Communities of Practice: The Institutionalisation of Informality

January 05, 2011 | bit-wartung | SDC Experiences |


Rating: none

Tobias SommerBy Tobias Sommer
Communities of practice in their traditional definition are every manager’s dream: on individual initiative, experts form informal communities out of pure interest in a subject to exchange their experiences, talk about new challenges and learn from each other, pushing their field forward and developing new solutions. Ideally, all of this would happen alongside normal work, stay out of organisational structures and employment agreements, costing the company not time nor money nor energy – sort of a holy grail of knowledge management.

Sadly, as with anything that seems too good to be true, this hasn’t worked for a long time. In many organisations and companies, informal networks have lost their drive for several reasons, one very important of which is the overwhelming wealth of instantly accessible information through the internet, evoking the illusion that one can learn more from a computer than a colleague. Managers now face the question whether they want to continue without these prolific expert CoPs or if they want to reinforce them by institutionalising them and dedicating the necessary resources and attention to them. In their article “Harnessing Your Staff’s Informal Networks” (Harvard Business Review 03/2010; article can be purchased here) Richard McDermott and Douglas Archibald vouch for the latter option and – based on an extensive study of such networks – give an overview of what institutionalised networks can look like, and in what they should differ from the traditional communities of practice. As SDC’s thematic network organisation is a commitment to the institutionalisation of expert networks too, this is a good opportunity to compare their findings with the experience of our own networks.

The authors recommend four principles according to which institutionalised networks should be organised:

  1. Concentration on important subjects
    Active communities should work on issues that have been defined by top management. The example of Pfizer is mentioned, where 9 so-called councils work on crucial themes of pharmacology. Additionally, there are 12 less formalised networks that work on subjects of secondary importance and can be upgraded to councils if demand is rising.
    SDC networks do fulfill this requirement: Their foci lie on clear thematic subjects as defined by the directorate. Contrary to Pfizer’s model however, there is no hierarchy of these subjects; they all have the same institutional weight.
  2. Defining goals
    The traditional conception of Communities of Practice that exchange ideas and experiences without any clear-set goal becomes obsolete when these communities become institutionalised. Setting explicit and measurable goals and requirements can stimulate the exchange of information and make networks collaborate more efficiently. These goals however should differ clearly from goals set for project teams in that they are of a more general nature and long-term oriented.
    SDC networks operate in a different logic. While general guidelines for the role of networks in the organisation have been defined by the management, thematic goals are defined by the networks themselves. Thus their ambitiousness, measurability and degree of concreteness can vary considerably.
  3. Creating clear structures
    For a good integration into an organisation, networks should have strong and clearly defined connections to senior management. The authors however claim that just appointing senior managers as tutors of networks isn’t enough; it is essential that the appointed managers understand the value of such networks and can dedicate enough time to mentoring them.
    In this respect SDC networks work in an almost inverse logic: the thematic experts in charge of our networks have to report their networks’ progress to the directorate and promote their importance in order to assure the directorate’s support.
  4. Setting higher requirements and expectations
    This is tightly linked to the definition of clear goals: Networks should be required to play an important role in the organisation and to take certain responsibilities. In the construction company Flour for instance, management expects the organisation’s networks to be the first source of information, and to elaborate normative standards and processes.
    SDC networks have no normative power and set their goals themselves. Thus their influence on the operative level depends largely on the level of activity, self-promotion and the decision-makers’ appreciation of networks as pools of thematic expertise. Higher and explicitly defined expectations could improve their influence and be a way to tap their full potential.

The critical question we need to ask ourselves is: what dynamics does the current organisational configuration of SDC networks lead to? Are networks in a competitive setting, and if so, does this competition improve the networks’ results or does it produce artificial activity that does not necessarily contribute to a more efficient SDC? One main lesson we have learnt from our exchange with GTZ, KfW and Solution Exchange (read Patrick Kalas’ reflections on Steve Glowinski’s visit to SDC in 2009) is that career advancement is considered the most important incentive for staff to actively participate in networks. From this perspective the latter possibility could well be the case. A more flexible network structure with two or more levels – like Pfizer’s model mentioned above – with some strongly institutionalised core networks focusing on 5 to 6 priority subjects and a number of less formalised networks dealing with matters of less importance could prove more efficient and adapted to SDC’s thematic priorities.

In the second part of their article, McDermott and Archibald enumerate four ways to improve the efficiency of networks:

  1. Giving network leaders enough time: on average, network leaders spend 17% of their work time for the management of the network. This activity and the amount of time it requires needs to be an explicit part of the job description; otherwise it is likely to be the first item on the task list to be put back when other pressing issues are pending.
  2. Training leaders for their role: leading a network is not the same thing as leading a team. Therefore network leaders need to be trained specifically for moderating a network. This includes skills in mobilising people for e-collaboration, dealing with the transorganisational nature of a network and finding ways to have an impact on operative units without having normative power.
  3. Organising face-to-face meetings: e-collaboration software makes it possible to cooperate across geographic distance, but it does not replace the creative power of personal meetings, nor does it foster personal trust that can be essential when it comes to sharing experiences and learning from each other’s mistakes. Face-to-face events are therefore crucial for the initiation and continued activity of networks.
  4. Using simple IT-tools: when it comes to e-collaboration, simplicity and familiarity are much more important than sophisticated cutting-edge functionality.

Regarding these four recommendations, SDC networks are mostly up-to-date with McDermott and Archibald’s findings. Thus the biggest potential for improvements of SDC’s networks might not lie in their internal configuration, but in the way they are embedded in the organisation and how they are connected to and appreciated by the directorate. Currently the major challenge lies in finding an organisational setting where operative decision makers consult networks despite their lack of normative power.

Learning from and exchanging experiences with other organisations can be a precious source of inspiration to rethink these features, as we have learned from our exchange with regiosuisse’s Knowledge Communities, UN India’s Solution Exchange networks and former GTZ‘s knowledge management division.

To learn more about SDC’s experiences with communities of practice and networks, go to

No Comments

Leave a Reply