Engaging tensions of knowledge management control.

Abstract

Any proficient activity calls for balance, which normally entails striking the right median between extremes. Management needs to strike the right balance when it comes to deciding how best to control resources, processes, and assets under its purview. Managing knowledge management programmes is no different, and entails a judicious balance in managing the tensions of centralisation and decentralisation. The important question is: When should one choose a centralised mechanism for knowledge management versus a decentralised one? The answer lies in three aspects: type of knowledge process, type of knowledge workers, and type of knowledge we are seeking to manage.

**********

Of the many definitions of the term 'tension', the following most closely captures our connotation of it in this paper: "balance maintained in an artistic work between opposing forces or elements" (Merriam-Webster Online Dictionary, 2005). This definition is apt for two reasons. First, the term 'artistic work' accurately describes the process of constructing a knowledge-based organisation, or re-architecting an existing organisation to one that is more knowledge-centric. In addition, breaking down the term 'artistic work 'into 'art' + 'work' helps bring out a primary tension: knowledge management is both an art and work. The art element has to do with having imagination and vision about what a knowledge-based organisation should look like, identifying its attributes and its dimensions, and examining how each component of the organisation fits into the overall architecture. The work element has to do with constructing the vision. One might think of work as the job of transforming the architect's plan, a logical design, into a physical reality like a building or piece of machinery. Unless one can strike the right balance between the art and the work product, there will be disconnects leading to a poorly created knowledge organisation.

Two common disconnects are (a) when the art conceptualisation is excellent but is followed by poor work, and (b) when the art conceptualisation is poor but is followed by excellent work. In the first instance, we have a good design on paper but it is implemented poorly; in the second case, we have an unimaginative design that is implemented correctly. To promote art, we must allow for creative work through decentralisation and looseness in control. Conversely, to facilitate work we must strive to control the process of engineering for which we require centralisation and tight control dynamics.

The second item of interest in the definition of 'balance' is the concept of 'opposing forces'. Tensions that we describe in this paper can best be characterised as opposing answers to important governance questions such as: Does an organisation exercise tight control over its knowledge management process or allow it to be conducted loosely? Should an organisation choose centralisation over decentralisation? Should order be achieved through top-down or bottom-up directive mechanisms? Questions such as these are important and must be considered when embarking on a knowledge management programmeme. These tensions exist due to the monotonic manner in which most organisations operate knowledge management programmemes. The naive organisation will attempt to apply a one-style or one-flavour approach to all knowledge management efforts--this is a recipe for disaster. Tensions also exist due to lack of appropriate imagination about the complexity of managing knowledge. What works well for one aspect of knowledge management, let's say knowledge creation, will be futile in another context. The challenge, as we see it, is for managers to be imaginative and identify ways to engage knowledge to best meet organisational needs.

Managing Knowledge

Managing a knowledge management programme is no easy feat. While there are many reasons for this, we will focus on a single aspect: how to best control and command the knowledge programmes of the organisation. We feel that executives face an unusual surge of tensions in this area, all of which can be reduced to one of control dichotomies.

At the fundamental level, executives think they must choose between centralised or decentralised approaches to control. Centralised approaches to control can be characterised by tight structure. Decision fights are normally restricted to one individual or a select few. Order is therefore imposed in a top-down fashion through directives from top management, or from senior personnel with decision rights. These individuals normally determine what is permitted, how the organisation should conduct work, and how order is ensured and governed. Decentralised approaches are the opposite; here decision rights are available to individuals across all hierarchical levels of the organisation. Order emerges mostly from a bottom-up direction, rather than being imposed in the top-down mode. This is because individuals have the right to act on information and knowledge to make decisions--decisions that affect not only their own work, but also their surroundings, such as their team or their department. As we expand the concept of surroundings and the number of decision makers, order emerges from these actions at the lower level. The question becomes: When should one choose a centralised mechanism for knowledge management versus a decentralised one? The answer lies in three aspects: type of knowledge process, type of knowledge workers, and type of knowledge we are seeking to manage.

