Knowledge and Process Management

Volume 15 Number 1 pp 59–71 (2008)

Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/kpm.299

& Research Article

A Process-Based Classification of
Knowledge Maps and Application
Examples
Martin J. Eppler*,y
University of Lugano (USI), Lugano, Switzerland

This paper discusses pragmatic ways of classifying knowledge maps to give an overview of their
application contexts and formats. In the paper, we show where and how the term knowledge
map has been previously used and what criteria must be met in a sound and useful knowledge
map classification that can support knowledge management processes. Various classification
principles are presented and discussed. A table then matches map formats to knowledge
management purposes and knowledge-related contents in order to serve as a selection and
organizing framework. Examples of some of the main types of knowledge maps are presented to
illustrate the varieties of knowledge mapping present in the classification. The paper concludes
by discussing its limitations and future research questions in the area of knowledge mapping.
Copyright # 2008 John Wiley & Sons, Ltd.

INTRODUCTION ful in eliciting (Hodgkinson et al., 2004) or referen-

cing implicit knowledge (Meyer, 1991; Sparrow,
An early step toward understanding any set of 1998) and thus can help to make knowledge more
phenomena is to learn what kinds of things there are in widely and easily available. The terms knowl-
the set—to develop a taxonomy. edge map (Eppler, 2002; Sparrow, 1998) or knowl-
(Herbert A. Simon) edge mapping (Wexler, 2001) have been used to
designate a wide variety of approaches to organize
The advantages of visual representations for the and structure knowledge sources, knowledge
field of knowledge management have long been application steps, insightful concepts, expert net-
recognized and discussed (Eppler, 2002, 2003; works, or communities of practice (Vail, 1999).
Newbern and Dansereau, 1995; Sparrow, 1998; Vail, These varieties of knowledge maps have so far
1999; Wexler, 2001; Wurman, 2001) and include a never been systematically classified and compared
better overview, a faster access, and a more efficient in terms of their characteristics, unique features, or
and memorable representation and communication application parameters (for a first tentative classi-
of knowledge assets (such as experts, practice fication see Eppler, 2002).
documents, communities, patents, etc.). Visual There are numerous benefits that can be achieved
representations have also proved particularly use- through a classification of knowledge maps: first, it
can provide a descriptive overview of the domain
(Bailey, 1994, p. 12) and can function as an inventory
*Correspondence to: Martin J. Eppler, University of Lugano or repository (ibid., p. 13) like a structured toolbox.
(USI), Via Buffi 10, 6900 Lugano, Switzerland.
E-mail: [email protected] In this way a classification can also become a
y
Chair of Information and Communication Management. problem solving heuristic (Dherbey, 2005, p. 68) that

Copyright # 2008 John Wiley & Sons, Ltd.

