Conversations
with David Snowden
Practice and
Communication
of 3rd Generation Knowledge Management
David Snowden
Director
of the Cynefin Centre for Organisational Complexity
IBM Global Services
United Kingdom
Synthesis
by Chris McGuire
Administrative
manager
ReservoirTeam Ltd
Adelaide, South Australia
Editor's
note:
This is a synthesis of the "Conversations with David Snowden"
held in January, 2002, the first episode in a parade of KM luminaries
who served as guest moderators during the second year of the
monthly AOK STAR Series.
Prologue: Jerry
Ash, AOK Chief Executive, posed a two pronged question to David
Snowden: How is "context" different from "content"?
and: What is happening in the workplace that supports the notion
that the focus is shifting from knowledge as a "thing"
which can be captured and cataloged to the management of knowledge
ecology?
These two threads
have been separated in to Part I and Part II respectively. Part
III presents an exchange between Snowden and Ash on a variety
of related subjects.
Part I: Early in
his reponse to the first question, a further debate was spawned
surrounding issues of jargon/expert langauge/business English
in the relatively new discipline of KM. This broadens into an
analysis of online disussion forums.
David Snowden's
responses were, of course, shaped by the at times heated debate
that was running in parrallel. As far as possible Part I stands
alone and can be read without delving in to the challenge and
rebuttal of Part II.
Part II: The debate
over use of "expert language" continues. To this, it
emerged that case studies were not David's preferred way to illustrate
organisational change. Rather, he provides readers with something
more insightful; being his own approach and the strategies deployed
when working within an organisation. It becomes apparent that
this method of response is 3rd generation KM itself and, together
with Part I, David's vision becomes clear.
Table of Contents (Click on list item to go
directly to each topic)
Part
I: 3rd Generation Knowledge Management
 Part
I: 3rd Generation Knowledge Management
I always know more
than I can say, even after I have said it, and I can always say
more than I can write down.-- David Snowden
Introducing David Snowden
Jerry
Ash, AOK chief executive:
It is altogether fitting that we begin the 2002 STAR Series with
David Snowden, who has teamed the past year with the Series'
first guest moderator -- Stephen Denning -- to carry the message
of storytelling as a Knowledge Management tool around the world.
Storytelling, it
turns out, was an integral part of the second generation of KM
-- a progression from Content to Narrative to Context. David
leads us quickly through the three generations of KM in "Preparing for Conversations with David Snowden,"
and then leads us into the new millennium and the next generation
of KM:
As we move into
the third millennium we see a new approach emerging in which
we focus not on the management of knowledge as a 'thing' which
can be identified and cataloged, but on the management of the
ecology of knowledge. Here the emphasis is not on the organization
as a machine with the manager occupying the role of Engineer,
but on the organization as a complex ecology in which the manager
is a gardener, able to direct and influence, but not fully control
the evolution of his or her environment. We are also seeing a
refreshing move away from programmes which seem to manage knowledge
for its own sake, to those that tightly couple knowledge management
with both strategic and, critically, operational priorities.
David Snowden is
Director of the Institute of Knowledge Management for Europe,
the Middle East and Africa. One of the founders of "Organic
Knowledge  Management," he is an acknowledged expert
on the management of tacit knowledge and has developed a series
of pioneering methods including the use of anthropological techniques
for knowledge disclosure through the ASHEN model, the use of
stories as an advanced form of knowledge repository (based on
six years of research into story telling cultures around the
world) and the Cynefin "Just in Time" model of knowledge
transfer between formal and informal communities.
A gifted speaker
and educator, he is in regular demand as a keynote speaker world
wide. This STAR Series discussion is a prelude to his presentation
on Third Generation KM at the AOK/Delphi Summit, March 2002.
Please join me in
welcoming David Snowden to the AOK network. We look forward to
moving forward with you, David, into the next generation.
Back
to top
How is Context Different
From Content?
Jerry
Ash: I would
like to begin the discussion by telling you that AOK members
welcome forward thinking on the development of the knowledge
strategy and they like stories that connect theory to practice.
Therefore, I ask: How is "context" different from "content."
David
Snowden:
Thanks for the welcome and I'm looking forward to the discussion.
