Dave Snowden: Cynicism and Serendipity

I've never heard Dave Snowden speak before so I was excited to finally get a chance and he did not disappoint.  His good reputation is certainly well earned; Normally the idea of listening to someone speak for 90 minutes at a conference would fill me with icy dread but Dave held us spellbound as he wove ideas around us faster and faster.

He's been described as a 'radical Welsh philosopher' which is a label he warms to, he revels in the ancient celtish practice of ribbing the English.  Nevertheless there is no shortage of philosophical underpinning to what follows.  Please also note that this isn't a literal transcription, I may have inadvertently introduced errors or changes of meaning.

Cynicism and Serendipity: A just in time approach to KM

The basis of successful KM is an understanding of philosophy and cognitive psychology.  More important than computer science or business management which he describes as dead disciplines.   Understanding of human nature is the gateway to the future.  "Re-establishing the importance of human beings."  He describes himself as an optimist and sees that, whilst there is a long way to go, things are moving in the right direction.

Dave described a little of the adventure behind the founding of the Cynefin centre and how, at times, IBM execs had tried to have him and his group thrown out of the company.  But he also said that, paradoxically, it can be much easier to innovate in large companies like IBM because there is no tight management.  You can often hide your project away by not arguing with the processes and justing getting on with things 'fying under the radar.'  Getting funding and agreement whilst operating this way requires a lot of human knowledge.  He said that John Pointdexter spoke of IBM as making the government look beauracracy free.

However from nearly being thrown out they had worked to get to the position of being asked to train IBM's top 100 in a new initiative.  This lead to a joint project between IBM and the University of Cardiff and the founding of the Cynefin centre.  Dave was very enthusiastic about this development seeing an opportunity to bridge the divide between academics and practitioners.

Moving on to the subject of their work Dave suggests there is a lot of confusion of properties with qualities.  The typical behaviour of management consultants is to attempt to generalise from a hypothesis about company behaviour:  e.g. successful companies have 'flat management structure' ergo flat management structure leads to success.  Dave suggests this is logically incoherent that it is like a man who steps off a boat in France and see's that everyone is wearing glasses.  He assumes that all French men wear glasses.  The management consultants then go on to assume that wearing glasses will make you French!  Life is more complex than this.

Next Dave suggested banning of the term "best practice".   He went on to suggest that best practice is past practice, and usually a bad thing in a fast changing environment.  He also lamented the way Government keeps copying industry when the forces at work are completely different.  Without the profit motive there is no pressure to force things that don't work out of the system, and many of these things are incompatible with the service culture.  Why does it happen anyway?  Why does government keep taking up things which industry has abandoned as useless?  (E.g. balanced score card).  He suggests that consultants who have learned that business won't buy their methods look around for some poor dupe and government is it!

At this point Dave said something that really resonated with me and the work that Paolo and I are doing:

  • Major consultancies know that knowledge transfers through informal networks!

About the name Cynefin (pronounced kin-nev-in):  They choose the word from a "Language that only a few other million gifted people speaks" because it's meaning was less likely to be overloaded and confused.  Literally translated it means 'habitat' or 'place' however the word has a deeper meaning: 'The place of your multiple belongings.'  This is suggestive of the idea of multiple paths which profoundly influence who you are but about which you are only dimly aware & can never hope to fully understand.   This is reflective of the idea that methods based on predicting the future based upon cause & effect are fundamentally flawed because you are dealing with uncertain pasts.

There are two schools of management:

  • Technofetishists who believe that people are just there to enter data and that everyone wants to spend their lives in virtual chat rooms.
  • New age fluffy bunnies who believe that technology is the spawn of Satan and that everyone should hug at the beginning of a meeting.

What is required is a more holistic approach that balances the spiritual with the empirical.  This view acknowledges that technology is a useful tool (by which we mean it fits the hand and can be used to do meaningful tasks).  But if you have to reengineer your hand to fit the tool then something is wrong.  Cynefin is about applying real scientific rigour to soft aspects of organisiations using the new sciences arising from biology & chemistry: complexity and chaos.

