denotes premium content | May 19 2013 Feature
posted 1 Apr 1998 in Volume 1 Issue 5
Thresholds of Acceptable
Uncertainty
David Snowden , IBM looks at achieving symbiosis between intellectual
assets through mapping and simple models.
The mapping of an organisation’s
intellectual assets at a high level provides an understanding of what they know
and , more importantly, what they need to know. The preceding article in this
series identified a means by which organizations could rapidly create a
knowledge map of their organisation by focusing on the practice of decision
making over a concentrated period of three to eight weeks. In this article some
of the potential uses of such maps are identified within an overall business
context: firstly, the context of uncertainty thresholds in decision making; and,
secondly, the context of effective organisation and utilisation of tacit and
explicit knowledge assets in new partnerships.
In the first of this series of musings
on the nature of sin in the discipline of Knowledge Management I examined a
mortal sin of commission; failing to initiate tacit knowledge projects in the
early stages of a Knowledge Management Program.
That article outlined some early work
and thinking on knowledge mapping and proposed a taxonomy of tacit knowledge. I
also identified the need to separate the two dimensions of knowledge management:
the asset dimension, which is the most common application area; and the
dimension of judgement, which is the most neglected by current practice. From
this we can see that four types of knowledge management activity are possible,
the first two of which are the most common:
a) Sharing explicit knowledge.
b) Making tacit
knowledge explicit.
c) Sharing tacit knowledge. This cannot be achieved through an
Intellectual Capital Management System (ICM) and is as, if not more, important
than sharing explicit knowledge.
d) Empowerment by creating trust
mechanisms for the more extensive use of tacit knowledge. This is potentially
the most powerful Knowledge Management activity and requires the greatest degree
of sophistication.
In this second article I want to introduce the four types of tacit
knowledge using an Uncertainty Matrix to provide context. This will complete the
themes developed in the last article, moving from observing decisions to
identifying and classifying knowledge assets. I will then propose a basic model
for deriving greater value from knowledge maps. This requires developing a
pragmatic route forward that places in perspective the principal knowledge
management practices at present in the market. This will bring me full circle to
the starting point of the first of these articles, the development of effective
models for tacit knowledge exchange and learning in organizations.
The
business context of knowledge management
This concept of differing levels of
uncertainty is key to understanding the appropriate balance between tacit and
explicit knowledge. The uncertainty matrix pictured above illustrates this. The
further we are to the right of the matrix, the more uncertain we are about our
objectives. This can arise from the speed of change in the marketplace, high
levels of competitive activity, entry of new forms of competition etc. The
further we move down the matrix the higher our uncertainty about how to achieve
our objectives. All organizations inhabit different quadrants of this matrix at
different points in time.
Models of rational decision making assume that any problem can be moved
to the top left quadrant. All that this requires is some more data capture and
analysis, and using the right method to find the correct answer. Leaving aside
the validity of these pseudo-rational models, the problem is that in uncertain
markets they are never timely enough. Consequently individuals and organizations
need to develop a tolerance of uncertainty. Both have a natural, intuitive
threshold of uncertainty at which they are comfortable to operate, though where
this is set can change according to the circumstances. Entrepreneurial companies
in bio-genetics have a very different profile to established players in
defence-based engineering. However, few organizations articulate where this
threshold lies. The articulation of this would provide the where and how of
using tacit knowledge - unless this threshold is articulated organizations will
lose the opportunity to manage and refine the use of tacit knowledge.
In the last article I
identified a category of tacit knowledge use, the arrogant little shit syndrome.
In this case the decision making process has been mystified by an individual or
group of individuals trading off their past reputation in decision making (or
more often straight good luck). This is operating beyond any sensible threshold
of uncertainty; it has crossed the border line into chaos.
The where and how of tacit
knowledge types
The uncertainty matrix allows us to see the contexts in which the four
different types of tacit knowledge can be used.
1. Intuition, out
of the box thinking. This represents a high tacit response to the
conditions of uncertainty. The uncertainty may result from the business
environment or from the organisation’s desire to seek new opportunities in a
stagnating market. In this area both experience and inexperience are necessary
components of the management of knowledge (the use of the term the management
here is contrasted with the use of knowledge.). New thinking is not generation
bounded and can be stimulated, and definitely validated by the presence of more
than one generational perspective. To derive the most value from this approach
it is essential that the process of management is appropriate. Task focused,
time constrained communities of purpose (see later) are the most effective. They
provide both the intensity and commonality of purpose necessary to create the
right environment for the effective exchange of knowledge. Peer group review
processes are not appropriate in this quadrant as they generally stifle
creativity and tend towards preservation of the status quo.