You ask a broad question -- at its simplest I see knowledge management
as creating shared context -- if the two sides of a communication
do not share any context then no information can be created,
it's all data. Shared context creates shared meaning. Content
is abstracted from context, and specifically from people. I see
knowledge management as creating information from data by the
provision of shared context; I reject and despise the linear
model of Data-Information-Knowledge-Wisdom. I have sent a summarised
version of a much larger paper to Jerry for posting -- this provides
more thinking on the 2nd (post 1995 or Nonaka) generation and
my 3rd generation approach.
Abstract
and a PDF of the full paper
Back
to top
Evolution From Newtonian
Physics to Complexity
Debra
Amidon, Entovation International:
First, welcome to this forum . . . it's great to have you 'here.'
Perhaps there is
some alignment that can/should be done on all the generation
management material to date, such as Senge's 5th Discipline,
and Charles Savage's 5th Generation Management. We also published
a 1996 article on 5th Generation -
and
the tool for analysis. The analysis is done according to
Performance, Structure, People, Process and Technology; and the
dimensions evolve organizations with focus the assets to be managed:
1st -- Product/Technology, 2nd -- Project, 3rd -- Enterprise;
4th -- Customer and 5th -- Knowledge. Each generation is described
in more detail in the article as well as references of interest.
More germane to
our topic here is that in the same article, we analyzed the evolution:
- Stage I. Technology
Transfer (moving from one place to another; the"passer/receiver"
language applied to labs, within consortia or country to country).
- Stage II. Technology
Exchange (technology transfers through people; the "contact
sport" analogy; dual communications links; dialogue among
parties; ideas from either side).
- Stage III. Knowledge
Exchange (Shift to viewing that which is transferred from "widgets"
to ideas and insights as a function of human interaction; realization
of something beyond "information"; timely access provides
the competitive advantage).
- Stage IV. Technology/Knowledge
Management (recognition that the "process"cannot be
left to serendipity; organizations must pay "sweat dues"
in addition to the enrollment fee; emergence of a new discipline:
the management of technology; builds staff and mechanisms to
manage the process).
- Stage V. Knowledge
Innovation Systems (realization of the dynamic nature of the
total process of innovation; emergence of the "virtual"
research enterprise without functional, industry, sector, or
geographic borders; takes systemic view of =ECknowledge =EEflow;
shifts focus from monitoring discrete deliverables to creation
of a learning system designed to enable profitable growth).
There is a similar
tool published in the article as a PowerPoint slide, but not
yet available on the Web. Just contact me directly to request
a copy.
David, of course,
is defining the various stages of managing knowledge as the asset.
I wonder if there is some similar analytic discipline we can
apply to help people understand that, indeed, we are living in
5th Generation (albeit the 3rd wave thereof) 'kaleidoscopic change'
and managing with 2nd and 3rd generation management technology.
Is this gap a familiar one to our AOK practitioners?
P.S. By the way,
with Leif Edvinsson, we have also been researching 6th.
Back
to top
Generation -- Managing the
'Future' as the Asset!
David
Snowden:
I've sent Jerry two other articles -- designed for a general
business audience, one of which is fairly old but covers the
management of informal communities, the other is more recent
and describes narrative databases. I've put these in for those
who want to open up a more pragmatic dialogue.
Liberating
Knowledge
Narrative
Knowledge
I like Debra's idea
of the third wave in the fifth generation with some restrictions.
One of which is I that I would favour Tom Stewart's view that
we have seen three big "movements" -- Quality, Process
and IC in business thinking and I prefer that to the Savage's
model. The main one though is that I do not think that complexity
based work (my third generation) is not going to represent an
incremental, albeit step change, in thinking in the way that
we went through Product/Technology, Project, Enterprise, Customer
and Knowledge.
I think complexity
represents as large a change as the shift from Newtonian to modern
physics, it challenges the conceptual underpinnings of process
management and scientific management in general.
In particular it
means that we can not build predictive and prescriptive models
of the future; managers cannot by intentional acts achieve organisational
goals in all but the simplest of cases. For example, the work
we are now doing with the US Government and elsewhere, involves
creating alternatives to scenario planning that create multiple
perspectives on a problem, rather than extrapolating future possibilities.
This means that a decision maker entrained in the culture of
the US has a model by which they can look at a situation from
an Islamic perspective, one that is impossible in an analytical
framework.
To use a biblical
metaphor, new wine needs new wine skins. Other than that point
Debra is arguing that you should read the referenced articles
and her existing and forthcoming book -- a recommendation I'm
happy to endorse.