It is worth noting that systems thinking, complexity, and Chaos are fundamentally different approaches.

Dave introduces the story of Mulla Nasrudin and the falcon (from the exploits of the Incomparable Mulla Nasrudin).  These are Sufi stories.

Nasrudin found a weary falcon sitting one day on his window sill.
He had never seen a bird of this kind before.
"You poor thing," he said, "how ever were you allowed to get into this state?"
He clipped the falcon’s talons and cut its beak straight, and trimmed its feathers.
"Now you look more like a bird," said Nasrudin.

The purpose of the story is to highlight what happens in organisations when people are faced with things that they don't understand but need to pay attention to.  When faced with something successful people want it to look like something familiar.  This is a big problem.

Dave claims that worst practice systems are far more useful than best practice.  What spreads fastest: stories of success or stories of failure?  Spreading stories of failure is a more important learning exercise.  We are cognitively detruned to stories of success.   Best practice doesn't work!  As an alternative he suggests that narrative databases create learning environments based upon serendipitous encounter.  These spread the net wide to get at experience but are not directive.  He asks: Why are we building other types of systems when they do not work?

The word manage comes from the French meaing 'the ability to ride a horse in a dressage event.'  The horse executes perfect movements according to a plan.  Business consultants want to replace the horse with a mechanical horse that can execute these movements even more perfectly.  But managing complex knowledge requires disruption of expectations.

The importance of the difference between categorization & sense making.

Q:Why do we do KM?
A: To improve effectiveness of decision making and create conditions for innovation.  Hence: action & innovation.

Too many in the field have lost sight of this.  Dave says that knowledge management is a transformatory process shifting from Taylorism to the "new age".   The Cynefin centre looks across all discliplines.

How do you improve decision making in conditions of extreme uncertainty and change?  How do you avoid the talon clipping problem?

For 20-30 years we've operated a model of the human brain closer to cybernetics than neuroscience.  The assumption is that thought is a logical, rational, linear process.  This is wrong.  So is Myers-Briggs and all these other attempts to put people into boxes.  It is reminiscent of Brave New World:

"Alpha children wear grey. They work much harder than we do, because they're so frightfully clever. I'm awfully glad I'm a Beta, because I don't work so hard. And then we are much better than the Gammas and Deltas. Gammas are stupid. They all wear green, and Delta children wear khaki. Oh no, I don't want to play with Delta children. And Epsilons are still worse. They're too stupid to be able to read or write. Besides they wear black, which is such a beastly colour. I'm so glad I'm a Beta."

The human brain is adaptive.  The way we see the world changes according to context.  Disruption changes brain patterns and the key thing in human intelligence is patterns.  We match stimulus against patterns to know how to act.   The brain creates patterns.  Hence KM has a problem: We cannot codify patterns for use in text books.

As proof Dave offers an anecdote about how he met a bunch of macho developers at a seminar who wanted him to come do an intensive week with them.  Not wanting to spend a week in the Mid West of America with a bunch of macho developers he said "Okay.  It'll be £5,000 a day plus a 1st class air fare."  To his horror they agreed.

During the week he found a lot of resistance to the idea that patterns cannot be codified.  So he offered them a challenge:  He would take their executive team for a day.  If at the end of the day they still didn't believe, he would refund his fee.  If they did, they would double it.

That day he took them to a builders yard set up with a wall that needed plastering and all the correct tools and materials to do the job, along with copies of "10 easy steps to plastering" (an acknowledge good text book).  The challenge: Plaster the wall.   Towards the end of the day, when there was more plaster on the them and the ground than on the wall and they were arguing the ethics of bringing in a sanding machine to make the wall flat Dave introduced them to an old guy.  In 40 minutes he had a fully flat wall.

They paid double.

Dave calls Nonaka; the model that launched a thousand failed KM projects.  If you imagine you can make tacit knowledge explicit you only have to try the following test.  Use Google to translate a paragaph from English => French => English => French => English.  The process of making something explicit loses context.  You end up with rubbish.