RESEARCH ARTICLE Knowledge and Process Management

relates possible mapping solutions to knowledge classifications. In this section, we thus review the
management challenges. Thus, a classification key requirements discussed in classification and
reduces the complexity inherent in choosing a categorization literature (Bailey, 1994; Bowker and
knowledge map format for a particular application Star, 1999; Dherbey, 2005; Lakoff, 1987; Minto, 1995;
context. As a further benefit, a map classification Wurman, 2001) in order to apply this research
helps to recognize the similarities and differences method adequately. A classification, according
among different types of knowledge maps. It helps to Bailey (1994) is the ordering of entities into
to compare different types of knowledge maps along groups or classes on the basis of their similarity.
pertinent criteria. As a side-benefit of developing a Classifications minimize within-group variance,
classification, one has to develop an exhaustive and maximize between-group variance (Bowker
description of the variables that define a knowledge and Star, 1999). In other words, a classification
map’s application context. Finally, a classification of should maximize the homogeneity within a group,
knowledge maps may also reveal new forms of as well as the heterogeneity among groups, thus
knowledge maps that so far have not been applied. facilitating analysis, organization, and assessment
The classification may systematically go beyond the (Bailey, 1994). Classification (according to Bowker
current state-of-the-art practice of knowledge map- and Star, 1999) can also be described as a
ping and show potential future formats. Having spatial-temporal segmentation of the world (or one
listed the benefits that can be expected from a aspect of it). Taxonomies, as a special kind of
systematic classification of knowledge maps, one classification, are tied to a purpose, in the context of
should also note the potential disadvantages of this this paper allocating the right type of knowledge
research approach. Such disadvantages are the map to the right type of knowledge management
focus on description, rather than explanation. problem. Bailey (1994, p. v) points out the difference
Classification may lead to reification (Bailey, 1994, between taxonomy and typology: whereas a
p. 15), that is to say to pretending that an ideal typology is conceptual, deductive, and based on
archetype does exist, when it is merely hypothetical. reasoning, a taxonomy is empirical, inductive, and
Tied to this criticism is the fact—relevant in many based on large sets that are examined and grouped.
knowledge management application contexts— The classification proposed in this paper is partly a
that classifications tend to be rather static and taxonomy derived from existing real-life knowledge
difficult to adjust as a domain changes and evolves. maps. In part, it is a typology, as it aims to point out
These disadvantages lead to the recommendation other possible forms of knowledge mapping that
that a classification system should not just rely on may not yet have been applied in real-life contexts.
one classification principle, but should propose A high-quality classification system that is fit for use
various, alternative classification criteria. In Section should be both consistent and manageable. Con-
‘‘Classification Principles and Typologies for sequently, it should meet at least two sets of criteria,
Knowledge Maps’’ of this paper, we will thus namely logic criteria and pragmatic criteria. In
propose different useful classification principles to terms of the logic or formal criteria that make a
structure the domain of knowledge mapping. classification sound, the classification has to have
unique classificatory principles in operation which
are not mixed at each level of abstraction or
REQUIREMENTS OF A CLASSIFICATION hierarchy (Bailey, 1994; Minto, 1995; Wurman,
SYSTEM 2001). This will ideally guarantee that the resulting
categories are mutually exclusive (non-overlapping)
Classification lies at the heart of every scientific field. (Minto, 1995), and that the classification system is
Classifications structure domains of systematic complete (the categories are collectively exhaustive),
inquiry and provide concepts for developing theories meaning that there are no items in the domain that
to identify anomalies and to predict future research cannot be assigned to a category (the classification is
needs. comprehensive). A classification should capture the
(Lohse et al., 1994) totality of phenomena supposedly contained within
it. In our case, there should not be knowledge maps
Before examining existing classifications of that cannot be classified within the proposed
knowledge maps (in Section ‘‘Prior Definitions system. Furthermore, a good classification assigns
and Typologies of Knowledge Maps’’) and propos- items to groups based on objective and stable
ing our own set of categories (in Section ‘‘Classi- attributes, so that there are no unclear category fits
fication Principles and Typologies for Knowledge of particular items. The labels for different groups
Maps’’), we should briefly examine the general (on a particular hierarchic level) are on the same level
rules or criteria that lead to valid and useful of abstraction (ibid.); the classification should not mix

60 M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management RESEARCH ARTICLE

specific (sub-) groups with very general ones. It PRIOR DEFINITIONS AND TYPOLOGIES
should also be made explicit what lies beyond the OF KNOWLEDGE MAPS
categorization scope: the boundaries of the classi-
fied domain or area should be clear (i.e., one should The map is the territory if people treat it as such.
give inclusion or exclusion criteria). With regard to (Karl Weick)
the pragmatic criteria that make a classification more
ergonomic to use, one must pay attention to In this section, we review prior definitions and
self-explanatory informative category names, as well classifications of knowledge maps and point out
as take into account the total number of categories their limitations and the resulting need for new
that should not result in an overly heavy cognitive classifications. The term knowledge map has so far
load for the targeted users. The process of item been used in at least seven different scientific
attribution to a category is made easier, if each communities: in education studies or in researching
category has a typical representative, a so-called instructional methods, such as mind maps, concept
prototype member that can act as a mnemonic device maps, and related graphic learning tools (Tergan
for the category and thus makes it more memorable and Keller, 2005), in organization studies (Huff and
(Lakoff, 1987). The granularity of the classification Jenkins, 2002; Huff, 1990) and in requirements
should be in line with its intended use requirements engineering (Browne and Ramesh, 2002) where the
(not more specific than actually needed). In term designates the elicited, visualized mental
summary, we can state that an ideal (sound and models of managers or IT users, in decision analysis
useful) classification should have the following to elicit crucial information (Browne et al., 1997) in
properties, of which the first six are formal (or information retrieval to designate interactive search
soundness) criteria and the subsequent four are result displays and search result browsing inter-
pragmatic (or usability) ones: faces (Coyne, 1995), in decision support systems to
designate, among other things, the informative
1. it consists of mutually exclusive categories (groups graphic rendering of decision variables (Smelcer
that do not overlap), and Carmel, 1997) in artificial intelligence where it
2. that are collectively exhaustive (i.e., together the can designate the conceptual representation of an
groups cover the entire classified domain), expert domain (Gordon, 2000), and in the knowledge
3. that are based on stable and objective grouping management community (Burnett et al., 2004; Vail,
criteria (in order to unequivocally assign an item 1999). In the context of knowledge management, a
to a category in a classification), knowledge map generally designates an overview
4. that have category names on a consistent level of on a collection of knowledge-related contents. A
abstraction (per hierarchic level), knowledge map typically consists of two main
5. based on one explicit, consistent, and informative parts: a ground or background layer which represents
classification principle per level of hierarchy, the context for the mapping, and the individual
6. for a clearly specified and delineated topic area or elements that are mapped within this context.
domain, The elements which are mapped onto such a
7. where the categories have self-explanatory, shared context range from experts, project teams,
informative category names, or labels, or communities of practice to more explicit and
8. and contain typical, representative (prototype) codified forms of knowledge such as white papers
members for each group in the classification, or articles, patents, lessons learned (e.g., after action
9. resulting in a well-organized system that does reviews or project debriefings), events (i.e., train-
not overload the users as it contains an adequate ings), databases, or similar IT applications such as
amount of groups that can still be managed by expert systems or simulations. Knowledge maps
short-term memory (the granularity of the group these elements to show their relationships,
distinction does not exceed the level of detail locations, or other attributes. Knowledge maps
necessary for the envisioned task that the answer questions such as: how do I find relevant
classification supports), knowledge, how can I judge its quality, how can I
10. a system that is hence understandable and usable make sense of its structure, and how do I go about
by the envisioned user groups. applying or developing it? Definitions of knowl-
edge maps that we have found in the literature
The 10 criteria compiled above can provide follow this logic. Vail, for example, defines a
guidelines for the assessment of current knowledge knowledge map as follows (Vail, 1999, p. 10):
map classifications and for the development of new
classifications. The former is presented in the next A knowledge map is a visual display of captured
section, the latter in the subsequent section. information and relationships, which enables the