Back
to top
Recipes
v. Heuristics (Plus Confusion Over Words)
Excerpt from James S.
Robertson, Managing Director, Step Two Designs Pty, Ltd, Australia, who was resistant
to new words for new concepts:
I have seen a large
consulting firm prepare a presentation to senior executives at
a government organisation. This was a "boilerplate"
presentation, with all the usual buzzwords and key phrases.
But practically
none of it was written using terms and concepts that the organisation
would understand.
An example of their
mistake: their presentation used the manufacturing of donuts
as an example of a complete product lifecycle.
And this was supposed
to mean something to a government organisation responsible for
the licensing of drivers and vehicles?
Thankfully this
was pointed out, and fixed before the final delivery. I still
note, however, that the presentation was not successful: the
steering committee simply didn't get enthusiastic about a "strategic
KM initiative."
David
Snowden:
I think that James S. Robertson defacto supports my point in
his reference to using a model of donut manufacture to government.
That's classic consultancy and something I fight against. I used
the word "contextualise" because I think it vital in
knowledge management not to use generic models, but to create
a model in context (i.e. contextualise).
If a model is rooted
in the stories of an organisation's histories and its possible
futures (narrative techniques) then the model has meaning to
that group. My approach is get the organisation to tell stories,
and then to populate a framework with those stories, draw boundaries
between spaces and then move forward to action.
This means that
the model and the boundaries within it are rooted in the organisation's
own narrative (and I'm not going to apologise for using the word)
context.
There is a key difference
here between two approaches to KM consultancy -- the recipe book
V heuristics.
In a recipe book
approach consultants study several companies to identify properties
(such as common vision) associated with qualities (such as profitability)
derive a generic model from that experience which results in
action to acquire the properties in order to achieve the qualities.
This makes assumptions that they have really discovered the reasons
behind the achievement of profit or whatever and is a logical
error in any event, but never mind, it's the way that consultancies
work. They also then use factory methods to roll over a standard
template. One of the reasons I never use slides is that I have
a hatred of consultants who just roll out their model regardless
of context.
In heuristic approaches,
a set of rules of thumb or operational principles are developed
which are applied to create unique, contextual solutions in each
case -- no universal model, the application of sound heuristics.
Here we have the chef, not the recipe book user with all the
differences in quality that metaphor implies. There is a place
for recipe books, but it's not a large one in a knowledge based
organisation. Context (and/or contextualise) and Heuristics are
amongst the most important words in knowledge management.
Knowledge Management
is the creation of shared context -- without shared context no
communication can take place. Information degenerates to data.
I do not see knowledge as some higher form of information, but
the means by which we create meaning out of data through communication
-- information.
Heuristics are a
vital knowledge asset and come in various forms. I did one project
with an oil company in which an engineer interviewed in his office
by a KM professional from IT explained how he solved problems
through a structured process. When I carried out an anthropology
project in the field (which was scary, go to an off shore rig
in a gale and you will see what I mean) and observed problem
solving in the field the structured process went out of the window
and he fell back to what he called "the music of the rig"
and combination of sounds and other barely articulated signals
meant that he knew instantly what was wrong. By asking him what
he knew in the context of his knowing it was possible to identify
over 12 heuristics that were governing his decision making in
practice - all capable of codification and distribution within
an expert community, none relating the results of the earlier
interview process.
Heuristics can also
be designed - my three for KM are:
- Knowledge can only
ever be volunteered it cannot be conscripted
- I always know more
than I can tell, and I will always tell more than I will write
down
- I only know what
I know when I need to know it
Those three heuristics
can be used to test the validity of a proposal for a knowledge
project.
If I had used a
word such as "eschatological," "teleonomic"
or "epistemology," I would accept the criticism, but
contextualise, heuristics and narrative should be part of the
vocabulary of any knowledge professional or any half way educated
human being for that matter.
Jerry
Ash: David,
you come to us thinking that we are K-work pros, and we are.
But we are also a microcosm of the K-world, a mix of the world's
deepest thinkers and a cross-section of those who range from
curious-about to challenged-by a new way of working. We are a
most perplexing audience both in the virtual and the real world.
We are speakers of the King's English, American English modern,
adapted English and English as a second language. We are all
intelligent, inquiring, and often confounded by the shorthand
often spoken by those "in the know."
If the knowledge
strategy is to succeed, it must find a common and comfortable
language in which we can all find understanding -- plain English.
