Library clips

Lately I’ve posted on the need for the enterprise to not use email for everything, and the enabling tools that are more appropriate, especially network centric tools that add the dimension of getting things done via informal circles without relying on just your close colleagues and your hierarchy.

Here are some links to these posts as background:
CoPs and informal networks
km 2.0 enablers: blogs, wikis, and social networks
Instead of sending an email

Since my last post on this topic I’ve been listening to more from Dave Snowden’s podcast archive (and plenty more to go).

Here are some slides that encapulate these new insights and methods.

The 5 podcasts I’ve just listened to…actually let me tell you, I’ve listened to some of these at least 3 times. Dave says some provoking stuff, yet it’s only because it goes counter to past knowledge management approaches, plus it has taken a while for me to absorb this great stuff as it’s a shift of mind set to a more practical or usable approach.

Dave alludes to an individual centric, fragmented, trust based, naturalistic knowledge sharing culture…more so just-in-time knowledge (and sense-making), based on small circles of people you trust, connected in a social network environment.
Of course there is a lot more to these podcasts, complexity and all, but I’m mostly focusing on the knowledge sharing aspect.

Another way of looking at it is moving from a content to a context approach.

Dave says that management science is based on information processing, which is not how humans make sense of the world. Instead he brings the natural sciences to management, based on a system/methodology more in tune with our cognitive process.

Here are the podcast links:
KM Australia 2007 Keynote
Jon Husband interview with Dave Snowden on Web 2.0
KM World 2007
KM Asia 2007 Keynote
Oil & Gas Exchange Houston September 2007

For a great review on Dave Snowden’s KM World 2007 keynote check out Stuart Henshall’s post, Dave Snowden on everything is fragmented.

NOTE: These podcasts bleed into each other, each is unique, yet each covers some same ground. From my notes, it has become a blur to me which podcasts the notes are from, and in fact whether they are my wording or verbatim/paraphrasing.

Here are some essential blog posts:
Reporting on sin…
sense-making & path-finding
Safe-fail probes
Whence goeth KM?
Natural numbers, networks & communities
Volunteer not conscript
If the world is flat, seek out the bumpy bits
Good judgement comes from experiences. Experience comes from bad judgement
Confusing story telling with narrative

In a post from a while back I used the term Organic KM (from one of Dave Snowden’s papers).
Back then, when I was new to this concept, reading these papers was not an ideal shared context for me, after reading the more informal blog posts, listening to audio and the process of writing this post, I understand the concepts much more.

This is the essence of what this post is about, it’s an organic approach (experimental more than designed), the system being a servant to the human, enabling things to manifest and emerge, which we then harness or dampen.

Jack Vinson has pointed out on his Future of KM post, that Dave Snowden is impressed with social computing and says that if these tools and their emergent freeform nature were around when knowledge management was kicking in, they would help achieve what we actually expected and wanted from managing knowledge.

The new KM (perhaps KM 2.0)

Just a quick summary of what this post is essentially about…

MOVING FROM
- traditional management science (social sciences)
- information processing
- knowledge things
- DIKW
- recipe model-copy and roll out-one size fits all (replicate outcome)/fail-safe
- codification (tacit to explicit)
- context dependent
- best practices
- formal communities (CoP)
- hierarchy

MOVING TO
- natural sciences (cognitive)
- pattern matching (sense-making)
- knowledge flow
- internalise, sense-making, pathfinding, execution
- safe-fail/complexity (impact based)
- narrative (anecdotes)/fragments/blogs (just in time)
- shared context
- tolerated failures
- informal networks/social computing (blogs, wiki, tagging, social networks)
- transparency/distributed cognition

The new thinking is does knowledge actually have to be managed, perhaps it being more on the flow and connection between people.
In fact Dave mentions that knowledge management may be in its decline, to a more naturalistic approach of sense-making “how do I make sense of the world and act in it”.

