- David Snowden — IBM
- Yasmin Merali — Warwick University
The survival of a business organisation depends on its ‘intelligence’ - its ability to call on the intellectual capital of its members or to capture new knowledge from outside itself. ‘Knowledge space’ refers to the store of possible information that may be generated within some cognitive system, constrained only by the language or culture of that system. Organisations have captive knowledge which may be ‘explicit’ in that it is written down or stored in some data base or ‘tacit’ (implied) in that it resides in people’s heads or in their practical skills or is unexpressed in some other way.
Business organisations are complex to the extent that their success depends on the dynamic interaction of teams or individual members which in turn depends crucially on communication or information transfer within the organisation. Information flows very fast and very freely on computer networks and the self organising clusters of individuals are a result of this information transfer. Yasmin Merali assessed the possible use of biological examples in explaining the dynamic nature of communication networks in virtual communities - i.e. those in contact via the computer network. David Snowden exposed some of the myths of so called ‘knowledge management’ techniques, introduced us to ‘action based research’ and offered a new model from the IBM store.
Seminar Notes On below Compiled For The L.S.E. by Geoffrey J.C. Higgs 19/3/00
Table of Contents
- Metaphors, analogies and models
- Birds, bees, ants and robots
- Information transfer in human organisations
- ‘What counts as knowledge?’
- Abstraction and self organisation
- The seven sins of ‘knowledge management’
- ‘Action based research’
- Four ‘realms’ by which a business can be classified according to it knowledge ‘management’
- IBM’s Model
Metaphors, analogies and models
The use of examples from biology is a well trodden path to understanding the behaviour of complex systems. The natural world is after all the global ecosystem in which we and all other organisms live out our lives and contains species which are either in competition or collaboration or symbiotic in their relationships. So it is natural for us to look for similarities between the way animal communities live and our own social organisations. We can use biological system as metaphors or analogies. If the metaphor seems to be working in pointing up similar features we test it as a metaphor for explaining the possible relationships that exist. Taken farther we may even build computer models which simulate to some degree actual behaviour. But there is a limit. True complex systems are evolving systems and as such are neither predictable nor can be expected to repeat their history in any exact way.
Birds, bees, ants and robots
Many animal social organisations display the characteristics of complex systems in which the dynamic interaction of individuals both with each other and with the environment give rise to emergent and often stable patterns on a macro level. Human organisations consist of highly autonomous individuals, but we can gain some insight into why they have the characteristics they do by looking at the behaviour of animals. Why, for example, do birds fly in particular formation, or bees swarm or ant colonies show a particular division of labour? In an ant colony there may be ‘stingers’ and ‘foragers’. When there is numerous prey there are more stingers than foragers. When the prey is thin on the ground there are more foragers. Stingers and foragers have different ‘rules’ of behaviour
but the number of stingers and foragers in the community varies according to its needs. Interactions between individuals lead to observed patterns of activity. ‘Rule’ one for a foraging ant might be ‘find food and take it to the nest but leave a chemical trail’. ‘Rule’ two might be ‘if you’re wandering about and you come across a trail then follow it’. ‘Rule three might be ‘if you’re not carrying food and you haven’t come across a trail then wander randomly until you find one’. Put these rules together for an ant colony and you get a particular kind of emergent behaviour. Models based on it offer a way to solve what is known as the ‘travelling salesman problem’ which involves finding the shortest route to service a number of randomly scattered clients. If robot ants are programmed to behave as described above the shortest chemical trail routes to food will be used by more and more ants. This is because those ants which use the shortest routes will get back to the nest and get out again sooner than the others. The observed feature is that there is a greater density of ants on the shortest routes to food sources.