"Story virus."  Never get a celt angry: Blood unto the 7th generation.

Human knowledge is stored in patterns far more than raw skills and artifacts.  Knowledge is in the fingertips and needs to managed in a different way.  This is both our power & our downfall.  It also means that you can't get knowledge from people by interviewing them because:

  • I only know what I know when I need to know it
  • The way we know things is not the way we think we do things
  • "I know more than I can say.  I will always say more than I can write down.
  • Knowledge can only be volunteers it can't be conscripted

This last is very important because it means that incentives and other attempts to make people share produce the wrong behaviour.  People will either camouflage, or dissemble.

3rd generation approach to KM (Post-SETI - Nonaka) separate knowledge into:

  • context
  • narrative
  • content management

Content: Very high cost associated with proper codifications.  Only where it's needed and we have stablity of knowledge.  Knowledge goes out of date before you complete the documentation process.  (A big problem in government).

Narrative: What I can speak but not write down.

Context: What I can neither "say down", nor write down.  Context is the basis for the success of apprentice schemes and the reason they are being re-introduced (e.g. Cynefin's IBM Inside programme).  The most effective way of doing complex knowledge transfer.  You only qualify once the master agrees you've got it, exams are not enough.

Cynefin is doing work in social network stimulation.   The idea is to design the informal network of an organisation for use in mergers.  They have identified that in a merger the the side with the most rigid processes tends to dominate.  This is because they tend to have an established beauracracy which can be made common to both sides.  This is a bad thing.  In the process of the merger informal networks get destroyed.  It can takes 4-5 years to build the networks back which is why mergers often fail.  The approach is to stimulate the informal network quickly.

You need all 3 pieces but most current KM are only content management systems.

It's a problem of perspective.  When electrons were particles, we looked for particles & found them.  When they became waves we looked for waves and found them.  In the same way knowledge is paradoxically both a thing and a flow.  But content manages only acknowledges flows.

Expertise management systems are effective.  Connect people when they need to be connected.   If we are asked a question about something which we know and care we tend to respond with honesty, some examples:

  • An old friend asks you a question.  You understand what they mean by it and how they will understand your answer.  Knowledge moves freely.
  • An unknown colleague asks the same question.  Normally you will help but you will be inhibited as you consider what they mean and to what end they might put your knowledge.
  • An idiot CKO wants to "know what you know."

Dave thinks that the knowledge is power motive not that strong, fear is much stronger.

How did KM throw common sense out the window?

The answers lie in cognitive psychology.  Nobody makes rational decisions.  What actually happens is a first fit pattern match with previous experience.  2 minutes later you have rationalised it retrospectively.  This is a key concept when dealing with complexity.  The nature of the decision is a pattern match.  Note not best fit, but first fit.  What happens when there is no pattern?  We hypothesise patterns until we find the first pattern that fits.

You can see this in the American reactions post 9-11.  There were no patterns.  Now patterns have been hypothesized.  But they are patterns for dealing with the past - the terrorists won't use planes next time.

The stories we develop as children & grow up with are very dominant.  Growing up in the 70's you learned that if you didn't occupy the University at least once per term you weren't a real revolutionary.  You could differentiate 16 types of marxist.  In Washington today anyone slightly to the left of Tony Blair is a raging communist.  You can differentiate 16 varieties of the religous right.  Patterns are based upon stories which determine how we see things.

Perspective is vital

We all suffer from pattern entrainment.  Dave offers up an advert to consider:

  1. At first you see a mean looking skin head with a police offer hot on his heels.  An obvious thug.  What does this suggest to you?
  2. Then you see that he is running towards a guy with a briefcase clutched to his chest.  Obviously a rent-collector stereotype, the case is full of cash.
  3. Finally you see the skinhead pushing the guy out from under a load of falling building materials.  Saving his life.

The caption is "For a different perspective, read the Guardian."