This makes sense (pardon the pun), in the web 2.0 world we often call what we do “social computing”, we are sharing thoughts and forming networks and communities. There is no effort involved, it’s totally natural, yet we don’t think of it as managing knowledge, instead we constantly share our thoughts/experiences/opinions, not just because it’s related to a deliverable, but because it’s a personal and cultural thing. When we want to know something we search the blogosphere or ask our network (we also monitor our interests by subscribing to blogs and connecting to people).

What I like about blogs in contrast to just documents, is that blogs can act as a quality filter for gems in the Document Management System (DMS). If you are wondering if a document on a topic exists are you going to go solo and search the DMS, or search the enterprise blogs to see if anyone has posted and pointed to such a document.
Your search in the DMS may result in 100 documents, your search in blogs may bring up 2 hits pointing to documents; the fact that someone is pointing to these posts may say something good or bad about the quality, it’s the fact they were worthy (good or bad) of a mention.
This approach is taking the first choice of dipping into the social filter before going it alone, and that’s not all, you not only get a pointer to a quality document, but you get extra perspective around the document.

The killer message of this post is “…creating a knowledge sharing culture is a misconception”.

But first, here are Dave’s 3 heuristics of KM, taken from here:

“1. Knowledge can only be volunteered, it can never be conscripted
- you can’t make someone use or share their knowledge if they don’t want to

2. We only know what we know when we need to know it
- knowledge is highly contextual and is recalled in context, not in abstract

3. We always know more than we can tell and we will always tell more than we can write down
- some things can only be known through experience, narrative forms of knowledge dominate, explicit …”

Focusing on knowledge sharing, these other points are mentioned:

1. If people need knowledge in the “context” of need it will always be shared
- people will share in the context of your immediate need

2. People don’t share knowledge in the anticipation that you need it
- if you ask people (perhaps someone you may not know) to put it in a common data store for a possible need in the future, on the basis you might need it…it just doesn’t happen.

So how are blogs different?

Firstly, you are not asking people to share it, as with these new social tools and attitude there hopes to be a naturalistic approach where the culture shares what’s happening in blogs as regular as they use email. In the open blogosphere I don’t consciously blog to share, I just blog as a natural extension of keeping notes, the great thing is people see these notes and discuss, etc…
Then as you go on, you find that you may indeed blog to share, poke, invoke discussion, as you have your soapbox, where you make sense of the world, people visit and help you figure out the world.

What’s lacking in codification

More from Dave:

Building systems built on idealistic models of sharing are fundamentally flawed

The tacit-explicit model fails to understand the nature of the human act of knowing.
The essence is lots when trying to codify tacit information
- too much effort in trying to express what you know
- shared context is assumed, but instead it is context heavy
- low abstraction with knowledge holder, not in tune with their thinking as you don’t have much in common (language, education, morals, values, beliefs, experiences)
- there is more accurate expression at time of knowledge compared to more abstracted hindsight (this is where field work/mentoring/shadowing/narrative/apprentice/social computing helps)
- in the end it becomes difficult to decipher without the knowledge holder

Social Computing and Narrative are the new “knowledge sharing”

From the above, social computing could be seen as the direct alternative to codification as they are both written down. But social computing enables and manifests a behaviour to share information as it happens, in a very quick and simple, and unstructured way…plus these items and people are all connected in informal networks.

What I like is that anecdotes are raw data, others can read/listen and draw their own conclusions, everyone may get something different (most often knowledge diffusion will occur). This is like the beauty of poetry and myths, anecdotes are not as subject to time (like codification), they last the ages (forever applicable), they can be useful to other disciplines…on second thought perhaps myths are more storytelling.

Trust/Informal networks

Dave continues…

Reality is; sharing is based on trust
Trust is nurtured and grown overtime, it can’t be designed
Trust exists in small network groups, not at an organisational level

If you want to manage knowledge, you need to manage small network interactions…you don’t manage stuff at a global organisational level.

A social computing environment enables you to network a circle of high abstraction in turn creating exchanges of shared context, this is knowledge flow…Dave says that web 2.0 “…makes the context in which you receive and filter information and knowledge more critical, partly because you can control that context by exchanging with people you already know and trust”

My post on informal networks hones in on exactly these points…mostly a shared context needs to exist for a successful transfer of information, and a circle of high abstraction is the ideal environment.

Web 2.0 achieves all these things without trying to be knowledge management:
- self assemble
- self organising
- RSS Reader and social network (filtered trust)

Dave goes on to say that Web 2.0 makes context in which you receive and filter information more critical:
- as you control the context, as it’s coming from people you know and trust
- it moves knowledge outside the corporate control
- rewards neglect the trust factor…need to share for the right reason

Informal networks are more successful over formal networks, as you are free to choose who you read and discuss (trust)…are we moving from CoPs to informal social networks.

So again, you can’t create a knowledge sharing culture:
- you can increase the interactions between people
- you can increase their interdependency
- you can increase immediacy of their KM request

This is largely a just-in-time concept that is able to execute based on knowledge flow (in essence blogging) and networks (social network connections and subscriptions-people filter).

Dave’s future of knowledge sharing is:
- it’s not something subject to organizational objectives
- the role of the KM dept will be absorbed everywhere
- the role of IT will be to create connectivity between people
(connect people and km will look after itself)

In essence you are creating the ecology of knowledge flow, the shift from object to context.

CONCERNS

Even though this is a more naturalistic approach, will all people blog and network?

I look at it this way, we already have email contacts, a network is making these contacts more dynamic into a connected system…it’s a no brainer. Those not taking part will feel disconnected and will therefore hookup.

Your productivity dashboard will have an RSS reader and a blog, these will be features of the network system…very non-threatening as a feature compared to scaring people with a brand new standalone blogging product.

Once in a knowledge flow networked system, you can explore, discover and capitalize on the social capital to get things done, especially your weak ties.
As I said in a previous post, I really don’t know what Outlook is waiting for…if blogging and networking were part of the familiar territory that is Outlook, these tools would be less threatening as they would be considered normal.

Busyness

Andrew McAfee mentions that new social tools may be seen as unproductive, and people may be too busy anyway.
Tom Mandel mentions people have always used new technology and social tools to get things done, eg. phone, fax, PC,email, IM, etc…and Luis Suarez chimes in that maybe we should move to group performance.
Joe McKendrick eloquently states the paradox, alluding to, are you going to risk the opportunity to make more widgets over spending time collaborating and networking which may result in making better and more widgets via a new innovation…the answer was in the air waiting for the social captial to synergise and manifest it.
Anne Zelenka brings home the comparison between busyness and burstyness workers types.
All of this talk leads to ROI, but that’s for another time.

The next 2 points are from a Kathleen Gilroy podcast interview with Andrew McAfee.

Transparency

Managers may only want contributions that are appropriate to their level on the Org chart.
They may not want someone lower to have input at the same level, or at the worst refine or overrule contributions…this is a decentralised decision environment.
For social computing (networking) and sense-making to take place in the enterprise, this transparency factor is a huge obstacle to overcome or swallow. Tom Davenport doesn’t think KM 2.0 alone will transform culture.
Andrew McAfee tends to agree that org charts will not be thrown out, the main benefit will be idea percolation, crowd sourcing, etc…this is basically a result of having bottom up knowledge sharing tools.
Tom Mandel adds that enterprise work (changes in the way we work ) doesn’t equal organization of enterprises (changes in org structure)…time will tell. Luis Suarez isn’t budging on the matter.

In comparison to the enterprise, web 2.0 and the blogosphere is an egalitarian environment, there is no org chart, even if there was, no one cares, all people are treated equal.
That’s not to say that cliques and A-lists will form, where the popular people only link to each other or an A-lister will get credit for breaking news that was already published earlier by a lesser know blogger…but this is all human nature.

Another danger for management is a revolt, see the Digg revolt.

Network Effects

On the open web there is room for the long tail as there are enough people to make it scale, but in the enterprise the long tail is too small (there’s not enough people for there to be a long tail).
We see network effects as the aggregated value from all the individual contributions, plus the distributed discussion propagates this as well, then we can look into emerging patterns, this is the beauty of free form personal publishing, it has a greater value.

Again we come to knowledge sharing culture, people need to contribute, not just consume, otherwise we will not get the network effect. If we don’t have a fuller participative enterprise, then the social content will not manifest into great things.
In the enterprise if we have only a 1% participation rate from 10,000 people that’s only a 100 people blogging, will this generate a network effect, it may for a topic, but not the system as a whole.
In contrast on the web a 1% participation rate may be millions of people, enough scale for network effects to happen.

So it comes back to visibility and coaching, and a naturalistic approach.

Knowledge sharing?

Do we call it knowledge sharing, maybe not, that’s the result amongst other things, but what we are actually explicitly doing is not sharing, we are publishing fragments, it just turns out its being shared at the same time.
This perspective takes away the task or conscription aspect of having to do something for the greater good, and turns it into personal publishing for your benefit that everyone can also see. I guess this is the del.icio.us lesson, besides ease of use, personal benefit must be the primary reason you are bookmarking or blogging something, public benefit is just a bonus, and what a bonus it is, we get discussion and emergent patterns.

We publish and collect, people subscribe to us as their trust filter and the result is sharing and discussion. But we may also explicitly and directly share by sending someone links or asking a question. Because of blogs and networks we may know who to ask the question to as we are in a connected environment where we are aware of what people know, or our circle of trust can point us to someone.
It reminds me of the web, but instead people are the hyperlinks, pointing others to another person (page).

Browsing some blogs via my social filter I came across a post by James Robertson (who I subscribe to anyway) titled, Knowledge sharing’ should be avoided. James points out that the jargon and the act of knowledge sharing just doesn’t gel with people, it doesn’t rock their boat, they are not enthusiastic.
I think this is due to there not being a direct benefit, we only do things-most of the time-where our time spent is towards achieving something (for ourselves or not).
Six months after codifying and knowledge sharing an individual looks back and thinks, why am I doing this, is it being used, I’ve gone to the trouble to write these items for any type of audience that encounters them.
James says we should avoid the term altogether and put it into a more focused actions like ‘reduce processing errors’.
Knowledge sharing needs to happen as a result of doing something you are trying to achieve, not just for the sake of knowledge sharing. When you blog, you publish news or updates, successes, difficulties on your project, you point to other places…you do all this for the here and now, and it means something to you, it helps you through your day, your thinking out loud…

Sense-making is where it comes full circle (best practices or pattern matching fragments)

More from Dave…

Humans act on fragmented patterns, not information processing.
When faced with a hard issue or task, is our instinct to go for best practices or round up several people with experience that you trust?
We don’t go to best practices, we go for fragmented anecdotes
- they are context independent, more resilient and flexible than codification

eg. when faced by a lion, we don’t remember best practices, we pattern match narrative fragments and blend them into our decision (and escape hopefully)…this is sense-making.

Then you blend these fragments with your own context to create new patterns and ways of thinking, ultimately creating decision pathway forward.
Whereas codified knowledge may not transfer to someone else as it is too context specific, structured and more relevant to the time…instead you have more a shared context with people you trust and know (as you have a higher level of abstraction with your immediate circle). Without shared context and common abstraction (eg. education, experiences, background, etc…), knowledge to someone else may be data to you…without the shared interpretive framework you can’t successfully internalize it as knowledge.

The example was given of being able to relate more to raw war stories and anecdotes over a codified document on the war.
This is key for me, even though I agree there is room for codified documents of a particular nature, when it comes to sharing personal knowledge and thoughts it’s not as effective in diffusion.

If I read a codified document (context heavy) I will find it hard to identify in order to get something out of it for my purpose or context. If these were anecdotes, they are more open and less decided or contextual, this gives more chance to trigger something for me, it gives my mind space to “jam” with what I’m reading, “riff” on it to spur other ideas.

Sharing anecdotes and fragments by blogging and sharing sessions are just that, you are not trying to codify, you are just sharing a moment, the essence is kept intact. The benefit of this is that an anecdote can be internalised by different people different ways, whereas a codified document is meant to be internalised or transferred one way and that’s it, if successful it becomes knowledge to you.
Whereas a blog post or anecdote (fragment) can spur innovation, they are like images (they are vivid), they are not trying to convey one narrow point, 10 people can get 10 different feelings or internal knowledge from an anecdote compared to a codified document which is trying to convey a fixed signal.

eg. sharing anecdotes of safety failures will have lots more impact and a lasting effect than just relying on a codified safety manual.

Also when we don’t understand something, an example or situation always makes the point clearer, an anecdote is even better, especially in recall…just recall the anecdote and then derive the concepts in the anecdote (rather than trying to remember concepts).

Dave gives a great example of a map (codified), a map may plot points, but someone who knows the area will know the quickest way to get somewhere, areas to avoid, etc. There is no way this dynamic tacit stuff can be codified into a map, this shared context only comes from people or fragments…the map assumes shared context which is not present.

Speed or Quality

The premise of this post is that codification (tacit to explicit) is a much more cumbersome technique in sharing what’s inside your head and being able to transfer shared context. Instead it’s very context heavy and has less profound recall of actual event, with the notion of knowledge as things. Whereas networks and unstructured content (freeform blogs) are simple and fast, story/fragments (more context independent), and based on the notion of knowledge flow.

Added to this is the dynamic nature of these items (comments/trackbacks), and a profile based expert system where people can connect to each other, it’s very emergent and self assembled.

As mentioned the premise isn’t to share knowledge, it’s just a by product of doing work eg. personal publishing, discussion and networking in an open system.

Just to mix it up a bit more, a recent paper by Knowledge@Wharton called Does Knowledge Sharing Deliver on Its promises says the choice between accessing people or codified documents depends on the task, and whether you want speed or quality.

From the paper:
“We find that using codified knowledge in the form of electronic documents saved time during the task, but did not improve work quality or signal competence to clients, whereas in contrast, sharing personal advice improved work quality and signaled competence, but did not save time,” Haas says. “This is interesting because managers often believe that capturing and sharing knowledge via document databases can substitute for getting personal advice, and that sharing advice through personal networks can save time. But our findings dispute the claim that different types of knowledge are substitutes for each other. Instead, we show that appropriately matching the type of knowledge used to the requirements of the task at hand — quality, signaling or speed — is critical if a firm’s knowledge capabilities are to translate into improved performance of its projects.”

See more from Mike Gotta, James Dellow, and Dave Snowden.

Dave points out that it compares codified documents to network advice, what about fragmented anecdotes and blogs compared to asking people.
The other point is that observation always produces different results to surveys (subjects understand the hypothesis).

More notes

This post is focusing on the knowledge sharing aspect of KM, but these podcasts went into much more, here are the rest of my notes:

Most effective knowledge exists in flows…contextually created in the time of need

And this beauty:
“Knowledge is the real-time assembly of multiple fragmented memories in a real-time context to create a new unique application”

You can’t knowledge audit fragments, rather than create an audit of what you know, create a map of dependencies of your core business processes or knowledge objects.
Then create a small portfolio of projects of
- single process highly dependent on many objects or vice versa
A knowledge object is anything coherent that we can see how we can manage

eg. gather anecdotes…index them based on types of decisions present in the anecdotes
- allows to create decision clusters (anecdotes to create context so people can understand what happens in a decision)
- then go to people who make those decisions and ask them what was involved when they made this decision (based on the ASCHEN model).
- what artifacts are in use, skills, rules of thumb, experiences people need to make a decision in this space, natural talent (what makes some people better at this)
- re-cluster this to knowledge objects
This is creating coherent clusters of things that can be managed…it’s designing structure on what happens and how things are done…it’s a bottom-up knowledge objects and top-down processes.

People will tag things more conceptually, where the terms may not even appear in the content.

Anecdotes and self indexing by the people gives us a tag cloud where we can click on a term (decision based word) and read anecdotes that used those types of decisions. The overall tag cloud (storyscape) gives a visual view of frequency of anecdotes and decisions, revealing patterns and weak signals.
Whereas best practices try to tell you what to do rather than put together the pieces yourself (which is also more empowering).

Paradigm change
- semi structure tagging
- fragmented blogosphere (designed like our brains work)

Relationship model…social computing is more cow and grass, rather than cow and chicken if presented with all 3.

Use catalysts to create attractor mechanisms
- if patterns form it becomes an attractor that sucks in activity
- reinforce good ones, and disrupt bad ones

Manage emerging coherence with attractors within boundaries

Not just telling people what to do, just creating an environment to see what happens…safe-fail not fail-safe

Fail-Safe
- good for process based system where you know user requirements in advance of building the system
- in KM you can’t know requirements until people use the system
- people don’t know system they want till they get it, then they want something else…as the capability of these systems exceed our capacity to imagine its use till we have used it
- need to switch from Fail-Safe Design to Safe-Fail experimentation (self organization)
- social computing allows experimentation as it’s so cheap or free

Traditional Management
- study successful CoPs
- then try and replicate outcome into other situations
ie identify best practice by study of current practice, produce a recipe for a desired outcome, then replicate successful outcomes in other environments…this is a copy culture (case study)
- you can’t replicate outcomes, but you can replicate starting conditions
ie you produce a contextually appropriate solution

Further to this people learn more effectively from failures, they have more impact.

Assumption of traditional Management Science
- if we see things we will make rational decisions based on available data.
- but humans aren’t information processors, we are pattern matches
eg. Gorilla present in a video clip is not seen if you are concentrating on processing the task ,meaning you have not scanned the environment properly as we are not natural or exhaustive information processors (unless you are autistic)
- the most you can scan of available information is 5 - 10%

KM assumption
- world is a series of dots and you join the dots from previous experiences
- make a decision based on patterns in long term memory and 1st pattern (not best fit) is applied

What is this all about anyway?
- Improve decision making (How do we get people to make better decision)
- Improve conditions for innovation

Knowledge Economy
- it’s not what you own, it’s speed of exploitation (how you use what you have access to)
- the more you share the more you get
- a philosophy of abundance not scarcity

So how do we remember all the information in these podcasts?

If you remember the anecdotes and examples, the concepts will then follow…wish I used anecdotes to remember concepts in exams when I was a youngster.

It’s quite clever that Dave Snowden is using the very concepts that he is teaching in the teaching process/delivery…there must be a word for that.

Here are some of the anecdote titles and concepts:

AUTISTIC - processing vs patterns
ROUNDABOUT - rigid structure vs complexity (attractors/boundaries)
GORILLA - processing based management vs patterns
HELICOPTER/OIL RIG - field context (shared context)
MAP/OPERA/TRAIN/TAXI - codification (assumed shared context)
CHILDREN’S PARTY - outcome based management vs complexity, emergent (safe-fail)
CHICKEN/GRASS/COW - relationships vs categories
CHEF/RECIPE - outcome based and case study management, best practices vs complexity, safe-fail, blending patterns (sense-making)
LION - best practices vs blending patterns (sense-making)

[ADDED 29/11/07: Check out Luis Suarez’s series of posts (1, 2, 3, 4, 5, 6) based on one of the podcasts]

[ADDED 7/12/07: More on the new knowledge diffusion]