Type of Knowledge Process: Knowledge Creation versus Knowledge Commercialisation

Knowledge management can be viewed as a two-phased approach. The first deals mainly with the creation of knowledge. This phase represents knowledge generation, and comprises of knowledge sharing, storage, transfer, and application, all geared towards the generation of new knowledge. One might consider this akin to the act of invention. In order to invent, we must be able to go through the knowledge creation cycle multiple times and, hopefully, obtain new knowledge with each cycle.

The second phase of knowledge management is the commercialisation of knowledge. This is where we would like to transform the invention into an innovation. The differences between an invention and innovation are not inconsequential (Schumpeter, 1934). Innovations are commercialised inventions that call for taking inventions to the market and being successful at transforming them to realise economic ends. Knowledge creation and knowledge commercialisation feed into one another. New knowledge created will, hopefully, lead to new inventions that should inform future commercialisation processes. Likewise, commercialisation of knowledge should lead to discovery of weaknesses in products and services, which should serve to inform future knowledge creation activities.

Each of the two processes requires different governance approaches. Knowledge creation calls for a fluid space--one that can promote creativity, rich exchange of insights, debates and dialogue, and also nurtures new ideas. As an example, consider some of the research and development labs: these are spaces where each scientist operates, for the most part, autonomously. Scientists share knowledge voluntarily, because the sharing will benefit each other's inquiries and experiments. Order is not imposed on them, but emerges from their mutual interaction. Take the case of scientific communities: in these spaces, norms and best practices emerge as knowledge is shared, dialogues are conducted, and debates surface. We can even assert that order emerges out of evolutionary cycles. Best practices are shared, and debated. In the process, some practices are refuted, others get refined, and new practices emerge with the strong ones surviving over time. Bottom-up order is the dominant style of operations.

To foster knowledge creation, management must ensure that they operate at a distance, and do not impose on creative cycles or actions. Management should not be involved with dispute resolutions and the emergence of norms: these should be left to the knowledge workers. Self-governance and self-organisation are important strategies to appreciate here. Often, management feels that if they are not directly involved in a project or participating actively, they are not doing their job. This is not so; management needs to act as a catalyst here, by promoting the creative environments, providing resources for knowledge creation to occur, and also seeking new ways to enhance the work processes of the knowledge workers. Decision rights, however, cannot reside exclusively with the managers but must be delegated to the knowledge workers on site. The best analogy we can provide to help managers envision their role in the knowledge creation space is that of a funding agency. Management is the provider of funding for interesting projects. Each knowledge worker has the right to collaborate with fellow workers or work independently and put together a proposal for funding. In compiling a proposal, the knowledge workers have the burden of clearly articulating why the project is important to the organisation, how it contributes to strategic objectives, what the deliverables are, when these will be delivered, and how they build upon the existing knowledge stock of the organisation. This is significant, as the knowledge workers are putting in place metrics against which they think they should be measured. For instance, if the project says a deliverable will be handed in on a particular date, and this does not occur, only the knowledge workers can be held accountable, since they put the proposal together.

Management must take a portfolio approach and try to balance out investment in the various endeavours and ensure that the most interesting and important projects are funded in a timely manner. Ultimately, through the rules of evolution, only the strong will survive in this approach. Knowledge workers who continually under-achieve and do not deliver on their promises will not be respected by their peers, and hence will be involved in fewer proposals and initiatives, eventually resulting in their exit from the organisation. On the other hand, those who are successful will ascend in the ranks, lure other workers to take on joint engagements and, hopefully, lead to better management of organisational resources.

Allowing order to emerge has another interesting benefit: the emergence of core competencies. In due course, the organisation will learn to identify areas that are popular and successful from areas that are not. It can then plan on how to better leverage the successful areas for organisational ends. Additionally, it can help address issues in non-successful areas, if the exercise benefits the organisation. In the final analysis, we must remember that management can easily destroy knowledge creation by trying to micromanage the process and seeking to control rather than direct knowledge creation efforts. The successful management approach is one that seeks to provide direction without direct imposition and interference with creative cycles.