To my way of thinking, that is not to reach down, but to reach
out; to meet the learner, the decision maker, the implementer
where s/he is -- at the well head.
Oh, goodness. This
sounds so philosophical. I'll stop now.
David
Snowden: I'm
sorry that I don't understand your comment, it sounds wonderful,
impossible to disagree with, but I'm not sure what it means --
rather like many a newspaper column. Yes we have to communicate
and to do that requires common context. Whether that is plain
English or not depends on who you are talking with. I certainly
don't think I've used any short hand -- although I have referenced
material that I sent as reading. The language of a pure academic
is too far in one direction, that of a 14-year-old too far in
the other. I feel comfortable with language that the average
University graduate should understand.
To over-simplify
is to lose meaning; to over-complicate is to lose audience. It's
a balance issue. In common place, easily understood issues,yes
go for very plain English. To explore new areas and create new
meaning, I'm afraid you have to put in a greater intellectual
effort. I also think you need to read the point I made about
the difference between oral and written presentation as well.
In respect of straying
from the point, I'm surprised at the lack of practical questions.
Third generation techniques include narrative databases, stimulation
of social networks, just-in-time Knowledge Management. Lots of
good practical stuff.
Back
to top
More About Communicating
KM Principles
3rd
Gen KM: Context Will Help Understanding
Excerpt from Jack Vinson,
knowledge manager, BioPharma, Pharmacia:
Context will help understanding. And this gets us back to context.
When we are trying to "sell" any project, it must make
sense to our colleagues and to the people who hold the purse
strings. To do this, we need to sell the financial end to our
accountants, the efficient work processes to our business process
people, and the work benefits to the people who will actually
use the results of the project. We can do this any number of
ways, and if I read Dave correctly, he is suggesting that putting
the tool into the right context will help people understand how
the project will fit into the company.
3rd Gen
KM: KM Worker or Working with KM?
Carol
Tucker, AVP, credit administration, Maryland (U.S.), Permanent
Bank & Trust:
At the end of the day, I have to communicate to my senior management
in terms that they will accept using the communication style
most likely to gain their acceptance. If I give them a presentation
that does not speak in the way that they want to listen, the
message that I am sending will be lost in the "noise"
of my presentation -- the medium will mean more than the message
and form will triumph over content.
The true value of
narrative, to me, is that it enables the presenter to frame the
point in terms that the auditor can assimilate and own.
Although I send
the same message in the lunch room that I send in the Board room,
I tailor my comments to match the communication styles of my
intended audience. If I fail to do so, then I lose credibility.
I believe that David's comments about his speaking experience
[as well as the slightly testy tone of his "allergic"
response] says that he does exactly what Joe and James reference
-- he tailors his comments to the appropriate audience.
However, David lives
in a different world than I do on a day-to-day basis, and in
his world what passes as everyday verbiage would be questionable
jargon in mine. David is truly a knowledge professional who utilizes
KM to be a better professional and who is sharing his fund of
insights to help others; whereas, I am a banker who utilizes
KM to be a better banker.
To the KM professional,
it appears that I operate in stealth mode. To the banker, it
appears that I utilize a different skill set than is commonly
seen, offering an alternate model of business practices. I am
not a KM worker -- I am a person who does my work with KM.
Excerpt from Fred Schoeps,
former training and KM program Director, IBM: I
was reminded of Ecclesiastes 3:1-8 "for everything there
is a season . . . ." What short story describing the future
(in the best Asimov sense or Fulghum sense) could you conjure
up to enable the reader to envision how business will do business
once the discipline of complexity is a natural part of how we
do business?
3rd Gen
KM: Traditional Theory Limited
David
Snowden:
Some good progress here. Jack has exactly summarised what I mean
by contextualisation and there is a link to Carol's comments.
Any good speaker
responds to an audience, probes and tests their level of understanding
before deciding what to say and how to say it. When we use models
such as Cynefin, the contextualisation takes the form of gathering
anecdotes (naturally told stories, around the water cooler etc)
from that organisation's own history, and using those stories
to create the model. Not an imposing model with examples from
previous consultancy engagements -- I think Joe was reacting
against this practice which I despise. Equally, the process I
described to link knowledge projects to process improvements
is designed to make KM relevant to business functions and to
create both an explanation of the project and a language to justify
that project which is meaningful to the audience. Jack has correctly
read this into what I said.