2. Gut Feel, unconscious
competence, compressed experience. This tends to be exhibited in
leadership. In conditions where the direction is unclear, or objectives
uncertain, the effective leader is able to choose either the correct option or
the most robust option. A robust option is one that keeps the maximum number of
possibilities open for the longest possible time.1.
Once the best course of action has
been decided, the issue for management is one of validation.
The issue within this
quadrant is to make the most resilient series of decisions within a complex and
changing environment. Humans are particularly well adapted for working in this
environment as they are able to make decisions on the basis of sparse
information. This quadrant always requires a judgement about whether tacit
knowledge should be made explicit. However it is unlikely, except in the most
stable of environments - or those with the longest lead-time - that this will be
advisable. If tacit knowledge is decomposed into explicit then more information
and data will be required to comply with the stages in the process.
Operating in this
quadrant does not necessitate the acceptance of as extreme a risk profile as the
intuitive process requires. Here peer group validation through review is
appropriate. Some companies use a three generation workshop approach to validate
investments in this area. If the current, future and past generations of
managers agree, then the risk of operating on the boundaries of accepted
uncertainty are reduced.
3. Decision making which can be made explicit within context,
but the context itself is shifting and creating uncertainty. In this
case the objectives are clear and articulated, however there is a range of
options that we can choose from. Within this decision environment the need to
experiment and to reuse past experience requires some tacit knowledge to be made
explicit.
The
strength of decisions support systems is their ability to provide the
decision-maker with hard information, allowing them to evaluate the options
before deciding how to react. The problem is that increasingly uncertain
environments require the use of tacit knowledge within a bounded area, filling
the gaps between the more explicit, analytically based tools. This boundary can
be partially established by encapsulating tacit knowledge in procedures,
reports, statistics etc. This arena is the one which most business schools train
for.
4.
Decision making in a stable environment. This is a situation where we
all know, or can easily define, the objectives and the way to achieve them
within acceptable timescales, using an agreed method. With some exceptions the
use of tacit knowledge in this environment is not advised, as the need is for
predictability and consistency of response. Process re-engineering is especially
suited to this environment and can be extremely effective in this context. The
type of decision made in this quadrant will always be explicit. What must be
borne in mind is that those existing, stable procedures can often be the area
where an entrepreneurial competitor can really change the game. Therefore, there
is a need to ensure that processes in the upper left quadrant are periodically
driven into the bottom right, to see if new technology or other changes enable
improvement or competitive edge. The role tacit knowledge performs is in
reviewing these existing procedures.
Articulation is
key
The
purpose of this type of model is to create a context for managers to understand
the effective deployment of intellectual assets and, just as importantly, to
create a language to enable then to articulate different and more appropriate
forms of action. A baby progresses from a general articulation of hunger to a
more precise (and, in the case of my six year old son Huw, constant) demand for
Pizza, chips and peas. Equally, as we move into domains of increasing complexity
we need to create a richer language for our organizations which enables them to
select the type of knowledge required for each particular context. The problem
with too many management movements is that they have been over-simplistic
responses to deep environmental changes, offering inappropriate, cosmetic
solutions - so while they have initially promised much they have ultimately been
unsatisfactory or unpalatable. It’s no surprise that they often leave an
organisation severely dissatisfied. This has led to a kneejerk culture - a firm
may end up rejecting the management theory before it has been completed, quickly
moving on to try something else.
The uncertainty matrix provides a
means by which organizations can segment the problems they face and deploy the
appropriate intellectual asset that comes with it. It provides perspective and
can move an organisation into a more sophisticated solution-rich environment in
which different styles and responses can be understood. However, we need to move
from this understanding of where different types of decision and associated
knowledge assets can be used, to an understanding of how we can identify those
assets, merge or optimise them, if appropriate, and fill in the gaps in our
knowledge.
From understanding decisions to mapping
knowledge
In the first of these articles I outlined an approach by which knowledge
assets were identified tangentially, through the observation of decisions and
associated flows of information. The basic premise of this approach is that
knowledge only manifests itself in action. Interviewing - in particular
hypothesis-based techniques - will never discover the nature of a decision or
the associated knowledge assets, because the interviewing process takes place in
an abstracted environment. Here two key principles are reflected in the
Knowledge Diary approach outlined in the previous article:
1. Proximity; I need
to directly observe the behaviour. This is achieved by placing an apprentice
with the decision maker over an extended period of time, or through the creation
of an artificial situation which simulates the original decision process;
2.