Let us contrast the requirements for knowledge creation with those for knowledge commercialisation. We are concerned with what it takes to turn an invention--the outcome of a successful knowledge creation exercise--into an innovation. Innovations can be internal or external to the organisation. For example, a group may invent a best practice such as a knowledge process, and then seek to commercialise it in the organisation by proliferating its acceptance and adoption. Similarly, the R&D lab may develop a product enhancement that needs to be commercialised for the external market. Interestingly, many of the 'rules' for successfully transforming inventions into innovations are similar, regardless of internal or external focus. This is because, in both cases, the invention must be taken from its 'original' space and supplanted into a 'foreign' space.

Successful knowledge commercialisation occurs when an organisation has a systematic process (Brown and Duguid, 2001). An architecture that is efficiently and effectively managed needs to be in place to take the invention and transform it into an innovation. Consider the analogy of a manufacturing plant whose goal is to transform a logical idea into an innovative product by marketing it appropriately, producing it efficiently and effectively, dispatching it to the end-consumers, and having provisions to take in feedback from the consumers on usage issues. Successful manufacturers have an optimised, repeatable, well-defined, and tightly controlled process for commercialising inventions. To ensure success, such organisations must forgo decentralisation for centralisation to ensure that the process can be better controlled and order can be more definite and static.

Let us look at the management of the human resource function of an organisation. What would happen if, every time we needed to hire someone, we reinvented the recruiting process? The organisation would be engaged in an ineffective and wasteful usage of corporate resources. The process of recruiting, for the most part, is fairly straightforward, and needs to be managed by exerting centralised control. The very purpose of human resource divisions in organisations is to represent centralisation in functions. Without this, each person will be left to his/her devices when recruiting, resulting in duplication of effort, lack of coordination, and a disorganised reality for the organisation to contend with. Centralisation of functions ensures that we have one entity responsible for conducting actions within a domain rather than having them conducted in a dispersed and haphazard manner. In addition to centralisation of the function, we also centralise decision-making rights. The goal is to have a coordinated and unified effort on management matters.

Commercialising knowledge calls for similar dynamics. We must put knowledge created, that is, inventions, through a well-regulated and systematic process. This process must continually be refined as we learn how to improve the aspects of knowledge commercialisation. Centralised management of the process is important to ensure appropriate control. As suggested by the theories of transaction cost economics, it is better to be vertically aligned and hierarchically structured when one is engaging in transactions that are repetitive, frequent, and where organisations need to reduce uncertainty associated with outcomes (Coase, 1937; Williamson, 1981). As argued by Ronald Coase, the Nobel laureate economist, one reason why organisations exist is to increase the efficiency of conducting transactions. Commercialising knowledge can be considered akin to conducting transactions where efficiency is critical. Consider this example: when an organisation is making routine purchases there is not much uncertainty associated with the process or the outcomes because there is a fixed protocol to follow. This is what knowledge commercialisation is about.

There are many cases of organisations that are successful at inventing but poor at innovating. Xerox is one such organisation that has been plagued by failures to profit from its own inventions. The researchers at PARC, Xerox's research hub, have been responsible for some of the most critical inventions in many fields, especially computing. For instance, it was the PARC researchers who devised the GUI, yet it was Apple that was successful in commercialising it to inform personal computer monitor designs. One of the reasons for this innovation failure was lack of recognition for the different engagement ivies for knowledge creation versus knowledge commercialisation.

Organisations that are successful at creating and commercialising knowledge normally impose the two necessary but different control regimens essential for success. Consider the example of knowledge creation and commercialisation in the defence departments of various countries. Knowledge creation is relegated to research labs that are funded by the defence budget. These centres assemble researchers of the highest calibre and give them space, resources, and time to come up with inventions. Most of these centres, such as the US Army Research Laboratories or the US Office of Naval Research work on the basis of grant funding. Researchers--the knowledge workers--seek funding for their projects by demonstrating how they contribute to the various agendas of the agencies. Researchers here not only engage locally with other scientists but also collaborate with academicians and private sector enterprises as required, to get access to knowledge and improve their chances of inventing successfully. The knowledge commercialisation side of the organisation is handled by the armed forces, which operate in a top-down fashion. Once innovations are accepted and introduced to the organisation, they are pushed down hierarchical ranks via directives. There is very little ambiguity in this process as soldiers get their orders, and are expected to follow them. Order-based procedures result in efficiency in the flow of information and the acceptance of innovations in a timely manner, thereby promoting effective knowledge commercialisation.

It is interesting to note that both sides, the research and the armed units, work in synergy with one another. The armed services implement the knowledge invented by the researchers, and in doing so inform the researchers on where to direct future knowledge creation so as to better their chances of victory in engagements. This results in the needed inventions being calibrated, that will, hopefully, be more successfully adopted by the armed services, since it helps them achieve their military goals. Success can be linked to the fact that the knowledge creation and knowledge commercialisation units are managed using different principles and are allowed to be successful by recognising the differences in decentralised and centralised control regimens.

Type of Knowledge Workers: Standard versus Radical

Knowledge workers in organisations range in types and specialisations. Consider a simple dichotomy--those who perform standardised tasks versus those who work innovatively and are radical in their use of knowledge. Both types of workers use knowledge, but the different ways they use it requires corresponding differences in management mechanisms.

Knowledge workers who apply knowledge in a standardised manner are normally engaged in routine tasks. These tasks can be monitored, regulated, and controlled since they are predictable in design. Success measures can also be easily calibrated. Consider the example of customer call center operators. These knowledge workers are bound by routine procedures that govern how they answer a call and resolve customer queries. Success measures to manage such workers are based on time and motion studies such as how quickly the call was resolved, if the customer needed to call back, etc. Knowledge workers who are engaged in standardised tasks are better suited for centralised management controls since they need to be monitored against well-defined objectives, and are also less skilled in independent application of knowledge. Therefore, decision rights on knowledge should be controlled and their application restricted.

Radical knowledge workers are those who work with knowledge in unconventional ways. First, the knowledge workers in these domains can be considered high-end knowledge workers (Davenport et al, 2002: Awazu and Desouza, 2004). They know their work better than anyone else, and hence are best suited to devise protocols on how to manage their work processes. Second, since these workers thrive in creative and loose environments, they do not do well under the imposition of external pressures. Third, since the work performed by these individuals is non-routine, it is not suitable for a tailored style of management. Time and motion studies are not suitable for these workers, as there is no single way to judge the quality of their work. Decentralisation is the best approach to ensure optimal knowledge creation.

Organisations cannot order people to innovate, but must actively seek out inventions and nurture them into innovations. In our experience, only a select few knowledge workers take the time. effort, and risks involved in conducting innovations. These individuals are normally considered radical knowledge workers. as they seldom follow the status quo and do not mind spending the time or effort to build a viable alternative. When such radical knowledge workers are successful, they are given glamorous rewards: when they fail, they are derided and ridiculed. The history of innovation tells us that the radical knowledge workers will fail more often than succeed. Failure, however, is not a bad thing. New knowledge is often discovered during a failed experiment. Ask any radical knowledge worker of the times he/she failed on an experiment on knowledge creation and you will probably hear an echo of Albert Einstein's famous words: "I haven't failed, I just found 100,000 ways that don't work."

High-end knowledge workers are best managed in decentralised forms. For instance, consider the recent surge in open source software development efforts (Awazu and Desouza, 2004). The unique characteristics of the open source approach distinguish it from existing commercial software development practices. Participants in open source are globally and virtually distributed, and usually never meet face-to-face. These geographically distributed participants successfully coordinate software development without traditional mechanisms such as design processes, and schedules (yon Hippel and yon Krogh, 2003).

In most organisations today, access to and ownership of knowledge is viewed as a source of power. As such, seldom is this power openly shared or parted with. Most of it is made available on a restricted need-to-know basis, resulting in a few individuals controlling much of the organisation's knowhow. This practice is akin to how most commercial vendors provide only object or binary software code. However, participants in open source networks share the source code, which not only allows for rapid improvements to the source code, but is also a key determinant of openness in the community.

Communities play an important role in open source. Individuals care about the reputation obtained from their peers. This concern can deter individuals from producing lower quality works. This kind of peer-level horizontal monitoring differs from vertical monitoring such as in the boss-subordinate hierarchy, because decision-making does not have a top-down approach. Rather, it is a bottom-up approach, where a peer-voting scheme is implemented to seek fair process results and consensus. Self-governance schemes enable morn open participation as the end-users and programmers implement order on the practice, rather than relying on an 'outsider', such as management.

