posted 1 May 1999 in Volume 2 Issue 8
Real & Virtual: Best of breed
environments for knowledge sharing in communities
In this article Eelco Kruizinga & Timo Kouwenhoven discuss the challenges that arise when one tries to introduce knowledge sharing in communities. They concentrate their thoughts on the transfer of explicit or codified knowledge in communities, as they believe that this category of knowledge creates many opportunities within virtual environments. A framework for understanding these challenges is offered, and the authors present a number of successful combinations of real and virtual components within a community support environment.
More and more organisations and clusters of organisations and even individuals show a pattern of loosely coupled entities1 that share their knowledge to be able to add value to their environment. Often, there is no formal obligation to do so. There is only the common interest or common practice that binds the people in these networks of knowledge workers. These networks are hard to pin down: there is no employee list, there is often no project account for them in the books of organisations, people work relatively independent of each other and indeed, there is no formal contract between the members of the community that enables individuals in that community to rely on each other’s responsibility.
Although knowledge sharing in communities of practice is generally seen as an opportunity for leveraging potential in organisations, those that seek to support this sharing with real and virtual environments are confronted with the following challenges: How can one make loosely coupled groups of people capable of sharing? People are normally willing to share (Szulanski, 1996), given a proper value proposition for the individual, but because they’re often not capable in terms of skills and facilitation the potential for sharing is left unused. What impact does this implication have on the design of real and virtual environments for sharing within communities?
How can one capture enough context of shared items for people to be able to understand and use? Sharing of lessons learned is only sensible in those circumstances where enough context is conveyed for the receiver of a lesson learned to gauge the applicability of the lesson in his own situation. For example, lessons learned like: ‘we learned that we should be friendly for the customer’ can be particularly meaningful for the author of this lesson, but can be of no meaning and therefore of no value to a user of this lesson, because the context of the item is missing altogether (in a sense, ‘the making of’ is often far more instructive than the result).
How can one capture and use content on a JIT basis? One of the lessons of early knowledge management is that any intervention that adds extra tasks for stressed knowledge workers is bound to fail, so it is essential to embed knowledge capture and use in the daily work cycle. How can real and virtual support environments implement this requirement? In essence, every potential user of the content that is embodied in a community weighs the cost of ignorance versus the cost of finding out.
How can one secure the quality of content? Basically, quality of content is one of the key sustaining factors for the success of a knowledge sharing community. Should one build on the self-regulating processes within the group or should one opt for an explicit programme for the management of quality of content?
How should one structure content in order to safeguard retrieval of content? One could opt for a fully unstructured set of content carriers (in most cases documents), thereby in general alleviating the life of content producers. It is easy to place documents into the knowledge sharing environment (be it real or virtual), but for ‘readers’ of that content it is hard to find things and to see relations between content carriers. If one opts for a fully structured set of content carriers (with the set obeying for instance a rigid document model), the burden is replaced from the reading audience to the author community. For virtual environments this trade-off has repercussions for the use of intelligent technologies for document indexing and retrieval.
How should one present content for people to be able to navigate? Large collections of content are hard to navigate for users if no guidance is given. This guidance ideally follows the mental models that content users have of their work and the use of knowledge therein. The design of user interfaces for virtual shared environments therefore should be made with these mental models in mind. This design advice also holds for supporting navigation by human intervention (i.e. knowledge stewards, librarians and other human help for knowledge workers to navigate content). These roles should be in sync with the structure of work processes of their customers.
Coupling real &virtual components to build community environments
The essence of sharing codified knowledge is transfer of content from authors to a ‘reading’ audience and then receiving feedback from that audience, thereby closing the loop. A model for understanding the elements of an environment for supporting this transfer process is presented in Figure 1. The central element is the virtual community memory. This is an electronic shared repository of the codified knowledge (e.g. insights, ideas, lessons learned, best practices) and references to knowledge (e.g. who knows what or where to find it?). It is also a virtual place where community members can discuss the elements in the repository.
The community memory is encapsulated within a layer of facilitating processes, such as the knowledge market workshop, after action reviews, knowledge stewards, moderators of discussion groups or a sharing reward scheme.
Normally, intranets can be seen as the vehicle for the virtual community memory. We have to be careful however to see intranets as a panacea for bridging the gap between authors and readers in a knowledge sharing community. Many intranets positioned as a tool for knowledge sharing in reality act as ‘knowledge attics’. Content is haphazardly captured and used. In order to be successful, these solutions should move to a much more active knowledge pump in terms of content collection and dissemination. For more information on the content processing modes of virtual memories, see Van Heijst et al. (1997).
Mediating between community members and the virtual environment: real facilitation
The mediating layer presented in Figure 1 is a crucial element in the migration from a knowledge attic to a knowledge pump type of support environment for communities. We think that the design of the components in this mediating layer largely depends on:
1. The binding factors in the community: In general, the facilitation components should compensate for any weak or missing bindings (e.g. speaking different languages, being employed by different organisations), and build on the strong bindings in the community (e.g. being part of the same project, having the same professional background).