The immune system is a useful example in understanding how organisations adapt to the environment. Antigens can be taken as the environmental condition that leukocytes react to by producing antibodies. Suppose we have autonomous robots that are programmed to deal with objects in the environment in certain ways. Each individual robot ‘experiences’ its own particular portion of the environment but in doing so at the same time feeds back information to other robots. This activity builds up a collective history and a collective ‘strategy’ which is constantly being modified depending on how appropriate it is to the strategy of each individual robot. In working from such examples we conclude that local responsiveness to environmental stimuli is important in gaining a collective strategy for dealing with change. Low level coordinated individual responses give emergent patterns to the whole and simple rules lead to complex behaviour. Such a system gains a ‘robustness’ which is not dependent on any single individual. If a particular individual robot becomes defunct another can always take its place. Valuable insights can be gained by building such models, though there is at present no mathematical theory that is sophisticated enough to describe them, let alone the kind of patterns that might emerge in social organisations in which there is a high degree of autonomy such as those of the human kind.
Information transfer in human organisations
Human organisations consist of individuals or groups of individuals that communicate with each other. There is no single identifiable network but a network of networks. Individuals have a sense of identity within the organisation depending on how it is organised. There is also a corporate identity which is perpetuated but it is difficult to say where the boundary lies - what or whom is in or out. The information that is passed between individuals or groups of individuals may be codified in many different ways and at many different levels of abstraction. Some networks are far more structured than others. Information that is relevant to the organisation’s operation may also be stored in many different ways.
People, highly autonomous as they are, are nevertheless constrained by the rules and culture of the organisation enabling it to be described and classified. Economists, for example, make assumptions in defining what they see as an economic community. They have to explain why communities grow up because of trade. They have to explain how and why businesses cluster or swarm together in certain locations, why they copy and collaborate with each other and how new businesses start up.
For example, new operating systems in computer technology lead to new software development which is very rapid. Suppose we have two technologies A and B, of apparently similar but as yet untried merit. What would be the reason that people would flock to A rather than B? They are not always obvious. If one organisation chooses technology A then another may hesitate to choose B because if it turns out to be inferior A will have a head start. Since success depends ‘being in the swim’ there is also a fear of being isolated so the overall result is that there is a ‘gold rush effect’. Positive feedback leads to innovation swarming. New discoveries and information come together to have dramatic effects.
Looked at from a complexity theory viewpoint we can say that there are certain ‘attractors’ operating, which means that if we were to plot the trends on graphs we would see a certain kind of convergence. The urban concentration in Silicon Valley occurred largely as a result of information ‘spillover’. Companies which had good competencies and good resources clustered because of collaboration and they were in turn exploited by companies having less resources and competencies and so on. The process becomes explainable or tractable if we make certain assumptions such as the ‘agents’ that bring the process about are ‘rational’. That’s how we do economics. We assume that agents are always motivated in a certain way and use that assumption to start defining the system. As we become increasingly sensitive to its complexity we have to pick out more and more attributes with which to define it and there is increasing difficulty as we attempt to give greater and greater autonomy to our agents.
In the case of human social systems, the problem of trying to be an ‘independent observer’ becomes even more acute and because we are in the systems ourselves we find the attributes we chose to define the systems are much harder to agree on. We happily describe ants as being ‘programmed’ to do things, or we say birds navigate by an ‘instinctive’ awareness of the position of the sun or the configuration of the stars but we don’t say that about humans so we have to talk in terms of information. But information involves the notion of meaning and with meaning the whole legitimisation of what constitutes knowledge becomes very subjective.
‘What counts as knowledge?’
If the problems are large in defining human organisations in a behaviouristic way then we have to go to language and culture in order to make sense of them. ‘Information space’, or the realm of legitimate knowledge inside the language and the culture is ‘multidimensional’ - which means it can be cut up in many different ways. The problem of attribution is a social one and at the microlevel comes down to value judgments. It can be looked at the macrolevel in terms of the collective sympathy or resistance which determines whether or not an idea is taken up. Knowledge management is therefore not really about ‘managing’ the knowledge or skills which people have within an organisation but in ‘making sense’ of why something happened and trying to influence the future by such intelligence. It’s no good expecting a metaphor to provide all the answers. Human social systems are unlike any other animal system and largely indiscernible because we live in them. ‘Biological evolution’, for example, may explain our evolution on the crude level of genetic mutation subject to environmental pressures but ‘fitness’ in anthropological terms is a very different matter. In the natural world we believe species get selected for their ability to deal with the changing environment. They persist if they have the requisite variety in their gene pool. Human organisation on the other hand may get selected on value judgments. We often go for the best regardless of diversity considerations. So we’re saying that human behaviour patterns are grounded in intentions. Looked at this way we can forge a connection between say, some environmental practice and why it is imitated. It is imitated because it is good. But if we try to unpack the idea of good environmental practice then we have a very complex situation indeed.