This is a discussion
group of people interested in KM, so the language changes from
a general discussion with business people. I am also a mix of
practitioner, academic and researcher with a philosophy degree
to boot, all of which informs how I engage in conversation.
Fred (I and IBM
miss you Fred, it's not the same since you left) as always provides
good insight. He asks for a story of a complexity based future.
I think the best way to do this is to use two metaphors one from
Modern Physics and the "Gardener v. the Engineer."
Newtonian physics can stand for existing management practice
and theory in which the relationship between cause and effect
is assumed to be knowable and manageable. When quantum physics
and all that came along, Newtonianism was not invalidated, but
it was bounded; the same is true for complexity in management.
The fact that we now know that significant aspects of human behaviour
-- especially those relating to knowledge sharing and learning
-- are complex behaviours, we know the limits of traditional
theory.
Traditional theory
works for process and significant aspects of KM, but it is also
limited.
In a future state
I think we will see a significant shift from recipe book approaches
to heuristics. Self-organising communities, able to rapidly assembly
and deploy knowledge which is never truly known in the wider
organisation, will complement the more structured disciplines
of content management. The role of senior management is to create
and manage the garden, rather than to design machines, but they
will still use machines in the garden. For knowledge management
this means three things:
- Sound well structured
content management systems
- taxonomies
- search engines, etc.
Back
to top
Narrative Databases Using
High Abstraction Indices
David
Snowden: Context
management -- interventions designed to create and manage the
channels through which knowledge flows, both formally and informally,
but which do not manage the knowledge itself.
Fred mentions Asimov,
who ironically wrote stories about both worlds. In the Foundation
Trilogy we have Hari Seldon, the ultimate Taylorist who through
mathematics can predict the future of human interactions centuries
into the future. In the 50's this was the dream of science --
everything could be known -- and remains the assumption of mainstream
management theory and practice. In the "I Robot" series
we see the use of heuristics, the three laws of robotics, that
allow complex patterns of behaviour to emerge through their mutual
interaction, contradictions and paradox.
Back
to top
Shifting Sands of Context
Jack
Vinson: I
am quite interested to hear about your ideas on third generation
KM here and at the AOK/Delphi Enterprise Learning and Knowledge
Exchange Summit. And your thoughts on context are interesting
too.
We have started
thinking about knowledge strategy as part of our ongoing review
of the organization, and I realize that one of the difficulties
with dealing with context for typical information technology
projects is that we expect the context to stay fixed over the
lifetime of the application. I have found that the context hardly
stays the same over the course of the implementation plan due
to changes of personnel, scope and corporate direction.
How do we change
the mode of operation of our businesses from point solutions
to solutions that are flexible enough to adjust with the changing
needs of the organization? This probably sounds like a bottomless
well of money for a savvy salesperson. What I am really thinking
about is that projects should include flexibility and continuous
improvement after initial rollout. Maybe I am being a little
harsh -- we have plenty of projects that we are continuously
improving. I wonder how we can ensure that understanding of context
for new developments becomes a natural part of the specification
process.
David
Snowden:
Jack makes some interesting points and asks a good question.
The issue of context stability is an interesting one. Yes there
are situations where context is constantly shifting but there
are also stable contexts. One of the unique aspects of human
complex systems is our ability to impose structure, and reduce
uncertainty through collective action (country legal systems,
organisations etc.)
The Cynefin model
(see previous document -- shows four
spaces: known, knowable, complex and chaotic. In the know and
knowable space, context is stable and we can afford more second
generation approaches. In the complex and chaotic states, context
is unstable; and, different techniques are required. When we
understand this, it becomes important to understand what each
space means for the organisation and develop different approaches
and models for each space.
There is no single
answer -- solutions lie in diversity, but a descriptive self-awareness
of that diversity exists so we know intuitively when we have
crossed a boundary and need to change out models. it's not a
difficult process; as for strategy, I use narrative techniques
to contextualise the model for a company so the heuristics and
boundary conditions are defined not in some abstract language,
but are rooted in the defining stories of that organisation.
There are also some
interesting new developments in project management, in which
we use complexity principles to allow teams to self organise
to bid out complete project segments based on date and cost.
This contrasts with approaches based on forward resource scheduling
based on constraints. It is achievement focused and introduced
flexibility based on human obligation, commitments and ability
to manage uncertainty.
Back
to top
Part II: Extended Discussion
on KM Communication
|