High-end knowledge workers are managed by peers, seldom by management, and in cases where they are so managed, conflicts usually arise. Consequently, the organisation must ensure that it can decentralise the management of high-end knowledge workers to communities. As in the case of open source movements, the community ensures the rules on member participation, what is accepted behaviour and what is not, how to monitor performance, and other details. We can see a prominence of self-governance rather than directive or top-down management.

The final issue in terms of tensions deals with how best to evaluate the efforts of knowledge workers. On one hand, we have process-based measures of control. These involve measuring a knowledge worker's performance against a set of outlined procedures. As long as all the steps are followed, the knowledge worker has acted in an appropriate manner. Such measures are apt for dealing with knowledge workers who perform standardised tasks. The other measure of performance calls for output-based evaluations. Here, instead of specifying a process to be followed, the knowledge worker is allowed to be resourceful, and judged on the final output. This measure is apt when workers are required to think creatively about how best to achieve objectives and do not require micro-management. Workers who need to be managed under the output-based performance measure are high-end knowledge workers such as doctors, scientists, researchers, writers, and poets. Process-based and output-based measures are at two extremes and we can have balances between them. We see these balances in most organisations, such as when most employees are required to adhere to certain process-based protocols, such as sending E-mails, or work etiquette requirements, and where output-based measures are used in their performance evaluations to measure success or failure in terms of objective achievements.

Type of Knowledge: Public versus Private Knowledge

Organisations are laden with different types of knowledge that needs to be managed differently in terms of centralisation and decentralisation. One categorisation that can be used from the organisational perspective is the difference between public and private knowledge. Public knowledge represents information that is made available to all members of the organisation. This knowledge includes organisational procedures, products, services, and other necessities needed to perform work. No organisation can exist without some level of public knowledge, also called common knowledge.

Public knowledge can be seen as the integrating mechanism that binds the various constituents of the organisation. For example, consider the process of how a cappuccino is brewed by Starbucks franchisees. If the baristas at the various franchisees did not share common knowledge on the process, each customer could get a slightly different version of the beverage. Moreover, Starbucks would find it difficult to move knowledge workers from one location to another, as the various franchisees would have different routines in place. This would be problematic for effective and efficient operations, and would result in the essence of the franchisee-based business being lost. Similarly, customer service representatives of the organisation must share common knowledge on how to manage grievances so that there is a unified method for interacting with customers. Management of common knowledge---or public knowledge in the context of the organisation is the focus of many knowledge management discussions. Such management, especially when it is put to use, is handled through decentralised mechanisms.

Public knowledge is allowed to flow loosely within the organisation, and can be freely used by its various members. Decentralised regimes are apt when we do not suspect workers will need extensive supervision and the knowledge that is in use is innocuous, that is, sharing it would cause no harm.

In contrast, private knowledge is information that is closely held by an entity and not made common to the public. In the context of organisations, we are concerned with knowledge that is private to the organisation and not available to the general marketplace or the organisation's industry. As an example, consider the case of the formula to make Coca Cola. This formula represents the organisation's private knowledge and is the source of its competitive advantage. This knowledge is of high value and can cost the company dearly if it is leaked or otherwise misplaced.

Managing knowledge that is highly private to the organisation is best handled via centralised rather than decentralised mechanisms. Opportunism--that may be defined as devious conduct--must be controlled to ensure that the value of private knowledge is maintained (Williamson, 1975). The use of centralised mechanisms will help preserve the proprietary nature of knowledge, because they can exercise strict limitations on access to the knowledge, how it is used, and who has decision rights on it. These limitations are difficult to achieve in decentralised regimes where it is difficult to monitor all who interact with the knowledge and how it is used. Knowledge that is confidential or sensitive should be available on a need-to-know basis (Desouza and Awazu, 2004).

