Talk:How is knowledge captured and synthesized from discussions

From KM4Dev Wiki
Jump to: navigation, search

See the original thread of this E-Discussion on D-Groups

Tony Pryor, 2008/12/03

I know there's quite a bit of overlap with the Outcome Mapping (OM) discussion group, but I wanted to call your attention to this post on how to synthesize discussion threads. This has been an issue for me with km4dev from the beginning; how do we archive/harvest threads which way have a relatively short half-life on their own. My own experience from years past, when I ran a email list for USAID called RFNET, was what I had to pay a knowledgeable reviewer about 3 weeks of work a year to synthesize these threads into monthly 1-2 page summary sheets which would be indexed and readily accessed. While an expense, this mail list had a broader "corporate" purpose (in this case promoting and institutionalizing internal reforms) that made that cost worthwhile. In fact, without that expense the value of a continuous exchange on a constantly changing set of issues was seriously degraded, once a month or so had passed.

This instant messaging world ironically can make knowledge even more fleeting and ephemeral, somewhat memorable for those directly in the thread, but a confused jumble of ideas for the rest of us - like trying to gain insights by attending cocktail parties for a month (well, not necessarily THAT bad!). I know we've done a very good job within the km4dev family using the wiki space for some of this, but I wonder what more can be done. This is an interesting approach Simon raises for OM; are there others out there we all can learn from? Is the voluntary approach suggested by Simon sufficient for all uses and all groups?

And by the way, under the auspices of the Knowledge Driven microenterprise Development effort at USAID, I've started a blog as part of the KDMD KM Lab; I've already highlighted some issues of interest to chat group, and was planning in fact to cover this topic - how to harvest from threads - later next week. But I thought now might be an opportune time to flag the blog, even though it's just beginning its maiden cruise. Please fell free to comment on any other of it; I will drag any relevant material over to either OM or km4dev if of broader value.


Lucie Lamoureux, 2008/12/04

Hi Tony,

I know that systematic synthesizing (try saying that 5 times fast!) would add a lot. The suggestion comes up time and again but the cold hard reality is that the two days per week I dedicate to KM4dev are spend doing far less glamorous stuff, such as helping people work through our current wonky tech tools, setting up wiki accounts, rejecting spammers, doing checks on the people who want to join to see if they ARE actual people, etc. Doing summaries does take time so we have been asking community members who initiate questions if they would be so kind as to put the synthesis of the responses in the wiki. The upside is that this means the community itself is building its own knowledge base... the downside is that it isn't systematically done.

I have been reading the CoP facilitator/coordinator/admin etc. thread with much interest, as you can probably imagine. For this community, I have become less of a facilitator and more of a "backstage presence", which given the maturity of the community (over 8 years old) is understandable. In the early days, there was more of a need to get discussions going; now that the community is established, dialogue flows easily... much too easily, some would say! KM4dev is a high traffic list, which is great, but that also entails a lot of work to get everything summarized. Unless we can get more donors to come on board - and if anyone knows of any, bring them on! - I'd like to make another plea to the whole community to ask me for a wiki account and pitch in with the synthesizing of the threads. I know it's not a perfect solution but I'll echo Tony (below) and ask that if anyone has any other solutions, please let us know.

KM4D Associates.

Simon Hearn, 2008/12/11

Hi Tony,

I just saw your reference to the OM list and thought I'd comment - albeit far too late (I'm sure the discussion has moved on a mile =ince this posting but I can't keep up with the rate of exchanges on KM4Dev these days - not that that's a bad thing, just a challenge - and a good on= to have J)

The practice of summarizing on the OM discussion list is actually rather poor. I, as the coordinator, occasionally summarize an interesting discussion, but I've not managed to develop the practice among users - unless specifically asked to do so. I actually take KM4Dev as by inspiration when it comes to this practice.

What I have done is managed to pull together a team of volunteers to trawl the back logs looking for interesting discussions to summa=ize and collate into a community publication, capturing the learning over a period of time. We've compiled such resources Outcome Mapping ourselves in the past - se= - but what would be ideal is if the process can be community led - a True community publication. But working with volunteers is, of course, difficult. I have to keep reminding myself that the aim is not solely to comify knowledge, but to build community in the process - even if it takes more effort from my side.

Tony Pryor, 2008/12/12

Cheers, Simon: Not to worry! I too am overwhelmed, and tend to flit in and out of these discussions!

I'm not sure there really is a good/perfect system in terms of a voluntary approach. KM4DEV works so well because I think the nature of the subject almost by definition attracts those who are about sharing and collecting knowledge; if the KM crowd won't try to save and archive =nowledge, who will? But you also have people such as Lucie and Nancy, who go way beyond teh call of duty. Any successful volunteer-based entity has to have someone like this, which brings to the task a level of ownership and enthusiasm that money can't buy (not that they wouldn't mind a doubling of the budget...!). BUT it also makes such a system =ulnerable to departures, illness, other work pressures and the like.

Might be useful to develop a tool that links such threads into a wiki (like Nancy does manually) that can help collect and collate such discussions. But I find that the best such archived results imply a subtlety of synthesis which go far beyond teh mechanical or automatic.

I like your last point re not forgetting that getting people actively involved is as important as the result. But I'd say that the opposite can also be true, especially when the content potentially has high value for others, or could be misconstrued or not fully used by future readers unless placd into context - getting the result may be as important/even more important that getting members actively involved. Active involvement can come about in many ways; it may be though that the synthesis function may not be the best element to use to promote involvement, when the value of the synthesis is high.

Having said that, wikis are remarkable tools for certain things, but I don't think they take the place of a synthetic/analytic function in living long term value to an ephemeral discussion.

Johannes Schunter, 2008/12/12

Hello all, very interesting thread.

I wonder whether the recently posted question (and the answers provided by KM4dev members) on outsourcing CoPs would apply here as well? If the management of a CoP can be outsourced to an NGO in India (as proposed by Satish), why not the consolidation of discussion threads into a wiki?

Satish, what's your take on this?

Jaap Pels, 2008/12/12

We need knowledge monks. Re-digesting e-mail treads. Annotate attachments. Clean / brush up the text. SocialText integrates wiki and e-mail; essentially an e-mail tread is a wiki page. Might be a good tool for knowledge monks.

Tony Pryor, 2008/12/12

ooh, I love that concept...!

Tony Pryor, 2008/12/12

Good point, bringing in that other thread.

I guess the same logic holds true for the "monk" function as the "moderator" function (as opposed to facilitator) that I mentioned earlier: it can be outsourced as long as 1) the person really knows the subject and can add value (but not just his/her opinion regardless of the thread) and 2) is trusted, and known, by the community as a peer. but given those 2 variables, no reason that the monk couldn't be outsourced.

And frankly outsourcing it may indeed be the way to go. I found that the work was essentially "piece work", it came up as threads came up, and it didn't justify having a person hired just to do that, and yet relying on volunteers is risky sometimes since the time management of the effort on a volunteer basis is not a minimal task, if the threads are fast and furious. What one ended up with then was a quarterly fixed price "summary" and then hour by hour jobs as work came up.

"Monks for hire" may not sound quite right, but it's a reasonable concept!

Grant Ballard-Tremeer, 2008/12/12

At the HEDON Household Energy Network ( we have tried making use of volunteers, young professionals around the world who receive a small stipend, as well as staff at my company to manage communities of practice and synthesize discussions. How well it works has been a bit patchy. In our experience:

  • It's crucial that whoever does the synthesis really knows the subject well. To some extent it is possible to manage discussions with only a basic understanding of the subject (as long as the person knows how to facilitate discussions), but writing up a good synthesis requires significant understanding of the topic.
  • People who are managing groups 'in their spare time' are often not all that reliable. We've found that having a very clear set of tasks (process guidelines) really helps with that since it gives a structure to what has to be done. We also hold a weekly group leaders meeting on Skype where each group leader reports on a few key indicators and the team discusses how to address any issues and improve group dynamics.
  • There's always a tension between encouraging community involvement, and getting things done, and I personally think this tension will always be there. For a long time I had a theory that good content encourages more good content, and have thus tried to 'prime the pump' by creating content and discussions in house or thorough people planted to encourage discussion. However this hasn't always been hugely successful - we've spent a lot of time and money of getting good content in some groups but active contributions from the community has remained low. Other groups are far more active. I still don't know the ingredients needed to reach the participatory tipping point, although I once thought I did!
  • We've had lots of people volunteer via our website, but as soon as we ask them to do something concrete they seem to disappear. This must have happened 20 times in the past year, with not even 1 person actually following up on their offer! I have no clue about why this is: the volunteer tasks mentioned are clear that volunteering will mean doing some real work, so I don't think it can be a mismatch of expectations. I have wondered whether it might be a 'Debian Linux' effect (where a few years ago someone in that community was paid to bring the next release out asap, which resulted in an immediate reduction in the willingness of people to volunteer: "why should I volunteer if s/he is getting paid?"). We have some successful voluntary inputs, but these are either once off, or from people we've known for years, and have a

personal connection to.

Any ideas would be appreciated.

Jaap Pels, 2008/12/12

And.... monks are outsource-able

Tony Pryor, 2008/12/12


Terrific post; I'd say we have had very similar experiences. Let me just add on to your very well-presented note:

  1. Where the importance to the individual in finding the thread some other time is high (a thread on repairing a car THEY own, say, recipes from top ten restaurants, etc.) there is interest in making sure the material is traceable/findable. Where the importance lies mostly with the reader who has not yet used the site (future students...) the willingness to store and archive among the initial participants voluntarily is lower (since the value is in the future, and not with the initial knowledge producer).
  2. Probably more important, we, like you, tend to define a successful community by the number of active posters. But in some cases, the site's heft may come from the impact of lurkers, reading the post and then acting differently, but never then posting anything (an expert's blog, or the discussion forums on the Defense Acquisition University's KM system, (DAU ACC)).

It was that latter KM community (I think probably the most active one that I know of in the US Govt) that made me realize that in some cases bringing expert advice and commentary might be exactly what is needed, not chats and dialogue. (Again, I think in part it's our own bias towards the power of collaboration and talking that sometimes understates the power of reading and then quietly integrating knowledge). Often their "threads" can have 800-900 people reading the material, but only 2-3 people actually posting anything. It made me realize that we all tend to undervalue "lurkers" and readers.

You all might want to see these two examples from DAU (with the understanding that some may find these a tad strange in terms of subject. But there really is much to learn by benchmarking successful communities wherever they are found).

This thing has over 25,000 members, and tons of very active communities, as well as another group of things they call Special Interest Areas (the latter differing from the former mostly in terms of being as community that was self-defined, and unlike Communities not necessarily aligned and actively supported by part of the DAU faculty. In terms of our "monk" discussion, paid monks could be allocated to communities by the core owner, in this case the University's faculty who "own" a community and have funds to support it. Whereas the SIA's "monks" would need to be on a voluntary basis, since while interesting to the specific community it's of less interest to the student cadre at the heart of DAU's core purpose).

And for you website design wonks (LINK NOT WORKING), look at this example, and the format they use on the front page of every community. The same overall feel, just different pictures and different topic headings. A very clever and very practical way to maintain a common approach and yet have each community's topic be reflected in the actual graphic.

Michel Menou, 2008/12/12

The "knowledge monks" used to be called librarians, or documentalists, information officers, and so on, before the wired era. Trust machines and technology they won't deceive us more than humans used to do, but at much higher cost.

Veneeta Singha, 2008/12/12

Hello All,

Further to Michel Menou's post, an almost sound approach is given below. (Also downloaded from the web and very useful in most work scenarios)

"Ways with knowledge: people, processes and technology

One popular and widely-used approach is to think of knowledge management in terms of three components, namely people, processes and technology:

People: Getting an organisation's culture (including values and behaviours) 'right' for knowledge management is typically the most important and yet often the most difficult challenge. Knowledge management is first and foremost a people issue. Does the culture of your organisation support ongoing learning and knowledge sharing? Are people motivated and rewarded for creating, sharing and using knowledge? Is there a culture of openness and mutual respect and support? Or is your organisation very hierarchical where 'knowledge is power' and so people are reluctant to share? Are people under constant pressure to act with no time for knowledge-seeking or reflection? Do they feel inspired to innovate and learn from mistakes, or is there a strong 'blame and shame' culture?

Processes: In order to improve knowledge sharing, organisations often need to make changes to the way their internal processes are structured, and sometimes even the organisational structure itself. For example, if an organisation is structured in such a way that different parts of it are competing for resources, then this will most likely be a barrier to knowledge sharing. Looking at the many aspects of 'how things are done around here' in your organisation, which processes constitute either barriers to, or enablers of, knowledge management? How can these processes be adapted, or what new processes can be introduced, to support people in creating, sharing and using knowledge?

Technology: A common misconception is that knowledge management is mainly about technology - getting an intranet, linking people by e-mail, compiling information databases etc. Technology is often a crucial enabler of knowledge management - it can help connect people with information, and people with each other, but it is not the solution. And it is vital that any technology used 'fits' the organisation's people and processes - otherwise it will simply not be used.

These three components are often compared to the legs of a three-legged stool - if one is missing, then the stool will collapse. However, one leg is viewed as being more important than the others - people. An organisation's primary focus should be on developing a knowledge-friendly culture and knowledge-friendly behaviours among its people, which should be supported by the appropriate processes, and which may be enabled through technology."

With Season's Greetings and the very best wishes for the New Year,

Jaap Pels, 2008/12/13

Al true; most 're-organizations' and 're-(e)volutions' etc start with new names. First we had weed-poison, later we called the same DDT 'crop-protector'. First we had a sanwhich in the canteen, later we had lunch and re-named the same space grand-cafe :-) But seriously, it is not about the label. There is a big differences since the monks era described in 'In the name of the Rose' by Umberto Eco. In some places on this globe technology helps very much in finding text / e-mails. The archiving / retrieval we can do with technology; re-digesting needs humans. On a lot of places on this globes good old fashioned librarians / documentalists are still needed.

Lucie Lamoureux, 2008/12/19

See message below from Joiske, who is impatiently waiting for Dgroups2 :-) BTW, who is volunteering to summerize this thread on the wiki? :-p