Classifying Knowledge Maps 61

DOI: 10.1002/kpm
RESEARCH ARTICLE Knowledge and Process Management

efficient communication and learning of knowledge by of cognitive structures (Huff and Jenkins, 2002).
observers with differing backgrounds at multiple The problem with this classification is that it is
levels of detail. The individual items of knowledge not based on one consistent classification
included in such a map can be text, stories, graphics, principle and not always applicable to knowledge
models, or numbers. [...] Knowledge mapping is management.
defined as the process of associating items of Novins (1997) distinguishes among three types of
information or knowledge (preferably visually) in knowledge maps, namely pointer models (pointing
such a way that the mapping itself also creates to the correct source, usually a person), linkage
additional knowledge. models (adding some meta information on the sou-
rces), and solution models (relating knowledge areas
A more recent definition by Renukappa and Egbu to business problems). Pointer models are knowledge
(2004) also stresses relationships, but adds the source maps that typically map experts. They can be
important element of maps referring also to tacit geographic or organized by topic. Linkage models
knowledge. It also highlights the important notion provide more visual context on how the referenced
of knowledge dynamics, next to knowledge stores knowledge can be used, for example by linking
or repositories: knowledge to a visualized business process. Still
more meta-information on the referenced knowl-
A knowledge map is a navigation aid to both explicit edge is represented in solution models. In this type of
and tacit knowledge, showing the importance and the dense and informative map, descriptive and pre-
relationships between knowledge ‘‘stores’’ and the scriptive elements are mixed. While this is a useful
dynamics. categorization, its groups are not fully mutually
exclusive and the classification only focuses on the
According to Ernst & Young (Novins, 1997), to main functions and content types of knowledge
take a practitioner’s definition, a knowledge map is maps and neglects other application parameters
a place to find the source of answers, a method and such as graphic format, scope, medium, creation
format for collecting and communicating where mode, or required skill level.
knowledge resides and is lacking, typically within an These existing classifications and distinctions are
organization, a visual representation of the knowl- relevant, but they may be limiting the potential of
edge content areas. Based on these typical defi- knowledge maps to too few areas. It may be
nitions, we can conclude that the minimal criteria beneficial to explore new, alternative, and concur-
for a knowledge map are that it is a graphic overview rent ways of classifying knowledge maps in order to
and reference of knowledge-related content that serves a explore and extend their application potential
knowledge management-related purpose. beyond the currently implemented or envisioned
Let us next look at how this domain has been solutions. Developing multiple classification
structured so far, that is to say which types of schemes may also improve our understanding of
knowledge maps have already been distinguished. the application parameters of different forms of
In knowledge management, the classical carto- knowledge maps. Such new classifications—that
graphic map types (aimed at representing infor- strive to meet the 10 categorization criteria listed
mation about a geographic territory) are less fruitful previously—are explored in the next section.
(see Peterson, 1995). Eppler (2002) proposes a
simple knowledge map typology based on knowl-
edge management tasks, namely knowledge creation
and development maps, knowledge identification CLASSIFICATION PRINCIPLES AND
maps, knowledge assessment maps, and knowl- TYPOLOGIES FOR KNOWLEDGE MAPS
edge application maps. The main problem with
this classification is that it is not comprehensive, You do not understand anything, until you under-
versatile, or precise enough to be of general use in stand it in more than one way.
knowledge management. (Marvin Minsky)
A different, more abstract set of map categories is
used by Anne S. Huff in her anthologies on the topic In the study of taxonomies, there is a general
of mapping strategic thought and knowledge (Huff rule that a classification should always be based
and Jenkins, 2002; Huff, 1990). In her mapping on key characteristics of its items, but—according
typology, she focuses on cognitive maps and to Bailey—there is no proven rule to find these
distinguishes the following map types: text and attributes (Bailey, 1994, p. 2). Nevertheless, cogni-
language analysis maps, classification maps, net- tive linguist George Lakoff provides insightful
work maps, conclusive maps, and schematic maps directions in his theory of categorization. A

62 M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management RESEARCH ARTICLE

classification according to Lakoff (1987) can be one While the purpose describes the knowledge man-
of four types. These types are: agement task supported by the map (frequently tied
to an application context), the content dimension
1. Purposive: categorizing by intended use; in our
describes the elements that are contained and refer-
context classifying maps by the knowledge
enced in a knowledge map. Usually, a knowledge
management purpose they serve.
map contains only one kind of content, although
2. Perceptual: categorizing by common format/look;
there are maps that contain information on experts,
in our case, by the graphic format of a knowledge
documents, communities, and databases in parallel.
map.
Table 1 shows sample knowledge map types based
3. Functional: categorizing by personal use or type
on these primary classification principles. The items
of the content of the knowledge map.
in the open lists are intended to serve as illustrative
4. Motor-activity: based on physical interaction with
examples and not as exhaustive options.
the content; in our case, the medium or
Other possible, but potentially less useful, stable,
application context of a knowledge map.
or objective classification principles include the
managerial application domain or functional area
These classification principles can also be found in
(e.g., maps for project management, strategy,
seminal taxonomies of visual representations: Shnei-
quality management, procurement, risk manage-
derman’s task by type taxonomy of visual repres-
ment, finance, production, etc.), the amount of
entation formats (Shneiderman, 1996) suggests both
resources (time, money) associated with generating,
application purpose or functionality (task) and content
updating or using a map, the required skill level of
(type) as classification principles. Lohse et al. (1994)
map users (from novice maps to expert maps), the
conclude that extant taxonomies of graphs and
size of the map (10 entries vs. 10 000 entries), and the
images in general are either functional, that is, by
medium of the map (paper, poster, or digital).
purpose, or structural, that is, by graphic form,
Having presented different possible classification
although there are also model-based taxonomies
principles and resulting typologies, we can now
(Chi, 2000; Tory and Möller, 2002), as well as other
combine the most relevant ones into a matching
possible classification criteria, such as social context,
matrix that can serve as a first, generic selection
or cognitive process, see Blackwell and Engelhardt
guide for knowledge maps (that evidently requires
(2002). An example of a functional taxonomy has
adaptation for specific application contexts). For
been developed by Tufte (1990), structural classifi-
this matrix, we chose the use of the knowledge
cations have been developed by Bertin (1973), Horn
map in knowledge management (the knowledge
(1999), and Rankin (1990). From these prior
management process) as the dominant feature, as
approaches, we derive our primary knowledge
well as the format of the map and the level and
map classification principles, which are by purpose,
content type of the map (the former two as table
by graphic form, by content, by application level, and
axes, the later two as table entries). These are
by creation mode. We believe that these classification
considered to be the most relevant dimensions as
principles are relevant for a pragmatic taxonomy of
they guide the actual implementation process:
knowledge maps as they all relate to the actual
knowledge content needs to be adequately
application of knowledge maps. The pragmatic logic
represented in a graphic format for a particular
of these classification principles can be illustrated
knowledge management process at a certain level.
(see also Wexler, 2001) by converting them into
In this sense, the following matrix (Table 2) can
questions or interrogatives; choosing a particular
serve as a starting point and discussion template for a
type of knowledge map necessitates answering a
deliberation on which kind of knowledge map may
number of key questions, namely:
be useful for a given knowledge management
1. Which knowledge management purpose do I process or challenge.
want to achieve with the map? (The ‘‘why?’’ of The table represents the following reasoning
the map.) (along seven KM processes): for the creation of
2. Which kind of content about knowledge do I want new knowledge, knowledge maps can help in the
to represent in the map? (The ‘‘what?’’ of the generation of new concepts by representing emer-
map.) ging topics in cartographic maps, in diagrams, or
3. Who should use the map in which context or through visual metaphors (as well as through
situation and at what level? (The ‘‘for whom?’’ sketching not included in this classification). While
and ‘‘when?’’ of the map.) tables could be used in this phase, they might not
4. Which graphic form should be used and who can represent a rich enough structure to uncover new
create the map in what way? (The ‘‘how?’’ and insights or elaborate concepts—they could, how-
‘‘who?’’ of the map.) ever, be used to highlight the possible combination

Classifying Knowledge Maps 63

DOI: 10.1002/kpm
RESEARCH ARTICLE Knowledge and Process Management

Table 1 Knowledge map classifications

A. Classifying knowledge maps by intended purpose or KM process (‘‘why?’’)

1. Knowledge creation maps: illustrate the planned steps to develop a certain (organizational) competence or
create new knowledge (i.e., a technology road map)
2. Knowledge assessment or audit maps: illustrate the evaluation of certain knowledge assets graphically,
for example, by a 2
2 matrix (axes: current ability and future importance)
3. Knowledge identification maps: provide a graphic overview on knowledge assets (experts, patents, practices)
and points to their locations/coordinates
4. Knowledge development or acquisition maps/learning maps
(a) Learning overview and learning path maps
(b) Learning content structure maps
(c) Learning reviewing/repetition maps
5. Knowledge transfer, sharing, or communication maps: show who transfers knowledge to whom
6. Knowledge application maps: show which knowledge is necessary for carrying out certain processes or steps
in a single process
7. Knowledge marketing maps: can be used to signal competence to the public in a certain domain
B. Classifying maps by their content (‘‘what?’’)
I. By (digital and analog) content formats: 1. websites (incl. blogs, portals, homepages), 2. documents
(incl. books), 3. databases or repositories, 4. learning objects or online courses (or modules),
5. other file formats (e.g., sketches, drawings)
II. By content types: 1. methods, 2. processes, 3. experts (incl. groups), 4. organizations/departments/institutions,
5. lessons learned/experiences, 6. skills and competencies, 7. concepts, 8. events, 9. patents, 10. knowledge or
communication flows or relationships, 11. interests or knowledge needs
C. Classifying maps by the application level (‘‘who?’’)
1. Personal knowledge maps (visualizing one’s own skills or expert contacts, see Eppler and Sukowksi, 2000 or
Burnett et al., 2004)
2. Dyadic knowledge map (to support knowledge creation, transfer, or assessment between two people)
3. Team knowledge maps (visualize the skills present or needed in a project team, like the T-matrix,
see Eppler and Sukowksi, 2000)
4. Departmental knowledge maps
5. Community knowledge maps
6. Organizational knowledge maps
7. Inter-organizational/network knowledge maps
D. Classifying knowledge maps by graphic form (‘‘how?’’)
I Table-based format (for an example see Heng, 2001)
1. Person by skills table
2. Skill area by people table
3. People by documents
4. Team by project experience table
II. Diagrammatic format
1. Structure diagrams
(a) Venn diagram, (b) concentric circles (with or without segments), (c) matrix (i.e., 2
2),
(d) network diagram, (e) mind map, (f) concept map (Tergan and Keller, 2005),
(g) cognitive map (Huff and Jenkins, 2002), (h) strategy map, (i) fishbone
2. Process diagrams (Galloway, 1994)
(a) Timeline, (b) swim lane chart, (c) flow chart, (d) event chain, (e) critical path method,
(f) Gantt chart, (g) cycle chart, (h) decision tree, (i) value chain, (j) flight plan (Eppler and Sukowksi, 2000)
III. Cartographic format
1. Geographic map: globe/continent/land/island/region, 2. informational map: park, 3. tube/metro
(Burkhard and Meier, 2005) map, 4. galaxy/stars, 5. sea/ocean, 6. building/architectural map
IV. Metaphoric format
(a) From the natural realm: 1. tree, 2. iceberg, 3. canyon, 4. mountain, 5. river, 6. . .
(b) Man made artifacts: 1. house, 2. temple structure, 3. radar screen, 4. bridge, 5. race track, 6. . .
E. Classifying maps by their creation method (‘‘how?’’ and ‘‘who?’’)
1. Maps that are automatically and dynamically generated by the computer (such as self-organizing maps,
see Kohonen, 2001)
2. Maps that are semi-automatically generated (automatically assembled and then optimized by analysts)
3. Maps that are designed once by domain and mapping experts and then used in the same way by all users
4. Maps that are iteratively created, modified, or extended by the map user(s) themselves
(community generated maps)

64 M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management RESEARCH ARTICLE

Table 2 A possible matching matrix for knowledge map parameters

K map format/knowledge I Table II Diagrammatic III Cartographic IV Metaphoric

management process/purpose format format format format

1. Creation of knowledge M, L, C M, L, C M, L, C
1–3 1–3 1–3
2. Assessment or audit of knowledge E, F, S E, F, S
1–5 1–5
3. Identification of knowledge M, E M, E, F M, E, F M, E, F
1–4 1–7 1–7 1–7
4. Development of knowledge M, S, C M, S, C M, S, C
1–7 1–7 1–7
5. Sharing, transferring, M M, L, C, S, F M, L, C, S M, L, C, S
communication of knowledge
2–7 2–7 2–7 2–7
6. Application of knowledge M, L, S M, L, S M, L, S M, L, S
1–7 1–7 1–7 1–7
7. Marketing of knowledge M, E, C, S M, E, C, S M, E, C, S
4–7 4–7 4–7

Knowledge map content types: M, methods (procedural knowledge, know-how); E, experts, organizations, groups, institutions etc.
(know-who, knowledge carriers); L, lessons learned, and experiences (know-why); C, concepts (declarative knowledge, know-what);
F, flows or relationships (i.e., communication flows, collaboration relations); S, skills and competencies (i.e., capability maturity levels,
expertise levels, core competencies, etc.). Application Levels: 1, personal; 2, dyadic; 3, team; 4, dept.; 5, community; 6, org.; 7, network.

of skills or concepts. To assess the knowledge of the task of rating knowledge assets. To easily identify
experts, groups, or departments, one can employ (as knowledge, tables usually do not provide a concise
shown in Figure 1) simple tables. Cartographic overview. Cartographic, diagrammatic, or meta-
maps cannot (because of their loose structure) be phoric maps can provide richer means of overview
easily used for ratings. Diagrams, such as matrices, in this case. In order to develop new knowledge or
however, can provide visual ratings easily. Visual acquire new knowledge through learning, one can
metaphors, finally, may be too playful and open for rely on the didactic power of cartographic learning

Figure 1 A table-based departmental knowledge assessment map of an IT consultancy

Classifying Knowledge Maps 65

DOI: 10.1002/kpm
RESEARCH ARTICLE Knowledge and Process Management

Figure 2 A diagrammatic, knowledge application map of medium-sized market research firm (Eppler, 2003)

Figure 3 An institution-centered diagrammatic knowledge identification map visualizing researchers in the area of e-learning

maps (such as trail maps outlining learning steps), already known. This reasoning is equally applicable
diagrams (such as concept maps or mind maps), or to knowledge sharing maps, although in this case,
metaphoric maps that convey additional insights tables may provide a simple format to share or
about the content, or relate what is new to what is communicate the main steps of a method. The use or

66 M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management RESEARCH ARTICLE

Figure 4 A Venn diagram-based knowledge identification map listing and structuring institutions that have expertise in the area of
health communication

application of knowledge can be supported by any of APPLICATION EXAMPLES

the above means: through tables outlining sequen-
tial action steps and corresponding documents, To illustrate key types from these classifications, we
concepts, or experts, as well as through cartographic provide real-life, interactive online knowledge map
trail maps, diagrammatic process maps, or meta- examples in this section (Figures 1–7). In terms of
phoric depictions (i.e., a ladder or a road) that show purposes, the examples include knowledge assess-
how to accomplish a goal, by referencing concepts, ment, application, identification, marketing, and
documents, or experts. For the final knowledge acquisition maps. In terms of content, they refer to
management task examined in this paper, namely experts, tools and methods, documents, institutions,
knowledge marketing, the attention of the target concepts, applications, and websites.
group is a crucial element. In order to draw the The maps range from the departmental to the
attention of potential clients to an organization’s inter-organizational level. None of the examples are
methods, experts, concepts, or skills, novel, original, automatically generated maps. As mentioned in (Vail,
and even surprising ways of representing the 1999) employing automatic mapping techniques
offered knowledge need to be employed. Hence, (such as e-mail traffic or questionnaire-based social
the table format may prove not to be attractive network analysis software) foregoes, the chance
enough. There is, however, a trade-off between map of using the collaborative mapping process itself as
novelty and clarity that has to be managed.1 a communicative sense-making and identity- or
In this way, the application parameters for a consensus-building process for the involved com-
specific knowledge map can be systematically munities of practice. The knowledge maps that are
gathered for a specific application context. The depicted in Figures 3, 4, and 7 can be viewed and
reasoning outlined above may, however, have to be explored interactively at http://www.unisi.ch/
adapted for specific application contexts. knowledgedomainmap.htm.
Table 3 compares these seven examples in terms
of their main parameters, in order to illustrate the
1
Chillimind (www.chillimind.de), a mobile commerce company, represented spectrum of knowledge maps applied
has marketed its skills through an imaginative cartographic map
of mobile commerce concepts. Upon first glance, the map seems in our projects.
to depict an island. Only upon close observation can one detect The table reveals that several of the combinations
that it is actually a jostle of key technologies, business models, from the matching table have been implemented,
and technical terms from the realm of mobile business. The same
holds true for KLM’s map for its alliance partners, outlining its but that there may be many other feasible
partnering strategy. combinations that are still left to explore and that