Faced with a thug you don't rationalise, the patterns you have entrained guide you quickly to a decision.  This is a big issue in decision theory: "When do you stand still?"  "When do you rationalise?"

Patterns allow faster processing.  Human language is not a serial change from animal languages.  Dave seemed to reject Chomksky's view that it differs by bigger vocabulary and accumulated complexity.  However cognitive science teaches that human beings start with abstractions and then move to the "read world."  Human beings are permanatly in a phenomenological state.  When the patterns go wrong there is disaster but the next group learns from this.  Decision making requires learning about patterns.

It's wrong to take more time to understand patterns.

How do you tell if someone has winked? or blinked?  It's a very important question:  In a bar late at night you can get your face slapped.  In international dimplomancy is can be the difference between war & peace.  In the Cuban missile crisis did Kennedy wink or blink?

Is behaviour intentional or unintentional.  Our assumption tends to be: Other people do things intentionally we do things by
accident.  This can lead to creating circumstances you try to avoid.  Accidental results of pattern entrainment, not rational intentional acts.  This has a huge impact on KM.

Blair can do no wrong.  Blair is a spin doctor.  Blair is a liar.  Patterns.  How humans work.

Human beings don't have unitary identity.  We are not ants.  We all have multiple identities which we move between without thinking about it.  This is not a sign of mental illness.  It's fundamentally how we work.  Dependent upon the context we will see the world a differnet way (father, son, husband, lover).

Fluid identities, non-rational thinking.   In contrast with current management thinking which assumes rational behaviour and unitary identity.

Strong attractors create patterns around themselves & the patterns create meaning.  Gas under pressure, becomes vapour.  Then water.  This phase shift happens suddenly.  Stability. Same thing in human interactions.  Not pre-ordained.

There is a distinction to be made between ordered and unordered systems.  Ordered systems lead to targets, optimization and a cause & effect view.  Unorder is characterised by possibilities, inconceivability, complex systems.  Unorder: "It's order Jim, but not as we know it."  Characterized by non-repeating relationships between cause & effect, i.e. things repeat as long as they repeat until they don't.

These asymmetric collapses create deep shock.  However once the phase shift has happened you can't go back - the energy required to return to the original state is too high.  Working on the past doesn't work because the patterns have changed.  You are learning the wrong lessons.   Terrorists next time they will do something different.  Human dynamics: Conflict is critical to detecting weak signals.  Without conflict you end up playing follow the leader and destroy the capability for innovation.  Unordered systems are matters of managing patterns.  This is the key to survival.

Tom stewart: Managing with your gut (Business 2.0)  There is a science behind 'gut feel'  The way we manage a complex space is the way we manage a party of 12 yr olds.  Do you have a process plan and incentive schemes?  If you do you're a very sad individual.  Instead most people draw a line in the sand "Cross it and die" with multiple interventions to stimulate the creation of beneficial patterns.  You manage a complex space by managing the patterns.  Create boundaries, do interventions, stimulate attractors.

Not the same as managing order but not abrogating managing.

Disorder:  The state of not knowing where you are.

This leads to the basic Cynefin framework:   A sense making framework not a 2x2 consultantany model.   They did a test with a group of management consultants.  They gave them 500 items of which only 200 would fit in the 2x2.  The consultants made all of
them fit and believed that they did.

Category thinking is closed down to new opportunities.  Tendencies to categorisation is rigid in management science.   Sensemaking is the process of creating boundaries between things so you can make sense of them.  Definied by their own histories and futures.  Boundaries emerge from context and are not pre-given.

Domains of order & unorder
Unordered Domains Ordered Domains
Complex: The domain of many possibilities: Cause and effect coherent in retrospect, repeat accidentally.  Looks like Empirically knowable from the past.  Empirically knowable: The domain of the probable.  Cause & effect separated over time but repeat.  The domain of experts.
Chaos; The domain of the inconceivable:  No Cause & effect visible.  You have to do something. Emperically known.  The domain of the actual.  The only place where best practice makes sense.  Inappropriate in other domains.

