Dave Snowden

He who works with his hands and his head is a craftsman

RSS Feed

The picture is of a medieval baker and his apprentice, a choice that will have a poignancy for veterans of knowledge management, especially those of us who really don't by the electronic bread maker idea.  That concept from Nonaka, like too much of KM ignores the social context of buying bread.  That triggers memories of cycling down a country lane early in the morning to the holiday gîte with a couple of ficelle sticks and a bag of pain au chocolate that you were told not to get, but you know will all be consumed in minutes. 

Memories aside, its time to get back to my post of yesterday and talk about the issue of standardisation, accreditation and professionalism.  I'm doing this is part with the whole Agile/Lean/Kanban/Scrum community in mind as well as the knowledge management community whose latest attempt in this area triggered me to write this two part mini-essay.  I actually think its too late for KM by the way, the time was years ago and the field is now a subset of IT, but I will finish off with some suggestions there.  For Agile the movement is new, but it is starting to fragment.  It faces a fundamental need to find ways to be seen by strategy and business operations as something more than a better way to meet their requirements.  Creating something in the early stages of a movements life cycle is a lot easier than at its end.

Given that, I want to start by describing a process I almost got funded by IBM in the early days of knowledge management.  I'm doing that in part because it never got recorded, but also because I think some variant of it is the only way forward in an emergent discipline.  Back then complexity theory and cynefin were a glimmer in my eye rather than a central part of my practice, but I'm pleased to say the basics were all there.  There will however be a certain amount of retrospective coherence here: I now know more of why the ideas would have worked.

What won't work and the LCD danger

Let me start by saying what will not work, namely a committee of people sitting down to write a standard or create a checklist to certify an education programme.   Aside from the issue of how you get people to give their time, the problem is that in an emergent discipline the field is moving too quickly for a publication based method to work.  In fact it will either make things worse (stifling innovation) or simply be ignored as practice moves on.  In knowledge management we saw with the BSI initiative of which I was a part and also the Australian equivalent.  The danger here is well illustrated by this comment posted to the KM thread referenced yesterday.

The standards will be set by consensus. The desire for consensus is based on a recognition that non-consensus hasn't helped our profession. Those of us who seek consensus will contribute to this effort

Now I think that would be a disaster in an emergent field, consensus is the enemy of invention, it would result in dumbing things down to obtain said consensus.  Those who seek consensus are generally not the inventors and thought leaders in a field.  The other idea in the KM group is that standards would be determined by vote.  Now that idea indicates a complete ignorance of the history of science.  If that was the case then I doubt Harisson's clocks (H1 the first, tested on a trip to Portugal in 1736 is illustrated) would have ever progressed and we would still be trying to measure the distance of the earth from the moon while on the heaving deck of a ship at sea.

The basic and fundamental lesson of history is that mavericks and accidents produce progress and consensus based approaches reward orthodoxy and act as a damping agent at best and a stultifying agent at worse, and as it happens in the main.

Any approach to standards in any new field is going to have to allow a lot of space for mavericks, it will need to create a process by which well researched approaches, which have been tested in practice can be visible quickly to the wider community.

My first experiment

IBM had just taken over DataSciences and I moved into a rather unusual R&D role.  I was sitting outside of both IBM Research (who were unhappy I did consultancy) and Consultancy (who were unhappy I did research).  IBM had a very consensus based approach and the sort of clear separation of functions that comes from committee led design. Given that services (and at that time KM) development required one to mix theory and practice the only home was Marketing who were keen to have anything, or anyone who would support the position that IBM was somehow different.   I managed that, in fact towards the end of my days in IBM I featured in double page features in the Wall Street Journal and the Financial Times as an On Demand Thinker. That involved spending a whole day in a studio flat in New York having my picture taken in various clothes.  We were told we could keep whatever they selected so I was disappointed when they decided my Rohan “suit” was the best and I missed out on a rather good stripped blazer.  

Enough on the memories; lets get back to standards.  Now this was all the best part of twenty years ago and KM was just taking off.  No one knew what it was and there were multiple sources.  In fact that has been KM's strength it didn't have one originating guru and book like BPR, Balanced Score Card etc.  A feature it shares with Agile, as well as the irony that both KM and Agile can trace there origin to a Nonaka paper, and in both cases with disastrous effects, but that is for another occasion.  So we had diversity, but also excessive fragmentation. &nnbsp;A few years earlier in DataSciences I had resolved that by making sure we were one of the founding members of the DSDM consortium to provide some credibility for RAD/JAD so some form of standards body seemed a good way to go.

However even then I knew a committee would not do the job.  There were very distinct schools of thought (which persist to this day), but the main players were limited in number and accessible.  So I came up with a novel solution.  The idea was as follows:

  • We would create an Academic/Thought Leadership that would review all methods for consistency and integrity.  The idea here was not that the panel had to agree with it, but that the method was properly documented, researched and consistent to some theory/literature.
  • All such methods would then be published on a web site and available for use by practitioners who would comment on those methods, write mini-case studies and score them, including links to other methods that seemed to do similar or complementary things
  • The whole model had a commercial engine underneath it which both paid the panels and allowed recover of fees for use.  It allowed for open source and for exchange of tokens to avoid money exchange if needed.
  • Training modules would go through the same process.

Now the principle here was that we would get two patterns of use and marking, one academic (theory) and the other practitioners (practice).  That would allow patterns of use to emerge, but also to change over time.  So the standards would be emergent.

IBM had agreed to fund this for the first year and I spent two years of my life meeting thought leaders, academics and lead practitioners to secure their agreement to participate.  We even had the first event in Florida then IBM pulled the plug at the start of what, I know fondly call The Knowledge Wars finally settled at a peace conference in Nice some three years later.  But it was too late.

That said, with some amendments I think the model stands, and with modern technology the set up costs would be low.  But, and this is a very big BUT, it would require the buy in of a critical mass of the major players.  That means big consultancy firms (for KM, less for Agile), thought leaders, leading practitioners etc.  It would never get of the ground if the starting group, however well intentioned, lacked gravitas in the field.  That may be unfair, but its reality.

The second big idea – workbooks

This starts to introduce the idea of professionalism.  If you look at accountancy (where I have some experience) then the process of becoming accredited involves an agreed course of study, but also or practice which is recording in a workbook and signed off by an experienced professional.   Now I think this is very important, time is required to gather experience.  You can't just go on a two or five day course, get a certificate and claim to be a professional.

So any attempt to create standards, professionalism etc. in either KM or Agile should (I think) include a requirement to go through a series of educational events and personal study, linked to practical experience under supervision.  Now while accounting has a defined field of study and an agreed curriculum this is not the case for KM or Agile, neither will it be in the foreseeable future.   Given that I would take a different approach to this.  The total package, which would take two to three years would have to include:

  • Formal study at a nationally accredited educational institution in a subject directly related to the field.
  • Self study (reading, supervision, defence of a thesis) in subjects relating to understanding humans.  I would put some anthropology into that, some cognitive science as well as the more traditional management subjects.
  • Completion of a series of more pragmatic courses in methods etc. from a range of vendors.  This would include SCRUM course for example in AGILE, and could include private courses.  In the early stages I would make no attempt to accredit those courses, but I would require reports from delegates and some participation in the the sort of web site envisaged in the previous section.
  • A diverse range of projects (and here there would be some need to create some board categories) carried out under experienced supervision with reports by both the candidate and the supervisor.

At the end of that process, the workbook would be examined by a panel, and the candidate questions.  If you think about it, and returning to my first post, this is an approach based on the development of a craft.  The quote at the start is one line from this:

“He who works with his hands is a laborer.
He who works with his hands and his head is a craftsman.
He who works with his hands and his head and his heart is an artist.”
― St. Francis of Assisi

Art is a level beyond what I have envisaged above, but it will emerge but generally with people who are stand out, who rebel, who think differently.  Make sure space is left for art to develop. 

Final thoughts

Now there is nothing revolutionary about the above ideas.  I've built on established process, thrown in some technology to make it more scalable and double checked it against complexity theory to test it would allow a partially constrained approach to enabling emergence.

Notice that although there is some need for an organising body, it manages process not content.  Its an evolutionary approach.   As I have said or at least hinted, I think the whole AGILE/LEAN/KANBAN/SCRUM area is ripe for some variant on this.  Like KM of a decade or so ago there are a limited number of major players from whom a critical mass of thought leadership could be assembled.   For KM I think it may be too late, but it might work.  For certain taking a consensus based approach based on voting will not.   Setting up a discussion group that within a few days, largely as the result of the behaviour of one of your leaders, drives leading experts to withdraw is not a good start.  Starting to discuss membership levels, before you have event got the buy in of the community as a whole is foolish.

Also you need to know what happened in the past.   To give one personal example here.  Someone in the KM discussion suggested that the organisers should have got some of the thought leaders involved early and my name was mentioned.  The response by the group leader was that this had been considered, but that I would have just torn things down and therefore was not considered.   Now this really pissed me off.  For a start I had not applied to join, nor had anyone talked to me.  Unlike all members of the starting committee I have successfully sat on, and seen to completion two standards documents in KM, one for the BSI and one for the EU.   I've also been a research director for the UK's EPSRC and sat on research panels for the US Government.   I have a track record of collaboration in these domains which is significantly more than any of the leadership of the group who in general have none.  

I suspect what they meant is that I (and others) might make them uncomfortable.  Well if you want to set up a standards body, claim to be able to determine accreditation then I strongly suggest you seek out people who will make you very uncomfortable fast.  You don't add further insults and when they contact you one to one (I tried this to calm them down a bit) you don't escalate the insults.  To loose one thought leader from your group within days of inviting them to participate is a shame, to loose two should give you a warning, to loose three really says you are doing something wrong.

So if this initiative is to get off the ground, it has to consult more, back off from some of the detailed plans and get more engagement.  With regret, I think that consensus building amount the leading players in KM is beyond the current leader, if the evidence of their contribution to the discussion is taken into account.  Someone less confrontational, more able to listen to differences and able to move things forward is needed.



I promised to provide some brief summary of what I think the KM group should do, so here it is.

  • Don't even think about boards and formal company structures yet
  • Set up events at the major KM conferences (KM World, KM Asia, KM Australia etc.) for delegates to talk about what questions need to be answered, create some virtual area for that discussion.  Those events will have the main thought leaders
  • Start to work as a group on Wikipedia to create lists, and generally improve the KM articles.  That will at least scope the issue and within some form of editorial control.  One of those articles could be a summary of standards etc. that are in place
  • Ask people with ideas as to how these sort of things could work to write short white papers that can be circulated
  • Aim to have an event in 2013 (or events) at which all of that comes together.  Think of the founding Agile event if you want an example.  But for that you will need the thought leaders
  • Find a few academic partners in the major continents interested in this
  • Go and talk with the KM leads at the major consultancy firms, see what issues they have

As you have probably gathered, this says build a foundation first