Classifying Knowledge Maps 67

DOI: 10.1002/kpm
RESEARCH ARTICLE Knowledge and Process Management

Figure 5 A metro-style knowledge identification map documenting the experiences, experts, and documents of a 3-year project
(Eppler, 2003)

may provide tangible benefits for current and suggests how different map formats can be used for
future knowledge management challenges. Some different purposes and contents. The lack of
of these challenges are discussed in the following empirical validation of this matching is a weakness
conclusion. of this paper, although the template was presented
as a generic starting point for application dis-
cussions, rather than a final result. Subsequent
CONCLUSION AND LIMITATIONS research should nevertheless demonstrate through
evaluation studies whether this matching is indeed
Sapientis est ordinare. It is the function of the wise correct and under what circumstances.
man to order. A further open research question regards the
(Aristotle) development of prototypes for some of the stipulated
map types in different industry settings, including
In this paper, we have made first steps toward a follow-up evaluations. Related to this question is
pragmatic—that is to say consistent and useful— another research endeavor, namely to match knowl-
taxonomy of knowledge maps that can be used in edge map types with adequate information tech-
devising knowledge management solutions. Our nology applications.2 A final open research route in
approach has been based on rules and guidelines of this rich domain of inquiry concerns trade-offs
high-quality classifications. As a main contribution,
we have shown various ways in which knowledge 2
So far, we have used the following IT applications to manually
maps can be classified. The benefit of such multiple create (which implies maintainability issues) online interactive
classification principles lies in sensitizing managers knowledge maps: www.lets-focus.com, www.inspiration.com,
and researchers for the application parameters and www.visio.com, www.mindmanager.com. There are, however,
also tools available that can automatically generate knowledge
requirements of different knowledge maps formats. maps, such as the solutions of Aurigin, Autonomy, or Semio, to
We have proposed a tentative matching table that name but three suppliers.

68 M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management RESEARCH ARTICLE

Figure 6 An animated, cartographic knowledge development map (learning path or file rouge map) from the www.swissling.ch
project (Armani and Rocci, 2003)

Figure 7 A metaphoric knowledge identification map of a European Research Center (EJO ¼ European Journalism Observatory)

among map types present in the classification: maintainable. Identifying and analyzing such
While some knowledge map formats (such as application trade-offs in knowledge map types
manually developed metaphoric expert identifi- seems like a highly relevant future research area. In
cation maps or diagrammatic knowledge structure conclusion, it thus seems that we have only begun to
maps) may be useful for communication purposes, chart this intriguing research territory in the
they may not be highly scaleable or easily knowledge management domain.

Classifying Knowledge Maps 69

DOI: 10.1002/kpm
 70
RESEARCH ARTICLE

Table 3 A comparison among the seven maps using the parameters of the classification

KM process Map format Map content Mapping level Main benefits Industry context

1. Knowledge assessment Table: domain by expert Experts Department Training planning, Telecom
staff allocation
2. Knowledge application Diagram: cycle chart Methods Organization Allocation of methods Market research &
and knowledge marketing and tools along a business consulting
process, documentation
3. Knowledge identification Diagram: Experts Inter-organizational Highlighting experts University/research
concentric circles and contacts & development
4. Knowledge identification Diagram: Venn diagram Institutions Inter-organizational Highlighting experts and University/research
contacts, fostering domain & development
understanding
5. Knowledge identification Cartographic: Experts, documents, Inter-organizational documentation of Corporate-University
underground map applications, knowledge during collaboration
websites, databases, a project
publications
6. Knowledge development Cartographic: Concepts, learning Inter-organizational Building knowledge University/e-learning
island map content/modules about linguistic theories
through an overview
and sequence
7. Knowledge identification Metaphor: bridge Experts Inter-organizational Identifying experts and University-practice
contacts in a collaboration
research domain