The average life span of a Fortune 500 CEO is now 11 months.  When they have been successfully was it what they did?  Or did they just get lucky?   An interest experiment by spammers:

Take 500,000 email addresses and split them 50/50.  To the first 50% claim a stock will go up, the second 50% claim it will go down.  Let's say the stock goes up.  Ignoring the second 50%, take the first 50% and split them.  Repeat the experiment.  You eventually end up with a small group who absolutely believe in your ability to predict the stock market: ready to be scammed.

Decision Models
Probe
Sense
Respond
Sense
Analyse
Respond
Act
Sense
Respond
Sense
Categorise
Respond


Understanding the phases is key, knowing whether you are in complexity or chaos.  The Cynefin framework leads to you being able to make the right decisions.  Many corporations don't understand the unordered domains and treat everything as belonging the ordered domains.  This is a disaster as the decision models are completely wrong.   The anti-terrorist problem is the the same as emerging market opportunties from a management perspective.

Network linkages:
Weak central
Strong distributed
Strong central
Strong distributed
Weak central
Weak distributed
Strong central
Weak distributed

The goal of management is to move from chaos to complexity to the knowable in a watchful fashion by sending out multiple probes and watching the results.

  • Patterns we don't like disrupt.
  • Patterns we like we reinforce.
  • Shift to the left to exploit it.

Mathematical approaches to complexity fail.  Humans can impose order, ants are condemned to be constantly complex.  Shifts to the right are temporary expediencies.  Most organisations oscillate between known and knowable.

Creativity is not innovation.  Creativity is neither necessary nor sufficient for innovation.

Trust is not a property.  It's an emergent property.  You can't make people trust each other.   You can't train people to have qualities.  It doesn't work.

To be useful Communities of Practice require regular sheep dipping into Chaos.  A CoP should be destroyed after 12 months.  If it's valuable it will go underground and reemerge. Funded CoP's become a huge conservative force.  Look at the histroy of science.  Innovation is not a logically planned process.  Dipping into chaos to create new complex patterns.

The Cynefin programme

Cynefin dynamics:  Innovation cycles.  Just-In-Time KM focuses on creating environments that allow complex systems to emerge and move into expert environments.  Complex acts of knowing.  Rounding level in large organisations is $250M.  $49M gets rounded down to 0.  A different world!

Shifting consultancy from a utilization model to a software model.  Fed up with teams of consultants.  Email detoxification: Cold turkey organisations to get them out of bad email habits. Surving in their own environement represented as a metaphorical model.   Sensing organisational structures which will work post-merger.

Culture. Trust. Collaboration.  Big problems that companies have but don't know how to solve.  Cynefin will tackle 1 problem per year.

Looking through a complexity lense puts things into a context where you can shift R->L, from looking at order to unorder.

Western canada is a fabulous example of aborigine integration.  Shamans are heart surgeons and nurses.

PERSPECTIVE. PERSPECTIVE. PERSPECTIVE.

New patterns emerge from complexity.

Flagship discovery program.  A huge mutli cultural experiment.

Corporate values.  Rituals in absence of belief.  Singing the company song.  Create rituals that align people with belief systems.

I got caught up at the end.

[I wanted to ask a question about patterns and pattern languages but felt a bit intimidated.]

Questions: Paradox program.
Questions: Knowledge Audits.  Never ask a direct question.  Identify knowledge use (decision, judgements, resolutions). Cluster to entities to create a meaningful context.
 Heuristics are more common than explicit knowledge.  What experience is vital to what you do?
 Create coherent knowledge objects which can be managed and organised to affect core processes & activities.
 Portofoli of knowledge managament (amortises risky projects in terms of the tangible benefits of low risk activies)
 2-stage emergence:  Dissolve patterns in the spce.  Stimulate the space to form new patterns.  Rinse & repeat.
 Never tackle the low hanging fruit in KM.  Tackle high vulnerabilit to loss of knowledge areas which core processes depend upon
Questions:

19/06/2003 11:12 by Matt Mower | Permalink | comments: