KM guru Dave Snowden was one of the highlights of David Gurteen's Making Knowledge Work conference last week in London, at which I also spoke (about creating a social interface to corporate information - warning: 6Mb PDF with pretty pictures).

Entering the room like a Knowledge Management rock star - bearded, late and with the slightest hint of condescension towards his audience - he proceeded to pull three (count them ... three!) flip charts together in the centre of the room, whilst knocking over the projection screen, in what I took to be a subtle reference to Keith Moon. Rock and roll!

"KM is dead .. We know this to be true because government is now its biggest customer"

"The concept in management science that best practice can be studied and replicated is WRONG ... It ignores non-causal systems .... [and] confuses correlation and causation."

By now I am swaying in the aisles, lighter aloft, mouthing the words to the children's party story that illustrates how we can influence but not micro-manage complex systems.

He went on to berate the Anglo-Saxon obsession (he is of course extremely Welsh) with taxonomy and classification at the cost of perspectives and relationships, illustrating the point with the Cow/Chicken/Grass test (the odd one out is not grass, but the chicken, for cows eat grass), skillfully glossing over an agricultural challenge from a fellow Celt based on the fact that chicken sometimes also eat grass. Never mind, the chicken, where are we headed?

KM is dead - what comes next?

According to Dave, KM is heading in three directions:

1. Techno-fetishism: where organisations focus on codification through technology solutions, which is little more than an advanced form of information management
2. HR solutions: where it becomes a servant of recruitment, retention and succession policy, owned by HR and run by IT
3. Sense making: where the focus is not on sharing knowledge but on enabling better decision making, creating the conditions for innovation and understanding the way we make sense of our world

Sense making is where he thinks we should be going. But do all technologies fall into the codification trap of techno-festishism? I don't think so. Can there be meaningful technological underpinnings for sense making? Social software is one set of tools and approaches to augmenting human processes and supporting the way we make sense of the world that has achieved some success in doing this. We'll come back to that.

How do we make sense?

"The only humans who analyse all the data and then make a rational choice are autistic, but economists insist this is the way we all work."

Instead, as Dave Snowden reminds us, we filter incoming data (typically processing about 10% of available inputs) and perform intuitive pattern matching to make sense of a situation. We also do this to make decisions, using the best first fit principle rather than analysing all data to find the optimum solution. Think of how pilots make decisions, for example, or how we drive a car. We cannot hope to process all relevant incoming information. Instead, we sense it based on patterns.

Pattern entrainment is a danger when we start to apply the same patterns to every situation, but individuals are able to use identity (paradigm) shifts to apply different patterns to a given situation. For organisations, this can be more difficult, and they tend to suffer pattern entrainment from time to time - becoming locked into a specific way of thinking or doing things - until that pattern results in catastrophic failure, opening the way for radical change.

Part of what the Cynefin Institute does is to use dynamics and social network stimulation to shake things up within organisations to avoid the problems that entrainment can bring. This is just part of its wider approach to complex systems that I heard about at Dave's KM Europe 2003 talk, and which (I notice) Matt Mower wrote up more diligently than I did at a talk in 2004. The Cynefin approach, which is refreshingly intelligent and more practical than it often appears, is presented in more depth in a series of papers they are making available as 'open source consulting' via their website under a Creative Commons license.

Feeding our peripheral vision

Is it true that we apply groups of patterns, organised in domains of identity, and use them to make sense of our world by processing (but not necessarily analysing) a large number of sensory inputs? Well, this model certainly reflects the way I read a newspaper or (I admit it!) some books - newspapers follow patterns that mean I can avoid processing the majority of words whilst consuming most relevant information. The physical layout provides contextual clues that RSS feeds cannot match, even if they are a better distribution mechanism. Thinking about it, we also use peripheral vision and pattern matching in many other ways, such as when we are assessing our position in a room full of people or just walking through Brixton. In each case, our attention is spread thinly across a wide array of inputs and we will zoom in and out when we need to focus on something in particular.

What does this mean for Computer Mediated Communication (CMC) in general and social software tools specifically? I think we can learn two important lessons here.

First, seeking to provide a linear sequence of compulsory process-driven information ("the right information at the right time" in old school KM lingo) can actually hinder sense making among knowledge workers, by treating them as information processing appliances. Instead, techniques such as social tagging, feed aggregation, presence sharing, etc. can help package up small pieces of potentially relevant information into feeds and trails that we can process with peripheral vision. The new Web 2.0 tools and services help create an ecosystem of connected people and information - as David Weinberger put it this week: The cure to information overload is more information, only it should become more ambient.

We need to let people organise their inputs by exposing all relevant information in granular feed form and then provide smart aggregation and tagging tools to create a personal eco-system of content, cues and links.

This is what we have been describing as a social interface to corporate information sources: create a layer of feeds and flows that reference content objects, and allow people to apply flexible personal meta-data within a social context to constantly reorganise the links into that content according to their day-to-day needs. This is what my own presentation to Making Knowledge Work was all about (warning: 6Mb PDF with pretty pictures).

Second, we should help people develop the skills and confidence to move from linear processing mode, where they feel a need to respond to our acknowledge everything (e.g. memos and the email inbox) to peripheral vision mode, where people make better decisions and connections by assisted by ambient information feeds, and where information grabs our attention only when it needs to (e.g. "reading" in an RSS aggregator, sensing importance of links through number of references or recognised trust relationships). This takes the courage because it applies the 80/20 rule to investing attention - you might miss something from time to time, but if it is important it will probably re-appear on your radar.

If the objective of KM is to create the conditions for innovation and support better decision making, then we have some useful tools at our disposal to help contruct a personal information eco-system, and these are opening up new ways of working in a variety of areas:

• Weblogs for collective sense making, narrative flow and connected conversations
• Social tagging for emergent meta-data and serendipity
• Wiki for co-working and shared spaces
• Presence and perspective sharing to know who is where and what they are doing
• Personal and group aggregators to create our ambient inputs

In conclusion, I think there is certainly a role for social software in augmenting human interaction and co-working without it being techno-fetishist, and I believe it has genuine potential in support of personal and collective sense making. People growing up with social software ('the kids') may not know much about KM or even IBM, but they are making knowledge work for them in some new and interesting ways. This experimentation is producing new techniques and innovations that, in many cases, provides useful learning for organisations who are trying to make their underpinning knowledge sharing systems more effective and more applicable to supporting humans rather than just business processes.