Cultural evolution proceeds according to how we change our minds and knowledge creation involves the kind of ‘scanning’, ‘codification’, ‘diffusion’ and ‘absorption’ that Max Boisot has worked on. It is language and culture that condition the kind of behaviour that takes place in human organisations and it is our language and culture which determine which patterns we pick. But in information transfer there is also the problem of ‘context dependency’. Observations of environmental factors are always about particular situations which for knowledge sake have to be abstracted and generalised. Too much attention to the particular situation and generalisations get lost. Too little description of the context and you inhibit information growth. So there’s the rub. But the proof of the pudding is in the eating; if the cognitive system is inadequate so will be its strategy and the organisation will soon feel uncomfortable. If we were thinking in scientific terms then we might say it was time to shift the paradigm.
But paradigm shifts are not easily accomplished from inside an organisation. The symptoms of an inadequate strategy, such as diminishing profits and disappearing shareholders are easy to spot but the cure is more difficult. Restructuring an organisation is not just an exercise in logistics. It involves changing the ‘mindset’ and to do this competencies and capabilities from outside the organisation may have to be brought in.
It is the information space that has to be explored and understanding the dynamic and relationships involved is important. Networks of relationships have to be forged so that
new order is achieved and a ‘phase transition’ is effected.
Within an organisation there is always a set of core competencies and capabilities and though it is often necessary to bring in outside expertise to facilitate restructuring everybody in the organisation is a knowledge resource. It is the members of the organisation that have to manage the parameters of what they are and what the organisation is able to do. Co-operation is based on long term payback and people will move and switch if it is unfavourable. Thus the idea of a collective identity is a key motivator. The assumption that everybody is motivated by money is crass but there has to be something which is analogous to the reason that ants forage or birds flock. Autonomy implies choice. People who volunteer knowledge need to be able to choose whether to do it or not, so at the same time it is essential to nourish a sense of commitment.
Abstraction and self organisation
Ants, birds and bees are supposed to be motivated by instinct - whatever that means. They are perhaps unaware of what the whole system looks like. People use ideas and language to understand what they are and how they fit into an organisation by a process of ‘abstraction’. It is this which creates the ‘observer’ viewpoint and enables them to see the identity and coherence within a structure. But this process is a social one of learning. Some animals learn collectively more than others. Birds such as tits co-operate to get the cream out of milk bottles whilst robins which are more individually territorial don’t. Chimpanzees learn to crack nuts or fish for ants within the social group. Those outside it don’t learn the trick. So social interaction leads to co-evolution and engenders the idea of expectancy or trust. But from customs and habits come rules and sanctions. It’s surprising how well the biological metaphor fits this. Tribal communities in Somaliland live in a dynamic and dangerous situation. In the atmosphere of intertribal warfare these clans are tightly self regulating and self organising and they are very robust. Violate the institutional rules and you get severely punished or worst of all, thrown out. Cyber communities might be said to operate in much the same way.
Identity is the way people see themselves as a member of the organisation and the value judgments made about it will condition the ways in which they will defend it. But there’s a trade off between a stable community and getting trapped in a ‘mindset’ where the dominant logic results in tunnel vision. We need to have the flexibility of strategy to cope with a changing environment. We need to have the flocking and clustering of new knowledge groups that spring from the autonomy of agents and we need to avoid the normative ‘communities of practice’ that like the medieval craft guilds were doomed to extinction when the economics of the society changed.
Learning in an organisation is not a process of introversion, it is combining the new in a creative way with the old. Things which challenge may be due to exploration outside the organisation or be generated from inside. Whilst the ‘topdown’ influence is important in perpetuating what an organisation is and what it does, the ‘bottom up’ influence that springs from the autonomy of agents and contact with the outside environment builds flexible strategy. Allowing agents to associate with people outside the organisation (‘outliers’) who are not doing what the core community does offers opportunity for diversity.
Competing within cyberspace means that we must handle dynamic networks and agents must be multi- skilled and able to change roles. Organisations come together at different times and places, disciplinal boundaries are crossed and it is necessary to create a large enough ‘footprint’ to ensure a stable space within the overall information network. If we have such ‘flocking’ and ‘swarming’ then the question arises as to how we recognise the boundaries of an organisation. How do we decide what attributes we need to define our identity? We need to use ‘abstraction’ in the informational space to move above constraining definitions to choose attributes that reflect the patterns that are emerging. We have to understand the ‘attractors’ and ‘phase spaces’. Information leads to swarming but it is also the cause of shape formation. Agents in the business environment are like organisms on a ‘fitness landscape’ their interaction touches the lives of everybody in the social eco-pace whether it be in terms of collaboration or competition.
The perpetuation of an organisation and its identity implies ‘closure’. How do we reconcile this autopoietic idea with that of the ‘complex adaptive system’ which implies ‘openness’? The perpetuation of an organisation does not involve closure to new meanings though sensitivity to new meanings may alter the nature of an organisation. Social systems are not autopoietic in the way that biological organisms are. The metaphor is inadequate if it implies a closing down of thinking. The attributes by which we define our organisation may be physically based, technologically based or culturally based which is why attribution in terms of what is meaningful and what works is so important..
The seven sins of ‘knowledge management’
Knowledge management as a prescriptive practice is based on seven fallacies:
That knowledge can be managed:
Because of a confusion between the nature of ‘tacit’ (implied) knowledge and that which can be made fully explicit people have come to assume that the former is easily convertible into the latter. This in turn has lead to the belief that knowledge or knowledge development can be managed. This is false on the grounds that we always know more than we can tell and we can always tell more than we can write down. In short there is always a loss of context and content when these processes are carried out. Failure to understand this has led to a huge waste of money invested in so called ‘intellectual capital management’ systems. The mistake is also to assume that the nature of knowledge development is a bit like collecting a number of bits which are then put together to make a whole, rather like building an aeroplane. Such an approach neglects the dynamic nature of knowledge acquisition. It involves the confusion between something which is merely ‘complicated’ and something which is ‘complex’. An aircraft is complicated but we can take it apart and put it together again and it is always the same thing - the whole is always an aggregate of the parts. A complex system, however, changes when we take it apart and changes when we put together again. The belief that knowledge can be managed in the way that we might manage the production line or the quality control of a product is wrong. It is imposing a mechanical system concept on one that is essentially organic in nature.
That organisations can be ‘designed’:
The belief that a human business organisation can be designed in the same way that we might design a piece of machinery is a case of misplaced generalisation. There is a tendency for management consultants to study five, ten or fifteen companies over a period of five, ten, or fifteen weeks or months and produce generalised models of ‘best practice’ that they assume are fully prescriptive and imitating them will result in success. This kind of ‘hypothesis based consultancy’ is fundamentally flawed. Its scientific metaphor is that of Newtonian physics in which ‘cause’ is always assumed to be separate from ‘effect’ and that by gathering enough data on the relationship between the two the effect can be determined by the manipulation of the cause. Again this assumes a mechanical universe.
It is also a process of unwarranted generalisation. A bit like turning up in Calais, seeing someone wearing glasses and concluding that all Frenchmen (or women) do. Thinking about how people work in an organisation purely in terms in terms of factory based production theory leads to the wrong sort of organisational models. We need a new model in which the manager is not the captain or navigator of the ship but rather the designer that sets the parameters within which the ship can operate. A gardener rather than an engineer. We can’t produce drawings of a future organisation and hope to build it. Once we see that an organisation is complex rather than merely complicated we must try to start ‘journeys’ rather than aim for goals to be achieved.
3.The myth of the rational agent:
Most ‘management science’ consultants assume that individuals (or even groups of individuals) behave on a rational basis if fed sufficient data. This imposes the current model of computer based information flow onto interactions involving people. It fails to recognise that the individual is part of a particular social network and does not make each decision on a ‘clean’ basis. The way people behave is not merely built up from their own interactions over time but the ‘stories’ passed down them by past generations. It is stories that largely determine attitude and in some societies the ‘script’ is so structured that it is impossible to break out of it.
This is the belief that everything an agent does in a community or organisation is based on the expectation of a ‘return’. In most economic theory all transactions are in terms of financial value.
Anthropological studies show that this is not the case in many societies (perhaps no society?) and most large business organisations tend to be tribal in nature which involves status and deference. IBM has an overall staff of 380,000 but this really consists of a network of tribes and dependencies or ‘shadow’ networks that support each other.
Then again some businesses are run on feudal lines. Feudalism says that if you control land, then everybody who lives on it and doesn’t have any control of it is your slave. In modern business organisations control of the budget is analogous to control of the
land - anybody who has no say in it has to do whatever they are told.
A belief in Utopia:
In a recent edition of the ‘Harvard Business Review’ was an article stating the ‘five qualities of leadership’, implying that if you didn’t have these qualities you would never be a good leader. Making such value judgments and then prescribing them ‘willy nilly’ is often found in current management consultancy. Underlying the advice is the message: ‘do it this way and everything will be wonderful. The best advice is that it won’t and rather than aim for such perfection we should accept that nothing is ever perfect and be happy that for now something appears to be working.
A belief in ‘best practice’:
This assumes that circumstances will remain the same and sometimes they do, though inattention to how fast the business environment is changing has brought down more companies than even the ‘business process re-engineering’ fashion that has now fallen into disrepute. It is as well to remember that ‘best practice’ is always past practice. Looking back in the business world is often a poor guide to the future.
The organisation is merely a collection of individuals:
We have already looked at this fundamental mistake but it is often manifested in the belief that if you optimise each individuals operation the organisation will be optimised as a consequence. This is an entirely false view of the way people interact and the kind of patterns that emerge from those interactions. Managing people to achieve individual goals is only useful if as agents they are defined as unalterable parts of an operation and that they work utterly independently. In reality working in an organisation involves internal and external relationships, the boundaries of which are sometimes rigid, sometimes loose and sometimes constantly fluctuating.
The aim of exposing these fallacies is not to say that all traditional methods of management science are worthless. If we happen to be sitting on top of the only resource of something in the world and all we have to do is to dig it up and sell it and optimise the process of doing so then fine. But even that kind of captive market might find something else to use. What we have to do is to see what part of our business process is accessible to a mechanical model and what is truly complex. There is always a tendency to believe that the latest theory should replace everything that went before, and people come out of Santa Fe believing we should all embrace complexity and abandon traditional forms of knowledge management and process re-engineering. We should remember that Newtonian theory wasn’t eclipsed by Heisenberg’s ‘Uncertainty Principle’ or the application of ‘Quantum Theory’. We need to develop models in which present knowledge management theory is a special case and that enables business executives to recognise complexity when they come across it
‘Action based research’
Helping an organisation to restructure initially involves ‘action based research’. The process of ‘action based research’ is basically to go in and introduce a new ‘model’ or mode of thinking which disrupts the status quo and then see how well the process can be restructured in terms of community, strategy, culture, communication and knowledge exchange. There are four realms or ‘spaces’ by which a business can be classified. But initially it is vital to draw the ‘boundaries of an organisation in terms of what it is and what it does. A large amount of sense making is filling in boundaries to make them distinctive. If we don’t have ‘solid’ boundaries we end up trying to manage in all four spaces. Many organisations tend to use a single model when they should be looking at divergent ones. There’s a handy metaphor to remember: ‘If you go to the top of a mountain and you find a plateau covered with mist you get lost easily. If you go to the top and its a sharp point you don’t get lost because the boundaries are clear. You can navigate down the ridges and deal with the spaces.
Knowing the boundaries is vital in enabling members of an organisation to make their own decisions and act. The poet Robert Frost tells a story of two farmers repairing the fences between the two lands. The work is hot and tiring and one says to the other, ‘why are we doing this, my pine trees won’t eat your apples’. But the other farmer only says, ‘good fences make good neighbours’. We should think about how we manage children. We don’t design their environment but we do hope to set them on a course with particular values and customs which we think will enable them to survive. We mustn’t draw the boundaries too tight otherwise we will get rebellion and we will lose authority, but if they’re too loose chaos is likely to ensue.
Four ‘realms’ by which a business can be classified according to it knowledge ‘management’
|1 Knowledge flow is sufficient to keep the business process going||2 Knowledge flow enables the process to be optimised.|
|4 Knowledge creation and exchange is chaotic.||3 Knowledge is created and exchanged in a complex way.|
- This state or realm describes an organisation where the current information flow is sufficient for the process. Increase the information flow by new knowledge capture and the business becomes less efficient. You are exploring rather than exploiting. Reduce the information flow and you may be able to exploit the knowledge you have to a greater degree but you reduce the possibility of diversification. Most of the time people have sufficient knowledge for the business to run smoothly but a cyclical phase of disruption should be introduced to avoid ‘entrainment’ and conservative thinking which prevents an advance to new markets.
- This is the realm of ‘hypothesis based’ consultancy. The information flow is sufficient to allow the efficient use of existing knowledge and the promulgation of new knowledge. But the working paradigm is insufficient if the external business environment changes fast (in much the same way that determinist physics is inadequate at subatomic levels). However if you don’t have a high dynamic of change then the pursuit of ‘best practice’ is a legitimate one.
- This is a business process that is complex in the way we have defined it. It is where the interaction of agents within the organisation gives a high information flow in existing and new knowledge but the organisation may still fail in dealing with environmental change because its sense of community, strategy, culture and communication is inadequate. Once a situation is recognised as complex it is possible to make small changes which cost very little money but can improve the situation enormously. A bank in Bangledesh provides an example. There was a massive problem with debt repayment and an investigation was launched into the patterns of repayment including a profile of the kind of person likely to default and the adequacy of controls and checks and so on. More and more questionnaires and employee guidelines were produced but the problem refused to go away. The reason was that the situation was being treated as merely complicated instead of complex. In fact the more rules were introduced into the system the easier it became to default. Eventually someone found the solution. Anybody could have a loan provided four other members of the village took out a loan at the same time and they all guaranteed each others loans. In retrospect the solution seems obvious. A person knows others in his or her village in a way no bank manager can know them. Each person has respect and social obligations in the village and would be ashamed to default. Nobody is going to support another in taking out a loan unless they really believe it will be repaid. Simple intervention produced dramatic results. One of the ways in which a complex process is detected is by reviewing the solution. In retrospect the solution often makes rather obvious sense but we could never have predicted it in advance. Rationalisation in retrospect is like the difference between understanding in a Newtonian type paradigm in which three dimensional space is absolute and understanding in one which is multidimensional and relative. From an information or knowledge standpoint we can say that ‘meaning is always retrospective but never predictable’. Meaning is an emergent property that arises from a certain discourse level in a community. Action based research involves increasing discourse levels and responding to the emerging new meaning.
- If a business finds itself in this realm it is important not to increase discourse levels because that will only increase the turbulence or chaos. David gave the example of a colleague who was asked to deal with an organisation in this situation. He didn’t do any analysis but rather made certain changes (which he didn’t articulate) and noted the impact on such things as cash flow, the order book, and the perception of the shareholders. The changes were simple but they introduced ‘baffles’ which redirected the flow and reduced the turbulence so that the business could be moved into other realms.
The principle of sense making is to understand into which realm a process falls and to what extent. It is more important to know ‘what we do not do’, rather than what ‘we must do’ just in the same way that we learn more from failure than we do from success. David’s work on ‘storytelling’ shows that stories about failure are more valuable than those about success. Communities which rely heavily on narrative to pass wisdom down the generations use archetypal accounts to convey key cultural mores. Some companies have even introduced fictitious archetypal characters so that employees can tell stories about failure without blame being attributed which is very important to a ‘lessons learnt’ program as opposed to a ‘best practice’ one.
Communities or cultures can be managed either from a ritual or a rule basis. If you define the individual in terms of his or her uniqueness then you tend to get a system of rules for the individual. If you define the individual solely in terms of the part he or she plays in the collective then you tend to get a ritual based community. Go to a North Eastern U.S consultancy and you will find the former. Go to a West Coast company and it’s ritual communication based.
There is a word in Welsh which in English is roughly pronounced ‘cunevin’ (spelt ‘Cynefin’). A literal translation is ‘habitat’ though this trivialises the word which actually
means ‘the place to which you belong, your community, history and spiritual home’. The relevance for us here is that everything we do has some connections with our collective history and community based perceptions. Interpretation of the present and action in the future is determined by perception of the past. Management consultancies tend to assume a ‘green field’, something they can design from scratch - which is an engineering approach. But in real human communities you get ‘cynefin’ which implies a certain continuity.
‘Learning’ in the context of what an organisation does is ambiguous between inculcating that which is already known and allowing for individuals to discover new things. Levels of language have a huge impact on learning within an organisation. We have to always assess ‘intellect capital’ in terms of knowledge transfer. In ‘communities of practice’ experts tend to ‘close’ the language because they assume other experts have been through the same initiation into the jargon. But such expert language enables communication at very high levels of abstraction. High levels of abstraction enable density of communication but the language is highly symbolic. In a culture references to past experience, values or beliefs are often unstated. Within a culture 80% of active knowledge transfer can be in the form of symbolic language. This may be at the national language level and still have an important impact on the organisation. Recently the number of people in the world who speak English as a second language exceeded those who speak it as a first language. Speaking English as a first language is now a disadvantage in a large international company because those that do communicate less well because of assumed cultural reference
IBM employees work in different countries and often from home using the computer. There are few rules as far as ‘virtual’ collaboration and to date there are 75,000 of these collaboration clubs and some 55 formal ones. So this ‘shadow community’ is one in which people are self organising and creating vast tracts of knowledge that the company could never hope to manage formally. In fact individuals sometimes do not wish to collaborate formally for fear that their ideas will be taken by those they do not know and who will give them no credit. When forced to do so they will often resort to ‘camouflage’.
Because IBM would neither wish to nor be able to manage all the ‘spaces’ and ‘boundaries’ of such groups they use a process which has been dubbed ‘just in time knowledge management’. The company has created structures which enable it to call up and move information across boundaries on a ‘just in time’ basis. This started more or less accidentally when it was ‘flagged’ that a virtual ‘teamroom’, including David Snowden in the UK were working on ‘story’ - the use of narrative to pass on knowledge or expertise. Once it was put out that a report was in progress and any information would be welcome there was a number of e mails from people interested which became a flood when an article was published in the Harvard Business Review. Although David was about to write what would have been a very sizable report the questions in the discourse enabled him to focus on what was really relevant and at what level of abstraction.
IBM now have a search engine called ‘tacit’ that can trawl the ‘team rooms’ on the intranet and pick up any key words that might give a clue to information on any current problem that they have. An e-mail is then sent requesting help and a task force can be quickly assembled. But privacy is respected in that only key words and not text are picked up. Also whether a person responds is up to them. If for example you are looking for an expert on ‘story’ you may not pick up David Snowden’s but he gets informed that you are looking. If he knows that you are a person that’s likely to steal his ideas he doesn’t respond. If you’re someone that he knows and trusts then he might phone. In a bureaucratic organisation the thief would prosper but in this kind of ‘shadow’ system he or she gets starved of the access to knowledge on which the exploitation depends. Again we have an example of a system that is complex in that individuals cluster and form their own information groups. IBM does not ‘manage’ this process, but by tapping in to it it has access to a huge intellectual capital.
In a rapidly changing business world the ability to tap into such business eco-systems is a key to survival. Using a tool like ‘tacit’ to identify where communities that are useful exist, enables the company to create a structure from the ‘shadow’ world and make it formal if it wishes. We cannot form a community of this size by management fiat, the clustering and swarming that occurs is a natural consequence of the dynamic though we could organise a voluntary weekend, issue a paper, organise discussion or chat lines and see what will swarm. Clustering and swarming lead to sustainable communities which can subsequently be made formal. Survival in business may well come down to how fast you can access your own information space and how fast you can access other people’s.