2. The level of codification of the knowledge held in the community: In general, with less codified knowledge, common exchange formats for this knowledge are lacking. Traditional knowledge elicitation has to take place in the mediating layer before entering the virtual community environment.
3. The opportunities of information technology: The mediating layer can become ‘thinner’ with the use of modern information technology. We will discuss some options in this article.
As we look at the facilitating layer between community members and the community memory, we can see the following successful components.
In the formative stages of a community, workshops that seek to chart the knowledge market are a quick means for getting members acquainted with each other. One workshop design is to have people present their supply in a market booth and to let people set up their shopping list. This process works well in community kick offs (see table below).
knowledge market workshop design|
Each participant in the workshop prepares a flip chart with what they have on offer for the knowledge market.
Also, they prepare a shopping list with what they are looking for on the market, using shopping list templates that are pre-numbered for later processing.
One representative stays at his or her market booth, while the other skims the market and check marks the shopping list. For those participants that only have one representative, a facilitator stays at the project booth, while representative walks around.
Draw a wall-sized picture with bubbles representing each of the traders in the knowledge market.
Invite participants to draw their arrows (directed) from their bubble to other traders, based on their shopping list
Ask participants to indicate their 3 most important interactions with others.
Discuss salient points in trading patterns, e.g. are there things on shopping lists that could not be obtained from market? This could be done by asking each of the participants to give 3 statements, e.g.:
One conclusion on commonalties
One conclusion on differences
One conclusion about blind spots
Ask for additions and clarifications
Work on finding topics that the community can focus on
The role of a knowledge steward is to facilitate knowledge sharing among (knowledge-) workers (Krogh, 1997). They actively search for lessons learned from individuals and project teams in order to make them available for a broader audience. For a project, together with the project members they define what should be learned during the execution of the project and suggest what kind of known experiences could be used in the project. At the evaluation of the project they take the learning stock, classify and disseminate the lessons learned. Furthermore they facilitate sharing days at which employees can discuss experiences without hindrance. Knowledge stewards can also be the moderator of (electronic) discussion groups. They can be receptive and can communicate changes and developments. A requirement for a knowledge steward is first of all having a certain reputation based on demonstrated competencies. Next to that communicative skills are of paramount importance. Furthermore the knowledge steward should have a good relation with HRM, the IT department, and the library and documentation centre. Using information technology e.g. intelligent agents to search the world wide web for resources shouldn’t make them feel uneasy.
issues facilitated by a knowledge steward|
What knowledge do we need?
What are we going to learn?
What knowledge products are we delivering/can we re-use?
Who needs to learn as well?
How are we going to consolidate and distribute our work?
What are we learning?
What did we learn?
Is knowledge consolidated satisfactory?
Have we informed everyone satisfactorily?
Lessons learned sharing
We have found that some structure and support should be given to the exchange process within the community. Although in many cases people are willing to share, (but are often not capable e.g. when codifying a lesson learned, it is often not clear to the author what the audience might be). Helping to capture the context by providing e.g. a lessons learned format and facilitation in writing down the lesson can help to improve the quality and applicability of the content of virtual community memories. This facilitation should be coupled with ‘natural’ moments and places of content creation and content use, for example by having a facilitated after action review session within project meetings. In terms of common exchange formats for content, we would like to refer to the work that has been done in the area of learning histories and story telling (e.g. Kleiner, 1996, Druxman 1997).
The Set-Up acquaints the reader with the mood, characters,and direction of the story. Here references can be made to knowledge areas touched by the story
The Catalyst, the event that “kicks off the story”
The First Turning Point, the plot must thicken. The story takes off into a new direction that eventually results in the learning
The Climax, the culmination
The Final Confrontation, the effect of the culmination
The Resolution, the wrap-up. Here the learning message of the story is presented as well as advice for those that will play a part in similar stories in the future
Components of the virtual community memory
If we look at the virtual community memory, the following basic components appear to be successful in practice.
Topic oriented, to-the-point on-line discussions can be a source of codified knowledge in the community. Responsive knowledge stewards as moderators are needed to chase contributions. Also these moderators should be equipped with flexible tools, so that they can easily summarise and link contributions to each other. In order to solicit participation in discussion, a notification mechanism that alerts community members to new entries in the discussion helps keep the discussion alive.
The virtual community memory can be equipped with a set of knowledge cards in which community members can present their knowledge areas and ambitions. These knowledge cards can serve as yellow pages of the community, and as advertising material in the knowledge market.
Of course, the greater part of the community memory will consist of documents. These documents can be made subject to a document model (i.e. each document should be set up in a particular format and should have predefined relations to other documents). Alternatively, the collection of documents can be regarded as a loose collection without any predefined structuring. The advantage of having structure is that retrievers of the documents get a structured response to their queries. The disadvantage of structure is that a burden is placed on authors of documents. They have to obey the structuring rules of the document model, and this may impede a free content flow from authors to readers.
