Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 What Is KM?

Larry Prusak, an engaging polymath who knows plenty about KM’s origins and history, and had a hand in introducing it to the wider world, argued its history and rapid development could be attributed to three trends: globalization, ubiquitous computing, and the attention to the knowledge-centric view of the firm (Prusak 2001). Hence KM’s most obvious feature – it is multi-faceted, many-sourced, and several-languaged and not yet a coherent academic field with an established body of ideas, methods, and target phenomena. Its pluralism has created considerable confusion that challenges those trying to map the field (Earl 2001; Mehrizi and Bontis 2009); indeed many think it no more than a passing fad, old wine in new bottles with no insights not findable elsewhere in simpler language (Hislop 2010). I am sympathetic to Prusak’s characterization, and will lean on it, but also believe his story can be re-framed within our evolving insights into human action, especially within organizations. At bottom, KM means managing the relationship between knowing and acting in organizational contexts, part of which is managing the processes of knowing and learning towards organizational ends. Organizations are not new, so neither is KM; Roman and Florentine bankers kept accounts; Josiah Wedgwood kept a close watch on production costs. Computers and statistics have added new flavor but we should not consider computing a ‘cause’ of KM; after all, computers only do what we tell them to do, they are KM’s powerful tools not its causes. So we need to look behind Prusak’s categories to clarify the knowing – acting relationships he intuited between globalization, computing, and the knowledge-based theory of the firm. The last is especially important for not all writers see KM as narrowed onto organizational contexts; yet it is crucial to see KM cannot and does not embrace the entirety of human knowing. It always hinges on a ‘theory of the firm’, a boundary concept that separates organizational knowing from broader epistemological matters. Inattention to this boundary is the primary source of our field’s confusions. The confusion is most obvious when KM writers set out by trying to define human knowledge – instead of starting by defining the firm as the context that gives ‘knowledge’ its particular and manageable meaning, and establishes how KM might create economic value.

The most familiar theory of the firm (ToF) is of the firm as a rationally designed goal-seeking mechanism to transform inputs (factors of production) into outputs, goods and services. Such mechanisms generate and consume data about their production processes as well as about the markets in which the relevant factors of production are acquired and into which the goods and services produced are delivered. Data is essential to managing this type of firm, and is all that is needed. If the mechanism model was the only ToF of interest to practicing managers KM would be never be more than the timely generation, collection, movement, storage, analysis, and delivery of data about the firm and its operations. Indeed the vast bulk of our literature takes this ToF for granted, even as it is seldom spelt out. It follows that if the mechanism model has been adopted KM cannot be distinguished from IT or what used to be called EDP (Electronic Data Processing) or, before the computer age, ‘managerial accounting’ – which would take us back as far as the Ancient Egyptians, and beyond. With a data-oriented mechanistic definition in mind, many authors argue KM’s objective is to make the firm’s data-handling more efficient, in particular to discover, collect, and protect data that is ‘hidden’ or ‘lying around’ overlooked in the organization. There is nothing ‘wrong’ about this re-labeling, except the term ‘knowledge’ is not as readily pinned down as the terms ‘information’ or ‘data’, so re-labeling EDP and IT as KM introduces considerable and unnecessary confusion; indeed anything that might be said about improving the utilization of the firm’s data can be said more clearly by avoiding the term ‘knowledge’.

Knowledge is an exceedingly challenging concept, yet the urge to talk about it in our ‘knowledge economy’ seems irresistible because the term ‘data’ does not address all our concerns (Powell and Snellman 2004). A glance through the KM literature shows that our field’s defining ambition, why it looks beyond EDP and IT, is to reach beyond the data-framed IT discourse towards other aspects of real world organizational practice – though the authors who take this beyond-IT position are in the minority. Nonetheless the rest of this chapter pursues this minority view in the belief that KM is not simply one of accounting’s or EDP’s or IT’s subfields but is a discrete intellectual discipline whose boundaries and problematics have not yet been adequately articulated. Our challenge is to do this in ways that clarify rather than confuse, and to show what might be achieved thereby. In short, we need to know what KM is before we can discuss its origins, history, and development profitably.