Camouflage; I aim to reduce to a minimum my impact on the subject, so
that my presence does not influence or change their behaviour. The use of a very
junior apprentice helps in this. However, more importantly, we are looking for
knowledge by asking for information about decisions. By looking at the decisions
across a range of decision-makers the identification of the associated knowledge
assets reduces the risk for individual discrepancies.
The use of apprentices is derived
directly from the practices of the medieval craft hall: this was perhaps the
last time that we actively managed tacit knowledge in western society. At that
time most knowledge was tacit; it was the knowledge of how to do things, learnt
through imitation and practice over an extended period of time. Elements of it
have persisted through craft apprentices into the 20th Century, but its
centrality to the heart of business has been lost. While this apprentice
approach runs counter to the cost reduction ethos promoted by BPR, the cost of
losing tacit knowledge is far higher than the human costs of creating a
community to ensure its retention and growth.
In the craft hall model each
apprentice was allied to a Master Craftsman. The apprentice was no threat to the
master; the difference in age and status was such that there was little
inhibition in knowledge transfer - but too often, in modern mentoring systems,
the person being mentored is a direct threat to the master’s future job security
and knowledge exchange is consequently inhibited. Once the apprentice had
reached a certain level of competence they became a journeyman, still allied to
a master but now capable of independent work. Once a journeyman, the erstwhile
apprentice was capable of challenging the master, a process that both allowed
for the creation of new knowledge and the deeper exchange of existing knowledge.
There are two
aspects of the medieval craft hall which are of particular relevance. The first,
the non-threatening nature of the apprentice, has already been mentioned. The
second, and in some ways the most important, is the structure of the craft hall
itself. Within the craft hall the apprentices would meet on a regular basis.
They would exchange their learning, swap stories (the most effective means of
communicating complex ideas) and exchange tips and ideas. A study group on an
MBA course is a similar community of co-dependency, one created around the
survival needs of a group of individuals faced with common objectives and
problems. In this exchange, new knowledge is created and new techniques and
ideas can be imitated and dispersed. Creating spaces in which people can
exchange ideas is one of the proven best practices in knowledge
management.
This
craft hall model can be used within organizations to move from looking at
decisions/information flows to looking at the underlying knowledge assets. The
apprentices, with some expert assistance, gather in a physical space for the
duration of the project. Using a large wall, hexagonal post-it notes and pens,
they can link their understanding of decisions and information flows, captured
in their knowledge diaries, to build a diagram of the knowledge flows for the
organisation being studied. As this map develops then certain decisions start to
cluster, on the basis of the elements making up the cluster having a strong
common link. As the clustering emerges then it is possible to deduce the asset
being used. It may be a database of information, a plan, a process or a set of
experiences captured in a key individual that allows them to make a particular
judgement. As each cluster of decisions is analysed the knowledge assets can be
identified and their nature, whether they are tacit or explicit, agreed by the
group. If necessary the asset can be decomposed to the lowest level necessary to
separate tacit and explicit elements.2
Success requires the creation of a
high level of energy and excitement within the physical space. For this reason,
an absolute rule is followed: the apprentices are not allowed to be on the
project full time . They are on the project part time so that they can remain in
touch with their natural community and networks. In adopting this approach they
are able to tap into the wider knowledge asset base of their own networks. Staff
allocated full time to the team tend to loose this connection very quickly,
losing the informal networking and communication which arises from working with
colleagues.
Towards the end of the process there is a ‘discovery point’; a moment in
time when the group as a whole starts to move towards a usable model. The mess
of post-its and arrows on the wall starts to be reproduced in simpler models
using the language of that organisation, enhanced by the new richer language of
tacit and explicit knowledge. At this point the information and knowledge flows
can be optimised. Unnecessary stages, such as ‘approval stage’ decisions that
use no new assets and create no new information, can be eliminated.
I know what I know and what I don’t know - so now what?
Knowledge mapping is an essential precursor to any knowledge management activity. To be effective it is important to avoid excessive detail. We are looking for a general route-planning map with the major cities and roads identified, rather than a detailed 1:25,000 map. Often too much detail is sought too early, at the expense of understanding and the ability to use the output. This is particularly true of some software-based solutions.
