Thursday 23 June 2016

How scaleable are communities of practice ?

Communities of practice are seen almost as a silver bullet in KM terms - a must-have item.  But does the concept of communities scale to all organisations?

Communities of practice as a concept is 25 years old, going back to the original work of Etienne Wenger and Jean lave, who define communities of practice as “groups of people who share a concern or a passion for something they do, and learn how to do it better as they interact regularly".

The uniting force for the community is that "something they do" - an area of practice - and the focus of the community is "learning through regular interaction".  By learning in a community, community members can both use each other as a knowledge resource, and also co-create knowledge within the community. Communities of practice are as much knowledge-creating mechanisms as they are knowledge-sharing mechanisms.

There are two end-member approaches to Communities of Practice - large online communities, and small face-to-face communities.

The large online communities are immensely successful KM solutions for multinational organisations. The big consulting houses make extensive use of online CoPs, as do the big oil companiesand engineering houses. I have CoP success stories from  Siemens, Merck, Halliburton, Fluor, Caterpillar, Orange, Conoco and Shell.

These CoPs are largely focused on problem solving, and responding to reauests for help ad advice from members. They work through online discussion, often with the help of a facilitator or moderator.  The bigger the better, as far as these communities of practice are concerned, as shown in the graph below. Some of the most effective communities have thousands of members, and there is certainly an issue of critical mass. Communities need to have a "buzz" - they need regular, and to have regularity, they need mass. Online communities of practice which are too small are unable to sustain enough activity, and die. (see this blog post about how much buzz an online community needs)

At the other end of the scale are small communities of practice that meet face to face. The type example here is the tech clubs from Daimler Chrysler; small groups of tens of people that would meet once every two weeks to discuss technical issues and solve technical problems.  These online communities are often very valuable knowledge sharing mechanisms, but seldom grow beyond about 50 people. This is because discussion becomes difficult above 50, and because larger groups tend to meet less regularly and, again, that feeling of buzz dies away.

It is between these two end members that Communities of Practice get most awkward.

Firstly you cannot scale down an online community of practice far before you run into the issue of critical mass. This is what is behind the decreasing satisfaction in the column graph above. Most of the surveyed communities were online communities, or had some online component, and the smaller they are, the less effective they are deemed to be.

Secondly you cannot scale up face-to-face communities. Above 50 people, they cannot have a good discussion.  They start to fragment. The effectiveness of face to face communities would show the opposite trend to the column graph above.

There is an awkward zone for community size, where they are too large to meet face to face, and too small to be viable online. This zone is probably about 50 -  250 people, and this is a zone where considerable effort will be needed from the community facilitator.  You will either need smaller face to face groups within a larger online community, or you will need to amalgamate several of your awkward communities into one larger community with the requisite level of buzz.

Communities of practice is a great concept, but neither the face to face nor the virtual community solutions are fully scaleable, and they leave an awkward zone where neither solution works properly. Therefore think carefully when scaling a community solution. 


Bruno Winck, Kneaver said...

It's an interesting point of view but I tend to think it's a bit artificial. I think that more than the modality (online/f2f) what count are
- The nature of the domain being discussed
- The connection with projects (occasional vs sustained interest)
- The quality of the community manager.

Nature: I've been invited several times at Daimler as an expert in surface design and 5 axis manufacturing. Whenever you get there (and it you are trusted) you are quickly asked to get into the workshops and touch, turn around the parts, the machines. It's a reflex. Even with VR it's not possible today. It's the same at Ford, Honda, Hyunday, Zodiac etc.. Every time it ends up face 2 face. It's also small groups for a simple reason: there are few experts. take the c++ group (it's a normalisation community but works the same), limited group, continuous online conversations, no facilitator. Because the domain is immaterial 2 meetings a year are enough.

- Project mode vs resource mode. if the CoP topic is central to on going projects it will work no matter what because it's just needed. You have this on small online CoP in IT like Express-G (web framework for NodeJS), WP-JSON PAI (future API of Wordpress). No need of f2f meetings, people will work together for 5, 6 years without even meeting once.

- Quality of manager. A new breed of community managers, born in social media are progressively taking charge of CoP. they know little of the domain but they are excellent at creating stimulating, cooperative atmospheres. You find them first in IT, Marketing, Sales CoPs but as they become more prevalent we will see them come in more old style CoP. You could make a community more active simply by providing a steady stream of curated resources, occasional though provoking questions, inviting guests, asking for feedback on events, creating contest etc. It's pure buzz creation, pure facilitations techniques. As opposed I have seen many communities die literally because of the limited social qualities of the initiators. Where your view and mind overlap is that often large communities are professionally managed.

So my take is that well managed communities, on topic worth the effort can range from 15 to 9000 (I'm in one with 9000, built in 2 weeks, over 6 continents).

Nick Milton said...

Thanks Bruno

you are of course right, and good management on topics worth the effort can get you past the "awkward" conditions.

Blog Archive