When the community memory is growing in terms of quantity, relatively simple components such as the ones listed above lose their effect. Content readers experience difficulty in retrieving relevant items, mainly because they don’t use the language that is used by the authors of content. One way to try to bridge this gap is to work with a knowledge model on top of the content. Adding such a knowledge model (for example in the form of a thesaurus) to an unstructured document collection can partially overcome these problems. Through using a knowledge model, retrievers of content do not have to use the exact vocabulary of authors to find relevant documents to their queries. However, we should not overestimate the role of such knowledge models. The development and maintenance effort of such knowledge models is often only cost effective in larger collections.
When supporting cross-organisational communities, such as multi-party projects and professional bodies, a number of extra challenges arise. For instance, in many cases there is no common language, so it becomes harder to transfer knowledge without some extra guidance. Also, IPR-related issues may hinder a free exchange. Often there is no commonly agreed working method for sharing. There is often a lack of a common infrastructure (such as meetings and a shared repository). We are currently involved in setting up a knowledge pump (www.kalif.org) for supporting knowledge sharing between 16 European research projects, and what we see is that in the first phases the focus should be on the outer layer of the model in Figure 1, only after that the virtual community memory is starting to work.
In order to meet the challenges for supporting communities that share codified knowledge, one should carefully tune the mediating layer and the virtual community memory to each other. This tuning depends on the binding factors in the community, the level of codification of knowledge and on the deployment of information technology. We have given (although limited) examples of components that can be successful in creating a knowledge pump in a community. The creation of content that is rich enough for others to be useful is the main hurdle to overcome when creating a knowledge pump. In this respect, it is more instrumental to invest in learning to transfer rather than in learning to learn.3 Finally, although in this article the word ‘design’ is often used, we do not mean that these environments can be blueprinted. The development of a support environment for communities is iterative and also a learning experiment in itself. What holds for knowledge stewards, should hold for the designers as well.
|Communities of practice: history
Knowledge sharing in communities is nothing new. A phenomenon that is interesting to compare with the contemporary communities of practice are the guilds of the middle ages.
During the 11th and 14th centuries the western commercial and industrial classes united (Boissonnade, 1902). These unions were formed by patriciates to give resistance to the noblesse i.e. they were a tool of power. Amongst the founders were business-men, organisers of transport, industrial capitalists, ship owners, money changers, bankers and rich urban landowners. These unions received a legal and official existence. Usually it was called guild or Hanse, sometimes fraternity or confrerie. The main goal of a medieval guildship was to help and protect each other and furthermore to arrange feasts. In most cases the craftsmen of a guild lived, worked and learned together in the same building i.e. the guild-hall.
Learning in a guild
There were three levels of craftsmen: first there was the master who attained this level by preparing a masterpiece after a long period of apprenticeship. The master was allowed to perform his craft individually. Then there were the journeymen; they were skilled apprentices who work in daily comradeship with masters, but didn’t make their masterpiece yet. The lowest level of craftsmen were the apprentices. They could be considered as students who learned under supervision of a master. Both the master and the journeyman played the role of teacher for the apprentice. Apprentices learned by imitation and correction of masters and journeymen. Furthermore they learned from their peers based on telling stories. Sharing of knowledge from master to apprentice wasn’t considered a threat since the way from craftsmanship to mastery was long and during this period apprentices were not allowed to perform their craft individually.
The guild hall as a sharing place
Both life and work of craftsmen took place in one building (i.e. the guild-hall), hence the way they learned was highly intense. Because the masters controlled and regulated the activities in the guild-hall. It was considered a safe learning
environment for its members. Paradoxically this form of protectionism led to the decision of the 2nd constitutional commission in 1798 to disband guilds. They became a fortress based on their own laws and privileges which enabled them to exclude certain people. Peasants where banned from mainly city-based guilds. Hence they didn’t have the opportunity to learn a craft.
The history of the guilds teaches us about ways to make communities more tightly coupled, by discerning explicit roles within the community, applying explicit and shared ways of exchange and providing an institutionalised common sharing place. We have learned from current practice that cross-organisational communities often lack these roles, shared processes and institutionalised knowledge exchange places.
Boissonnade, P., Les études relatives à l’histoire economique de la France au moyen age, 1902.
Druxman, Michael B., How to Write a Story...Any Story: The Art of Story Telling , The Center Press, 1997.
Kleiner, Art, Roth, George, Field Manual for a Learning Historian (version 4.0), MIT-COL (Center for Organisational Learning), October, 28, 1996.
Heijst, G. van, Spek, R. van der, Kruizinga, E., Corporate Memories as Tools for Knowledge Management, in: Expert Systems with Applications, Vol. 13, No. 1, 1997.
Szulanski, Gabriel, Exploring internal stickiness: impediments to the transfer of best practices within the firm, INSEAD white paper, 1996.
Krogh, G. von, Nonaka, I., Ichijo, K., Develop Knowledge Activists!, in: European Management Journal, pp. 475-483, 1997.
Eelco Kruizinga and Timo Kouwenhoven are with the knowledge management practice of Kenniscentrum CIBIT, Utrecht, Holland.
They can be contacted at: