Starting a CoP

From KM4Dev Wiki
Jump to: navigation, search

Starting a Community of Practice


What are the minimum criteria to start a Community of Practice and how do you start it?

Communities of Practice (CoPs) have been identified as a way for groups to learn and share knowledge together. In development, CoPs offer a way to develop and share knowledge, improve practice and support collaboration both within and across organizations. Consequently, many seek to start a CoP. This FAQ looks briefly at some of the elements to consider when starting a CoP. Key issues include:

  • what are the key issues to clarify around a community's domain?
  • what is the minimum number of people?
  • what are they sponsorship requirements?
  • what are the community leader/facilitator roles?
  • what might be the tools and technologies to support a CoP start up?
  • what initial activities help catalyze the community?


community of practice, CoP, learning

Detailed Description

There are a number of issues in thinking about starting up a CoP in development. From the KM4Dev discussions, some of the issues that have been discussed include:

  • Do you spot an existing community and support it or can you create one from scratch?
  • What is thhe donor or sponsor perspective which includes issues of
    • how much control a sponsor exerts on a form that is strongly voluntary in nature
    • if it is better that CoPs arise on their own, or even "fly under the radar" at the start so that they are built with member energy and commitment, not a mandate from above
    • the role (positive and negative) of incentives on community formation
  • What is the member perspective
    • What is the minimum number of interested people to start a CoP? Maximum?
    • What motivation and interest (self interest/group interest) is needed to start a CoP
  • What is the facilitator perspective?
    • How do people catalyze the social relations, conversations and practice needed to generate a lively CoP?
    • Must a CoP be facilitated?
  • What is the technology support perspective
    • For communities that are not co-located, what tools and technologies can support CoPs in development where there may not be uniform internet, hardware and software access, and tech skills
    • Who sponsors and supports the technology
  • What is the full community lifecycle perspective
    • Instead of just focusing on the startup phase, a community takes a longer view of startup as just one thing to pay attention to in the lifecycle of the communit

From a practice/practical perspective, here are the suggestions from members on starting a CoP:

KM4Dev Discussions

CoPs have been a frequent topic of discussion on the KM4Dev discussion list. A key discussion was the Peer Assist with Tony Pryor looking at how CoPs have (or have not) been embraced in development. Tony asks the question about how to support the emergence of CoPs that aren't just the same group of people talking about the same thing... how to truly use the power of diverse people to address a common practice.

Paul Whiffen on 09/25/2000 05:15:42 AM

Please respond to "Knowledge Management for International Development Organisations" To: "Knowledge Management For International Development Organisations" Subject: Stories and Communities of Practice Dear all, hope you had a good weekend,

  • stories are very powerful and effective at getting people engaged / excited and really interested in things, including knowledge management.
    • We did a 'Learning After' following the Tearfund response to flooding in Bangladesh a couple of years ago and stored the lessons. In fact this was the first conscious KM process that we tried. 'Storing' at that stage consisted of nothing more than putting the output lessons on a simple document which ended up in the Asia manager's desk draw. A few weeks later, it looked like Bangladesh was going to suffer in the same way again. The quote from the Asia regional manager was of great interest: 'The first thing I did was go into my desk and recover that key set of lessons we identified the last time.' Now we have a more organised and structured approach to the storing of lessons, but the principle is the same.
    • The other day the office canteen manager came to me (the canteen feeds 200+ persons daily). She asked me what this KM thing was all about. I explained to her the basic principles of capturing what we learn to re-use it again in the future. She was amazed by the simplicity, power and 'obviousness' of the approach and registered real interest in it for the way the canteen operates. I like this story because it shows how transferable the ideas and concepts are.

CoP FAQ's (Frequently Asked Questions). Courtesy BP Amoco.

1. What is a Community of Practice (CoP)?

- CoPs are informal networks dedicated to sharing knowledge among practitioners.

2. What is the role of a CoP in Knowledge Management? - The purpose of a CoP is to share knowledge and experience, so that each individual can operate more effectively. CoPs are the owners of knowledge in that particular area of knowledge. By exchanging stories, problems and solutions, the CoP can bring heir collective knowledge to bear on individuals' problems. The CoP can also take various experiences and solutions from around the organisation and build a knowledge asset representing best practice.

3. Who takes part in a CoP? - CoPs consist first and foremost of practitioners; specialists who perform the same job or collaborate on a shared task. The Community acts like an in-house professional society, cutting across team and divisional boundaries.

4. How big are they? - Intense face to face CoPs seldom grow larger than 50 people. However, small local CoPs can be bound together into a wider community by communications technology, and a membership of up to 100 or more is not uncommon.

5. What do they need in order to work? - CoPs often form spontaneously, driven by the need of the members for operational knowledge. A workshop or conference often provides the catalyst.CoPs can also be deliberately encouraged in areas where it is known there is a need for knowledge transfer. CoPs do not facilitate themselves. They need a facilitator ; someone they see as an 'insider' and who has the respect of the community. CoPs require organisational recognition to be really effective, and face to face meetings strengthen the communication and relationships.

6. How do they communicate? - They may rely on electronic communication. Email distribution lists and online discussion groups help strengthen relationships that have developed at face to face meetings and provide 'meeting points' for members.

7. Is a CoP the same as a network? - A CoP is a form of a network, but not every network is a CoP. A network could be considered a CoP if it is informal, open to all practitioners, works as a mutual help society rather than having a shared performance contract, and has a means of constant virtual communication rather than relying on occasional formal meetings.

- Make it people led, not technology led;

- Get the top guys on board

- Pick on a project or two that's going to demonstrate success from KM early on;

- Tell some good stories.

4. To me there is no tension between flexibility and storage. You are storing knowledge (in the form of information) for sure, but that doesn't mean that the next user of that knowledge can't take it and innovate it for his / her uses. There is no conflict between KM and innovation – indeed clear KM makes clear what possibilities there are for innovation!

5. In Tearfund, KM is aimed at everyone everywhere eventually. It needs to be part of everybody's mindset. I used to work offshore in the Oil Industry and an analogy is safety. Everyone has here own technical job, but safety is EVERYONE's job (or part of it). The rigs I worked on had posters above mirrors in the washrooms saying 'you are looking at your own safety officer'. Of course there are safety professionals on the rigs, but they were there to imbibe it to everyone else, not to be the only one interested in safety. This analogy always works for me anyway!

6. Ref doing all this with Partners, well that's early days for us – looking forward to learning about that one...! Our goal is to link up all of Tearfund Supporters, Tearfund itself, Partners and Poor in one large learning ethos. OK, it's not going to happen overnight...! Paul.

1) Allow the Emergence of a Fabric of Overlapping CoPs

I strongly believe that his statement on a anthropologically natural group size is at the centre of the issue. (However in my memory, the "natural unit" has been about 60 units (persons or households; the latter is a multiplier in size but also a denominator in complexity), where we are capable to keep the overview). At that size, there is a tendency to split up - as I would suppose for the good of the whole group. I remember that in Bellanet KM workshops you always pointed out that CoPs are not only a social fabric in themselve but on a meta-level they build a fabric among different CoPs. Key figures "agents", who participate in two of them and*selectively* transport issues from one to another.

This multitude of similar groups (CoPs) on one and the same issue is not a flaw / imperfectness. In contrary redundancy is also a base principle of social organisation. It contributes to real diversity.

It is clear that same issues will appear in different CoPs time and again. One could say that they are reinventing the wheel as another CoP probably already has found useful answers. Yet, I would suppose that this is not a disadvantage, if we consider CoPs as learning communities. Learning most effectively takes place if people work on an issue (constructivist learning paradigm). To spare people from this process of struggling with an issue means to prevent learning (internalisation). So too much efficiency might turn out to have negative side effects.

2) Gardens of Wisdom and Trees of Knowledge

Once again I want to bring in our metaphor of gardens. A grove of trees has the purpose to produce fruit. Even if the grove (and its purpose) remain the same over time, the individual trees in there go through a life cycle. Why not having a "mother KM4DEV CoP" (the grove), in which you occasionally plant a tree for a specific purpose (e.g. for a specific group / major issue)? These groups would be created ad-hoc and people who share the specific interest would decide to participate (i.e. sit down under a specific tree to palaver). When the tree (the ad-hoc group) has reached the end of end of its life span, it would be felled (i.e. closed down) and people would return to the grove (mother CoP), which would continue its existence.

What is the benefit of this setting? The mother CoP would guarantee the necessary continuity and communality (commonality?). However by swaping out defined parts, the mother CoP would be relieved from to many noise of limited interest. This could also be combined with the idea of "gatekeepers" as suggested by Ian a few days ago. When such an ad-hoc CoP reaches the end of its life cycle, it is dissolved and people reintegrate into the mother CoP (respectively they have never left the grove).


Lesley, It seems to me that if you are starting at a *single* site (as for us at Tearfund HQ, Teddington, UK), then a KM model that helps to win early 'traction', or buy-in, with the organisation (by demonstrating captured learning for re-application) is the Learning Before During and After model, which builds knowledge assets of learning.

It also seems to me that if you are an organisation that has *many* sites and your priority is connecting them up to help people help each other and reduce the extent of re-inventions of the wheel, then the model of Comunities of Practice (CoPs) or Thematic Groups is the best way to gain early traction.

However, whichever one of the two you start with, you ultimately need both. It's as if each are incomplete without the other.

We started with the LBDA model, and only relatively recently have begun talking about CoPs as we start to think about connecting up with other Tearfund offices and encouraging our 400 overseas partner organisations to connect - and we still have a long way to go in this area....

However, what is interesting is that the two models seem to have an attraction to each other inasmuch that CoPs start to own the Knowledge Assets, and they are two parts of a whole rather than separate things.

For example, a recent CoP set up in Tearfund is the Desk Officers' forum. This is an informal network of the Desk Officers in different regional teams to help each other and compare notes, thoughts, insights and knowledge as they each deliver through their own projects and regions.

However, it was also decided that they needed a knowledge asset around 'Being a Desk Officer' to help orientate new staff when they arrived. This needed to contain all the informal and human stuff that new people need to hear (like managing stress / workload, the key things and relationships to get right etc) - ie, the kind of help that they are not likely to find in the formal job descriptions.

So, I interviewed 4 of the Desk officers to find out this kind of knowledge. Out of it came all sorts of practical, human advice on 'Being a Desk Officer'. This asset is now owned by the Desk Officers Forum CoP.

So, all I am trying to say is that the CoPs can be very powerfully linked to knowledge assets and they gravitate towards each other. Conceptually too, it makes sense that KA's should be owned by networks rather than individuals where possible....

This helps to make the move from an organisation that is shaped / led by structure (departments) to one that is led by core activities delivering against a strategy - and where support to teams (which will often be cross-functional teams) will come from CoPs and Knowledge Assets.

Also, I see no reason why this argument cannot be extended outside a single organisation to a collection of organisations working in partnership with each other on projects and initiatives...


CoPs within the organisational context of SDC require two things: (1) time, space, resources, as Sibylle said and (2) the firm committment by the management and the interest, to consider theoutcome, the result, the recommendations made by the CoP. One may as wellcall it a "management response". This contrasts sharply with task forces, where the management specifies all terms. We can look at such a "framework provided by the organisation" as an important incentive. This is true as well for CoPs in an "interorganisational" set-up. Manuel Flury en&p_phase_type_id= Learning by Visiting as a Means of Building Technological Capability

Hi all -- I researched a term paper for a graduate course earlier this year that might be of interest to some reading this thread. The paper, entitled "Learning by Visiting as a Means of Building Technological Capability," looks at the idea of "learning by visiting" and specifically at its application to small enterprises in developing countries. Certainly a promising way to spark and foster the growth of networks and CoPs, and seemingly an under-researched subject.

The paper is actually available online at en&p_phase_type_id=6 for anyone interested in taking a look. I'd certainly appreciate comments from anyone who takes the time to read it. Experted from: "Virtual Communities At Caterpillar Foster Knowledge Sharing", Vicky Powers, Training and Development, Jun 2004, Vol.58, Iss. 6; pg. 40.

Also see: Tenth Annual Report to the Prime Minister on the Public Service of Canada (2003) – Donor program should not narrow a CoPs approach and worldview to its own "bias". We have to change that donor perspective of a CoP.

- Some donors are very uncomfortable with diversity and complexities of approaches in development, constantly searching for generic frameworks. Development is always complex and its diversity is its most valuable asset.

- Try to keep discussions "agency" independent. It is more complicated when it comes to financial and human resources but allows independence of donor worldviews (example: iNARS e-discussion list in ICTs in agricultural development).

- The most interesting discussions lie at the intersection between different disciplines.

- A CoP must be by and for its participants and members must perceive real value in their involvement. The perception of a CoP as a "PR vehicle" for its donor's policies can undermine its authenticity. Likewise, a CoP that is set up as a focus group to help inform policies risks becoming a one-way process. Occasionally, participants and donor needs can happily coincide.

- Ironically, a CoP "by and for" its members can find it hard to attract funding (example: Aid Workers Network). It is difficult to promise hard deliverables since - by the very nature of voluntary participation - there no control over the participants' actions. Trying manage a CoP in the same way as a project with more defined deliverables can significantly diminish its potential. By controlling less it might be possible to achieve more.

- Realistic indicators of success might be the number of participants, the level of participation and the benefits reported by participants.

- Listen carefully to the concerns and priorities of the full range of stakeholders represented. Good facilitation can be key: clearly define the overall goal, encourage all lines of enquiry and new perspectives, do not pass judgment on contributions, and ask open questions that encourage diversity of input. Lucie Lamoureux

- Subscribe people interested in the topic and get them to invite other people they know to join.

- A CoP would be "open" without people who necessarily agree on everything. We should seek diversity of opinions, views, concepts, perspectives.

- As a matter of principle, we should try to only participate in open communities and shun those that are closed even if invited.

- Look at alternative approaches such as the Aid Workers Network's "honeycomb", where participants can segment themselves into different special interest groups that are in close proximity to each other and can "spill over", so members encounter people from other disciplines. They also encourage discussions to be held openly so sub-groups do not become insular.

- Focus on the quality of the list and its ease of use for readers and participants will come by word of mouth from everywhere.

I was involved in a few communities of practice that were defined as such (in fact we are all part of lots of CoPs' all the time). What helps as incentive is to sponsor some secretary role. That is what keeps a platform like this ticking.

All other outside incentives tend to corrupt the thing, turning it into a project with a closed character.

Bill Snyder wrote about it, but I am sure you have read his work. You can find it on the net (William Snyder, )

George de Gooijer

Examples in Application

A case study of knowledge managment in multi-agency consumer-informed "communities of practice": implications for evidence-based policy develpoment in health and social services John Gabbay, Andrée le May, Harriet Jefferson, Dale Webb, Robin Lovelock, Jackie Powel, Judith Lathlean Health: An Interdisciplinary Journal for the Social Study of Health, Illness and Medicine Vol 7 (3): 283-310, 2003

Abstract: We report a study that facilitated and evaluated two multi-agency Communities of Practice (CoPs) working on improving specific aspects of health and social services for older people, and analysed how they processed and applied knowledge in formulating their views. Data collection included observing and tape-recording the CoPs, interviewing participants and reviewing documents they generated and used. All these sources were analysed to identify knowledge-related behaviours. Four themes emerged from these data: (1) the way that certain kinds of knowledge became priviledged and accepted; (2) the ways in which the CoP members transformed and internalized new knowledge; (3) how the haphazard processing of the available knowledge was contingent upon the organizational features of the groups; and (4) the ways in which the changing agendas, roles and power relations had differential effects on collective sense making. We conclude by recommending ways in which the process of evidence-based policy development in such groups may be enhanced.

Related FAQs

[Insert links to related FAQs]

Further Information

  • Etienne Wenger’s home page. Wenger is the researcher and consultant who coined the term along with Jean Lave. [1]

An innovative approach to delivering public policy, Michael Kelleher

  • Anecdote"I now have a simple test to gauge whether a community of practice might form. When someone says, “I would like to start a community of practice.” I ask, “Can you describe the potential members by completing the following sentence? I am a …..” If they can fill in the blank in a way that people can passionately identify with the descriptor then there is a chance a community might emerge. Let me give you an example. I was helping the Department of Defence design a community of practice for project managers. ‘I am a project manager’ was a strong descriptor and so we knew we had a chance. During the design process the client has another job type for which they wanted a community to support simply called ‘technical’. ‘I am a technical’ didn’t inspire so we knew it was unachievable. The ‘I am a …” test is easy and effective."

Original Author and Subsequent Contributors of this FAQ

[Mention name(s) of author and subsequent contributors]

Dates of First Creation and Further Revisions

FAQ KM4Dev Source Materials

Discussion of CoPs at CARE

Storytelling and CoPs

Using surveys in surfacing CoP issues

Balancing community discussion with creating resuable repositories

Buckman Labs emphasis on connecting before collecting

Underground or clandestine communities and more on clandestine nature of CoPs

Post with link to document on "teasing" readers into participating

Peer Assists as CoP stimulant or catalyst

Industrial cluster development paper

Support for emerging Cops [ Incentives for CoPs and More on incentives and More on incentives and more...

UNDP Thematic Network start up story

Documents to support CoP startup from Elsie Importance of critical mass of relationships for CoP development

USAID's CoP wiki manual