Knowledge mapping, using the method shown above, takes no more than 2-3 months, and can be achieved in 2-3 weeks. It works on the 80-20 rule, whereby a 20% investment of effort achieves 80% of the requirement. Once completed the results and any further analysis need to be put into action plans so that detail can be built up over time. The whole process must be continually reiterated, as any map is only as good as its most recent survey.
Having created a map we are now in a position to approach knowledge management projects with a better understanding of what is achievable and where our priorities lie.
The knowledge map above represents an idealised process that I will describe as a series of steps:
1. Complete the identification of the knowledge assets from the decision/information flows and categorise them as either explicit or tacit assets. If the exercise is being carried out for more than one organisation (internally, externally or a combination of both) then an attempt should be made to group the assets from the different maps. This is best carried out through the same craft hall model as before; and, if the exercises take place in parallel, it will be even more powerful.
2. For explicit knowledge the question is how to identify the artefacts which currently contain that knowledge. These artefacts may be processes, filing cabinets, databases etc. Once identified they need to be collated, any duplication eliminated and processes established to make these artefacts available to those communities that need them. This may be through an Intellectual Capital Management System, a Data Warehouse or simply the consolidation of manual records. There is also a need for a supporting infrastructure, but I will return to this later.
3. In relation to explicit knowledge artefacts there is always the danger of over-perfection. In a recent assignment I came across a room full of card indexes which had been created and maintained by successive generations of field engineers for over a century. They were messy - hand drawn sketches and tips in different coloured inks. Unfortunately, a process improvement study had identified that it would be too expensive to digitise the cards and that they could not be maintained without the ‘unnecessary’ expense of field computing. The field engineers still used the cards, but with regret, knowing that the asset was falling into extinction through lack of maintenance. There was nothing wrong with the card system. It had evolved over the years and could not be sensibly digitised with the current state of technology - but why abandon it? The useful knowledge of previous generations was being wasted in the pursuit of some idealised and mechanistic command and control environment.
4. Question One. Tacit knowledge assets: Can we make this asset explicit within an acceptable timescale? The addition of the time scale question avoids the debate about whether all tacit knowledge assets can be made explicit. If it is not possible then we need to identify the key individuals who hold that knowledge and create communities of competence to develop and share that knowledge.
5. Question Two. Tacit knowledge assets: For those assets that we think could be made explicit, we need to ask the fundamental question: Should they? Partly encouraged by the fashion for data mining, there has been a presumption that all tacit knowledge should be made explicit. However, if the tacit knowledge is specific to a business context and that context is not stable, the act of making it explicit will reduce the capability of the organisation to respond to change. We have seen this in some process improvement exercises. The desire to structure tacit knowledge into regimented processes effectively created dinosaurs linked to the ecology in which they originally evolved and consequently unable to respond quickly to change, whether cataclysmic or not.
For those assets that can be made explicit, we need to create meaningful artefacts. These artefacts are best established through experiments or prototypes, if it is not possible to use existing artefacts that are preferable because they have evolved within the organisation over time. Prototyping allows artefacts to be validated and tested to establish the best means to store and communicate the piece of explicit knowledge. It should not be assumed that there is just one ‘best artefact’ in an organisation. A combination of artefacts which complement each other is better if the knowledge contained in it needs to be communicated to a variety of different groups.
6. Create communities of competence. This is one of the key knowledge management skills and the one open to the most interpretation.
7. Tacit knowledge is held by individuals or by already existing communities, the key to its retention and articulation being the ability to enable its sharing within and across communities. The question is how this can be achieved. One way to approach this is to use communities that already exist; two of which, communities of co-dependence and affinity, are naturally occurring.
a) Communities of co-dependence are the most effective natural means of knowledge exchange, but are often the least visible. These are communities of individuals who share goals, values etc. and who need to collaborate to achieve these goals (either survival or visionary). Some of these communities will be identified during the early knowledge mapping exercises.
b) Communities of affinity are also effective means of natural knowledge exchange and benefit from high visibility. Communities of affinity comprise individuals with a common role - get a group of engineers around a coffee machine and they will exchange significant levels of knowledge. However, the fact that they share this common role can also lead to rivalries, inhibiting full exchange of knowledge.
c) In contrast, Organizational communities rarely share knowledge, even when the processes attempt to force them to do so. As they are comprised of individuals forced together, it is difficult for the requisite level of trust necessary for knowledge sharing to develop. This is exacerbated by the fact that, as they are highly visible, they are therefore highly political.
d) Communities of purpose, tiger teams, SWAT teams etc., are a powerful means of knowledge exchange and achieve good visibility. However, they lose their effectiveness over time; you don’t use elite troops for a protracted siege!