M. J. Eppler
DOI: 10.1002/kpm
Knowledge and Process Management
Knowledge and Process Management RESEARCH ARTICLE

ACKNOWLEDGEMENTS Heng MSH. 2001. Mapping intellectual capital in a small

manufacturing enterprise. Journal of Intellectual Capital
Funding for this research has been provided by the Tasso 2(1): 53–60.
Horn RE. 1999. Visual Language. MacroVue: Lexington.
Foundation.
Hodgkinson GP, Maule AJ, Bown NJ. 2004. Causal cog-
nitive mapping in the organizational strategy field: a
comparison of alternative elicitation procedures.
Organizational Research Methods 7(1): 3–26.
REFERENCES Huff A (ed.). 1990. Mapping Strategic Thought. Wiley: New
York.
Armani J, Rocci A. 2003. Conceptual maps in e-learning. Huff A, Jenkins M (eds). 2002. Mapping Strategic Knowl-
How map based interfaces help the contextualization of edge. Sage Publications: London.
information and the structuring of knowledge. Infor- Kohonen T. 2001. Self-Organizing Maps. Springer: New
mation Design Journal R Document Design 11(2/3): York.
171–184. Lakoff G. 1987. Women, Fire, and Dangerous Things. What
Bailey KD. 1994. Typologies and Taxonomies: An Introduc- Categories Reveal About the Mind. University of Chicago
tion to Classification Techniques. Sage: Thousand Oakes. Press: Chicago.
Bertin J. 1973. Sémiologie Graphique. Mouton: Paris. Lohse G, Biolsi K, Walker N, Rueter H. 1994.
Blackwell AF, Engelhardt Y. 2002. A meta taxonomy for A classification of visual representations. Communi-
diagram research. In Diagrammatic Representation and cations of the ACM 37(12): 36–49.
Reasoning, Anderson M, Meyer B, Olivier P (eds). Meyer AD. 1991. Visual data in organizational research.
Springer: Berlin; 47–64. Organization Science 2(2): 218–236.
Bowker G, Star SL. 1999. Sorting Things Out: Classification Minto B. 1995. The Pyramid Principle, Logic in Writing and
and Its Consequences. MIT Press: Cambridge. Thinking. Pitman Publishing: London.
Browne GJ, Ramesh V. 2002. Improving information Newbern D, Dansereau FD. 1995. Knowledge maps for
requirements determination: a cognitive perspective. knowledge management. In Knowledge Management
Information & Management 39: 625–645. Methods: Practical Approaches to Managing Knowledge,
Browne GJ, Curley SP, Benson PG. 1997. Evoking infor- K Wiig (ed.). Schema Press: Arlington. 157–180.
mation in probability assessment: knowledge maps and Novins P. 1997. Knowledge Representation, Presentation
reasoning-based directed questions. Management held at MKO Semi-Annual Conference, September 8.
Science 43(1): 1–13. Peterson MP. 1995. Interactive and Animated Cartography.
Burkhard R, Meier M. 2005. Tube map visualization: Prentice Hall: New Jersey.
evaluation of a novel knowledge visualization appli- Rankin RA. 1990. A taxonomy of graph types. Information
cation for the transfer of knowledge in long-term pro- Design Journal 6(2): 147–159.
jects. Journal of Universal Computer Science 11(4): Renukappa SH, Egbu CO. 2004. Knowledge mapping:
473–494. concepts and benefits for a sustainable urban environ-
Burnett S, Illingworth L, Webster L. 2004. Knowledge ment. Proceedings of the 20th Annual Conference Associ-
auditing and mapping: a pragmatic approach. Knowl- ation of Researchers in Construction Management
edge and Process Management 11(1): 25–37. (ARCOM) 2004, September 1–3, Farzad Khosrowshahi,
Chi EH. 2000. A taxonomy of visualization techniques University of Herriot-Watt, Edinburgh, UK; 905–914.
using the data state reference model. Proceedings of Shneiderman B. 1996. The eyes have it: A task by data type
the IEEE Symposium on Information Visualization; 69– taxonomy for information visualizations, Proceedings of
75. 1996 IEEE Visual Languages, Los Alamos, CA, IEEE;
Coyne R. 1995. Designing Information Technology in the 336–343.
Postmodern Age, From Method to Metaphor. MIT Press: Smelcer JB, Carmel E. 1997. The effectiveness of different
Cambridge. representations for managerial problem solving: com-
Dherbey GR. 2005. Connaissance des choses naturelles et paring Tables and Maps. Decision Sciences 28(2):
des affaires humaines chez Aristotle. In La Connaissance 391–420.
des Choses, Samana G (ed.). Ellipses: Paris; 55–70. Sparrow J. 1998. The role of physical representations in
Eppler M. 2002. Making knowledge visible through knowledge elicitation. In Knowledge in Organizations,
knowledge maps. In Handbook on Knowledge Manage- Sparrow J (ed.). Sage: Thousand Oaks; 51–78.
ment, Holsapple CW (ed.). Springer: Berlin (Heidel- Tergan SO, Keller T (eds). 2005. Knowledge and Information
berg/New York); 189–206. Visualization. Springer: Berlin.
Eppler M, Sukowksi O. 2000. Managing team knowledge: Tory M, Möller T. 2002. A model-based visualization
core processes, tools and enabling factors. European taxonomy, Working Paper, School of Computing
Management Journal 18(3): 334–341. Science, Simon Fraser University.
Eppler M. 2003. Managing Information Quality: Increasing Tufte ER. 1990. Envisioning Information. Graphics Press:
the Value of Information in Knowledge-Intensive Products Cheshire.
and Processes. Springer: Berlin. Vail EF. 1999. Mapping organizational knowledge. Knowl-
Galloway D. 1994. Mapping Work Processes. ASQC Quality edge Management Review 2(8): 10–15.
Press: Milwaukee. Wexler M. 2001. The who, what, why of knowledge map-
Gordon JL. 2000. Creating knowledge maps by exploiting ping. Journal of Knowledge Management 5(3): 249–263.
dependent relationships. Knowledge-Based Systems 13: Wurman RS. 2001. Information Anxiety 2. Macmillan
71–79. Publishing: Indiana.

Classifying Knowledge Maps 71

DOI: 10.1002/kpm