access to six regional libraries. During thepast few years, it evolved to a substantial on-line library service. Searching for CFS libraryholdings can be done in any region or across all regions. Four search criteria are available:library catalogue (e.g., author, title, subject), publication date (s), language, or media. Initialsearch results are in the form of the number of records at each library. Clicking on a libraryyields a list of documents that can be ordered on-line through interlibrary loan.

Today, Metafore includes an array of on-line scientific search tools. There is a separate searchengine to access all electronic journals subscribed to by the library. There are links to a numberof global scientific search engines and databases, as well as links to search engines provided byother scientific libraries. Three “Current Content” abstracting services are also available. Finally,users can search bookmarks that have been preselected by a librarian as well as access a numberof reference tools.

3.4 Management: Knowledge Asset Inventory (KAI)• Problem: Knowledge has not been traditionally viewed or managed as an asset. It was

6

difficult to locate knowledge assets in the CFS.• Solution: Develop a process to inventory CFS’s knowledge assets. Develop a searchable

database to enable all staff find these assets by searching any field.• Status: Implemented (April 2005); 700 records; currently populating the database.

Knowledge workers enter information about their knowledge assets through a Web-basedtemplate. Help files guide contributors through the submission process, including describingtypical knowledge assets. Once entered, only the author can modify or edit an asset. Assetinformation includes: title, description, purpose, date of creation and expiry, lifespan, internal orexternal use, source, type of asset, location, and Web links. There is no management overview ofthe contributions; the quality of an asset is determined by the user.

All staff can search individual fields or all fields of the KAI using a controlled vocabulary or freetext. A search produces a list of all assets that match the selected criteria. The list can beiteratively refined. Clicking on an asset displays its full information, including author contactsand Web links.

Once the database has been populated, it will be possible to generate reports that will enable theCFS to begin managing it’s knowledge as an asset. Once we know what we know, it will bepossible to evaluate and prioritize its importance relative to the organizational mandate.

3.5 Policy: Access to Knowledge Policy (A2K)• Problem: Scientists are motivated to control data and information that could lead to

authored publications. This results in a reluctance to share unpublished material.• Solution: Develop and implement an Access to Knowledge policy that outlines roles,

rights, and responsibilities with respect to scientific content.• Status: Implemented (April, 2005); communication under way.

The CFS Access to Knowledge Policy establishes criteria and guidelines governing privilegesextended to CFS employees regarding their use of scientific content. It articulates a rationale fordifferent levels of access to organizational knowledge assets. The A2K Policy comprises ninedirectives: ownership, privileged access, partnerships, roles and responsibilities, levels of access,asset characteristics, openness, best practices, and cost recovery (Simard, 2005).

In keeping with the complex nature of knowledge, the guidelines emphasize common sense andjudgement rather than prescriptive rules and practices. They provide a framework to supportconsistent interpretation and application on a case-by-case basis. They also providerecommendations and suggestions, along with explanations and examples of how knowledgeassets should be classified. The thrust is to let managers manage and expect them to manage.

Implementation requires multi-faceted communications, starting with publishing the policy inpaper and digital formats. Regional centres were visited to describe the policy and explain whatis expected of employees. Then, the policy was officially announced to all staff by the Assistant

7

Deputy Minister. This was immediately followed by details of where to find relevantinformation and what to do. Finally, a four-step management approach was established: generalinvitation, friendly reminder, targeted contact, and one-on-one interview.

3.6 Project integration

Integrating KM projects can be illustrated with a scenario of preparing a briefing note in responseto a short turn-around request from the Minister’s office. The Briefing Note Database is the firststop for someone who is unfamiliar with an issue, needs to know the latest CFS position, or issimply updating a previous note. The Directory of Expertise and Skills can be used to findexperts who can help with new issues or new questions related to existing issues. Accessingregional libraries enables analysts to dig deeper and cite relevant scientific sources. Finally, theKnowledge Assets Inventory can be used to find unpublished material that can contribute topolicy analysis. What traditionally required a few days can now be done in a few hours.

4. Forestry Content

Content projects focus on using knowledge management approaches to solve forestry problemsin Canada. The projects are presented in the order in which they became operational. Development of content projects was led by domain experts, with some advice andrecommendations provided by the KM Division. The following examples represent only a smallsample of CFS content-based projects that use KM solutions.

4.1 Automated monitoring: Forest Fire in Canada • Problem: There was no national view of predicted fire danger or actual fire activity.

National resource mobilization was predominantly subjective. National-scale data wastoo massive to be processed with traditional methods.

• Solution: Develop a fully automated system to access weather data and large fire activityvia observation and satellite networks, integrate and process the data, produce nationalfire danger and fire activity maps, and disseminate the maps via the Web.

• Status: Operational (1994); this is the fourth generation of fire information systems.• URL: http://fire.cfs.nrcan.gc.ca/index_e.php

This system automatically acquires hourly observations from a national network of weatherstations and enters it into a GIS database. Using fire behavior prediction models, it transformsweather, fuel, and topographic data into fire behavior maps that are disseminated via the Web. Ahigher-resolution version of this system is used by many Canadian agencies to manage forestfires. This system has been adapted to map fire behavior in other countries, including Mexico,South-East Asia, and Russia.

Large- fire activity is monitored through satellites. Six satellite images are integrated into asingle geocorrected national mosaic. Fire “hot spots”are superimposed on the GIS database. Thisis used to produce map images that are distributed to the public via the Web. The common GIS

8

database allows fire agencies to integrate fire weather, fire behavior predictions, and large-fireactivity to support daily fire management decision making. This state-of-the-art system was thefirst of it’s kind in the world and has won awards for innovation.

4.2 Compiling national data: Compendium of Canadian Forestry Statistics• Problem: The cost of manually compiling and publishing national statistics had become

prohibitive. National tables could not be published until the last agency submitted data. Data had to be manually entered into analysis applications.

• Solution: Develop a Web-based template for submitting and compiling data. Updatetables as agency data are received. Publish the data in a down loadable digital format.

• Status: Operational (1997); the Compendium is no longer published in paper form.• URL: http://nfdp.ccfm.org/index_e.php

Web-based templates are completed by individual agencies as their annual data becomeavailable. This process greatly facilitates validation by reducing the number of calculation andincomplete data errors. The electronic database is updated and revised as agency data arereceived, thus insuring that the most up-to-date data are always available.

The Compendium of Canadian Forestry Statistics summarizes major statistical themes:inventory, harvest, fires, insects, products, silviculture, management expenditures, revenues, andpest control products used. An overview report is supported by dozens of tables and graphs thatare provided for each jurisdiction in Canada as well as at the national level. A key feature is anability to access and download numerical data in a variety of standard formats which users caneasily upload it into their applications.

4.3 Distributing knowledge: National Bookstore• Problem: CFS publications were distributed by six regional centers. Users had to send

separate mail requests to each center to obtain copies.• Solution: Develop a Web-based single-window site for ordering all CFS publications.

Enable direct downloading for digitized publications.• Status: Operational (2000); 17,000 CFS publications listed; 11,000 are down loadable.• URL: http://bookstore.pfc.cfs.nrcan.gc.ca/default.htm

The CFS Bookstore provides single-window Web-based access to all available documentspublished by the CFS. A 10-field search window provides considerable flexibility in locatingCFS publications. The initial search provides a list of documents and abstracts that satisfy thesearch criteria which can be refined as appropriate. Clicking on a document allows the user torequest a copy from the Web site. For electronic documents, the user may download thedocument directly. For print media, a “shopping cart” process enables users to order a papercopy via the Web. Registered users need only enter their identity to complete their order.

4.4 Enhancing Client Service: E-800 Service• Problem: Many common questions must be answered repeatedly, leading to redundancy

9

and duplication. A lack of links between query channels or responders leads toinconsistent responses.

• Solution: Develop a process to capture queries and responses from all channels. Developa process to publish Frequently Asked Questions on the Web. Develop a searchable Webdatabase of FAQs and responses.

• Status: Operational (2003); 1500 queries in the repository, 70 published questions &answers.

• URL: http://e800.cfs.nrcan.gc.ca/cfs/index.cfm?

The first page contains eight categories of questions. Selecting a category produces a list ofquestions that are automatically ranked so that the most frequent ones rise to the top of the list. A category may also split into sub-categories, such as disturbance - fire/insect. Questioners maycontinue to browse through the list or pose a keyword or plain language search. All FAQs arepublished in both English and French.

The key aspect of this Service is the set of underlying work processes that were developed tosupport it. One function involves assigning previously unanswered questions to appropriateexperts. Then, all queries and responses from all channels are captured and stored in a singledatabase (the warehouse). Frequency of occurrence guides a third function which edits andtranslates questions and answers into publishable form. The last process is to publish thequestions on the FAQ site and let them rise to their natural rank.

4.5 Preservation and learning: Forest Ecosystems of Canada• Problem: Thousands of forestry photographs taken over decades were at risk of being

lost. The photos were not catalogued and unavailable to anyone other than the originalphotographer.

• Solution: Develop a process to capture, catalogue, and archive the photographs. Developa web site to provide subject and map-based access to the photographs. Extend the site tosupport learning about Canada’s forest ecosystems.

• Status: Operational (2003); 14,000 photos archived.• URL: http://www.cfl.scf.rncan.gc.ca/ecosys/home_e.html

The “Snapshots of Quebec Forest Cover” provides access to more than 2,000 aerial photographsof forest ecosystems in Quebec, along with commentary. The photos can be searched via maplocation or up to six themes: type, stage, species, density, disturbance, or amplitude. A typicalsearch result lists metadata and a thumbnail of all photograph that match the search criteria. Clicking on any element yields the relevant photograph.

For learning, the site is divided into four categories: classification, disturbances, dynamics, andissues. Each category is sub divided into a number of topics. Each topic includes descriptive textoriented towards students, along with links to the photo database.

10

5. Conclusions

As with building a house, “one piece at a time” does not mean randomly choosing projects. There is a KM blueprint in place, in the form of a strategy and framework that shows all thenecessary pieces and how they fit together. There is a proper construction and assemblysequence, in that some things must precede others. Process projects are like the foundation andframing, in that they focus on organizational infrastructure by accomplishing one or more KMgoals across multiple content domains. KM processes are necessary regardless of the finaldesign, which is likely to evolve as organizations learn. Content projects are like the finish andfurnishings; in that they are what people see and relate to. They use knowledge managementapproaches to provide domain-specific goods and services to clients and to all Canadians.

Working at a project scale minimizes the risk of high-profile and high-cost failure. Multipleprojects further spread the risk, in that most are likely to succeed. Projects also minimize theneed for large up-front investment with returns that are difficult to demonstrate in advance. Projects can often be supported on the basis of near-term cost reduction and increased efficiency. It is also easier for people to see how projects relate to their specific knowledge work.

Over the long run, there is a potentially much larger benefit to a structured piecemeal approach.Culture change - which is at the heart of knowledge management - takes many years toimplement. Starting by attempting to change a culture has a high risk of failure. Change is tooscary; there is too much resistance; the goal is too nebulous. However, through the use of KMtools that help people do their work, culture gradually begins to change in response to the use ofthe tools without considerable overt hype or effort. Thus, KM by stealth would appear to havemany advantages and few disadvantages as a strategy for transforming organizations ofintelligent people into intelligent organizations that can succeed and thrive in the 21st century.

Bibliography

Buckman, Robert H. Building a Knowledge-Driven Organization, McGraw-Hill, New York,NY. 2004, 264 p.

Simard, Albert J., Managing Knowledge at the Canadian Forest Service, Natural ResourcesCanada, Ottawa, Ontario, 2000, 88p, http://bookstore.pfc.cfs.nrcan.gc.ca/default.htm

Simard, Albert J., Canadian Forest Service Access to Knowledge Policy and ImplementationGuidelines, Natural Resources Canada, Ottawa, Ontario, 2000, 30p,http://bookstore.pfc.cfs.nrcan.gc.ca/default.htm

Van Horne, Constance, Jean Marc Frayret and Diane Poulin. Knowledge Management in theForest Products Industry: the Role of Centres of Expertise, Computers and Electronics inAgriculture, 2005 47(3): 167-184, http://www.sciencedirect.com/