A community of competence is in effect a structure created to facilitate tacit knowledge exchange in a managed way. This means that it is artificial. The danger for companies is that they attempt to use one of the above four models, rather than merging the different types together to create something new, capable of moving from accidental use of knowledge to its conscious management. This is most effectively achieved by making this process evolutionary in nature. Starting to network communities of co-dependence, reinforcing that network with carefully selected SWAT teams built around affinity relationships, is a very powerful way of merging these models (it also the requires a far more detailed explanation than I can reproduce in this article.). The key to achieving this synthesis is the need to actively manage the ecology by understanding and accepting what exists and changing the environment - both subtly and dramatically - at key intervention points, to allow a different style of competence to emerge.
1. Having identified and started the process of creating new communities of competence to safeguard and grow our tacit knowledge base, we now need to do two things. Firstly, we need to target (particularly in a take-over or merger) key individuals and identify the commodities that are required to retain them - it may not always be loyalty bonuses. We then need to empower tacit knowledge workers within a community that acts as a control and check. This in turn helps build the community through the positive reinforcement of desired behaviour. Those communities that form first should be rewarded through the provision of an infrastructure to facilitate knowledge exchange. A host of formal methods for human and resource management, education and cultural change exist to support this type of process, therefore there is no need to reinvent them for knowledge management.
2. The supporting infrastructure is the link between tacit and explicit knowledge. Tacit knowledge requires the support of explicit artefacts. All organizations need to provide an infrastructure that facilitates the use of these artefacts and builds on the exchange of knowledge between workers by encouraging effective communication. Intellectual Capital Management Systems are a powerful way of achieving this. Before an ICM system is implemented there is a large amount of groundwork that needs to be done to understand the nature and use of knowledge assets and the language in which they are couched. Too many companies leap straight to the end point, without the necessary preliminaries.
One of the most effective early stage ICM systems I have seen was, in essence, a grey Internet built in lunch breaks and evenings by an individual who believed in knowledge and had enough senior management backing to permit use of company resources. The system had its failures but has now evolved beyond its original scope to the point where it needs to be formalised and possibly rebuilt. However, what is now built will take into account the lessons learnt from the low cost mistakes of the early experiment.
Artefacts & Communities: the essence of managing knowledge
The management of artefacts and communities is the essence of what we know as knowledge management. In order to arrive at this point we need to start with an understanding of what we have, its nature, what we need and, maybe, what we can discard. Once we know this we will have a powerful means of achieving competitive advantage through the effective deployment of those assets.
While this is important for a single organisation, it is even more so for a merger, acquisition or new supply chain partnership. We need to understand the intellectual assets of all parties in order to achieve a symbiosis, not a parasitical or predatory relationship. I believe, but have not yet proved, that the costs of this type of exercise would be more than recovered simply through the reduced requirements for crudely targeted loyalty bonuses - let alone the cost of intellectual capital which simply walks out of the door, or remains undiscovered and unused.
The process I have described over these two articles is both:
1. Easy; it is kept simple through application of the 80/20 rule. Knowledge mapping is organic not mechanistic. We do not require a full set of engineering drawings to get moving, or to get back on track.
2. Difficult; in that it requires a degree of wisdom in its application. We are still experimenting with the techniques and will continue to do so. Resisting attempts to standardise consultancy methods into neat boxes is important at this stage in the development of knowledge management.
I am, however, convinced that if knowledge management survives into the next century then knowledge mapping will be a standard activity in companies, each of which will have their own equivalents of the British Ordnance Survey to maintain those maps. If knowledge management survives it will depend on the community as a whole not falling into sin. Here there are two dangers:
 |
the venial sin of omission - failing to act knowledgeably in executing knowledge management |
| |
the mortal sin of commission - here knowledge becomes yet another management consultancy fad. |
In the next and final article of this series I want to return to my original point of departure and look once again at the sin matrix. I will use this to put the lessons of the previous two articles into perspective and from this formulate a series of golden rules and methods of action which need to be considered when implementing any knowledge management program.
David Snowden is at IBM. He can be contacted at:
snowded@uk.ibm.com