Many organisations have secure locations where they conduct knowledge management activities on highly sensitive knowhow. These facilities can be physical (such as a separate office building), or logical (such as secure spaces on corporate intranets that a select group of knowledge workers is privy to). It is important to note that decentralisation works well for knowledge that is innocuous and public to all organisational members, whereas centralised regimes are necessary to ensure that private knowledge remains so.

Conclusion

Managing tensions between centralised and decentralised management approaches is salient to building successful knowledge management programmes (see Table 1). Choosing the right management strategy depends on the type of knowledge, knowledge worker, and knowledge process in question. Each knowledge management programme is different and hence unique, and must be carefully examined before choosing a management approach. When it is essential to monitor the knowledge, knowledge process, and knowledge worker, a centralised management approach should be taken; when the opposite is true, management should be more decentralised.

Rather than viewing centralisation and decentralisation as extremes, we suggest that organisations should be engaged with the two realities and embrace them as warranted by environmental circumstances. We have suggested several possible strategies. First, encourage the use of decentralised management for knowledge creation, while employing a centralised regime for knowledge commercialisation. Second, manage high-end knowledge workers through communities of peers to foster decentralisation approaches, while managing knowledge workers who apply knowledge in a standardised manner through centralised approaches. Third, use a centralised approach to manage knowledge of a sensitive nature and a decentralised one for common knowledge.

When organisations are first formed, there are many informalities in how knowledge is generated and transferred. Formal mechanisms eventually replace informalities to enable efficient work and economies of scale. As organisations mature, they tend to favour formalised and centralised procedures over decentralised ones--this may be appropriate for most organisational operations where efficiency is important. However, it is not suitable for knowledge creation. Organisations need to tread the free line of balancing what is important for current efficiencies against how best to motivate resources to innovate for a successful future. This is what many have referred to as the ambidextrous organisation. Adroit balancing between centralisation and decentralisation management approaches is critical to achieving this objective.

References

Awazu Y and KC Desouza, 2004. "Open Knowledge Management: Lessons from the Open Source Revolution". Journal of the American Society for Information Science and Technology, Vol 55 No 11, pp 1016-1019.

Brown JS and P Duguid, 2001. "Creativity versus Structure: A Useful Tension". Sloan Management Review, Vol 42 No 4, pp 93-94.

Coase RH, 1937. "The Nature of the Firm". Economica, Vol 4 No 16, pp 386-405.

Davenport TH, RJ Thomas and S Cantrell, 2002. "The Mysterious Art and Science of Knowledge-Worker Performance". Sloan Management Review, Vol 44 No 1, pp 23-30.

Desouza KC and Y Awazu, 2004. "Need-to-Know: Organisational Knowledge and Management Perspective". Information-Knowledge-Systems Management, Vol 4 No 1, pp 1-14.

Merriam-Webster Online Dictionary, 2005. Website: http://www.m-w.com/

Schumpeter JA, 1934. The Theory of Economic Development. Cambridge, MA: Harvard University Press.

von Hippel E and G von Krogh, 2003. "Open Source Software and The "Private-Collective," Innovation Model: Issues for Organisation Science". Organisation Science, Vol 14 No 2, pp 209-223.

Williamson OE, 1975. Markets and Hierarchies: Analysis and Antitrust Implications. New York, NY: The Free Press.

Williamson OE, 1981. "The Economics of Organisation: The Transaction Cost Approach". American Journal of Sociology, Vol 87 No 3, pp 548-577.

Kevin C Desouza

Yukika Awazu

Institute for Engaged Business Research, USA

Kevin C Desouza is the president of the Engaged Enterprise. He serves as the director of the Institute for Engaged Business Research, the think tank of the Engaged Enterprise. Yukika Awazu is vice president of the Engaged Enterprise and a research fellow at the Institute for Engaged Business Research, the think tank of the Engaged Enterprise.

Table 1: Managing Tensions of Knowledge Management Control

                Centralised           Decentralised

Knowledge   Suited for knowledge   Suited for knowledge
Process     commercialisation      creation

Knowledge   Suited for managing    Suited for managing
Workers     standard knowledge     radical knowledge
            workers                workers

Knowledge   Suited for managing    Suited for managing
Type        private and highly     public and social
            sensitive knowledge    knowledge