Again, many authors start out with definitions of ‘knowledge’, and a common proposal is that knowledge embraces both explicit data and ‘implicit’ or ‘tacit’ knowledge that cannot be treated as data (Mehrizi and Bontis 2009). This is an epistemological distinction and drives what these authors means by KM. Others use different knowledge-typologies or ‘epistemologies’. Data, information, knowledge, and wisdom is popular – and there are several others evident in our literature. If the resulting confusion could be cleared away we would see that there are several different notions of KM, each contingent on the particular author’s chosen epistemology or definition of ‘knowledge’. Conversely, it is clear the majority of the field’s writers define knowledge as data, and this determines ex assumptio what they mean by KM – part of IT. In contrast, this chapter argues (a) managers’ concerns cannot be limited to data alone, and (b) there can be no satisfactory managerial definition of ‘knowledge’ and that it is ‘epistemologically naïve’ to think so. The struggle to establish an overarching unproblematic epistemology has been going on for millennia, and is not going to be over any time soon. A better way to grasp KM is to recognize and exploit the variety of epistemologies (notions of knowledge) already available to us. Thus to try and base an explanation of KM on a single definition of knowledge is simply a strategic error; it cannot work, and our discipline’s several decade history of failure should have made this blindingly obvious – the data is in. The alternative is to focus on the ToF that defines the type of knowledge to be discussed – to think ‘firm-first’. We have to know the firm before we can know the kind/s of knowledge it requires to exist and prosper. To repeat, if managers’ favored model of their own firm is a rationally designed machine, then data is the only kind of knowledge needed and KM is part of IT.

The most common move beyond the mechanistic ToF is towards a ‘learning organization’. The focus shifts from ‘knowledge utilization and retention’ and onto ‘knowledge generation’ (and ‘forgetting’). The mechanistic model does not lead to interesting explanations of knowledge generation – we know of no machine that churns out new knowledge as its crank is turned (though many write about ‘innovation management’ and Thomas Edison’s ideas about planned innovation remain relevant and interesting). Machines transform, they do not create. Likewise the extensive literature on ‘organizational learning’ is not as helpful to KM authors as it might be because it is does not successfully disentangle (a) the drivers of the knowledge creation process, such as environmental change or personal ambition, from (b) the processes of knowledge creation and (c) knowledge distribution (Dierkes et al. 2003; Easterby-Smith and Lyles 2003). A great deal of KM (and innovation management) is about ‘knowledge sharing’ rather than knowledge creation. It is also not clear whether it is organizations or people that learn. To propose an organization that can learn – perhaps by changing its ‘organizational routines’ or adding to its ‘capabilities’ is to propose a specific ToF that is (a) not general, and (b) needs to be spelt out if it is to avoid mere tautology as in “organizations learn by changing routines” where an organization comprises nothing but routines. The organizational learning (OL) literature is not yet helpful on these matters and its KM practice implications are not clear. An alternative approach presumes only individuals learn, so implying a specific model of the individual that differs fundamentally from the mainstream notion of ‘rational man’. But the OL literature is not conclusive on this ‘ontological’ point either nor does it give adequate attention to the extensive literatures on developmental psychology and educational theory that treat human learning as their research topic. Again, rather than stand KM on a definition of the learning individual, we might do better to stand it on a specific ToF that captures our intuitions about this particular firm’s practices.

So long as knowledge creation implies direct movement from a state of ignorance or ‘knowledge-absence’ into a state of knowing or ‘knowledge-presence’ there is little more to be said beyond “Do it!”. Whether organizational or individual, useful models of learning demand some specification of alternative modes of knowing, transitional between not-knowing and knowing. Many authors presume experience, something that happens while in a state of mindful action, leads directly to knowing. Others look to learning from others, or to reformulating knowledge already in mind, or to intuition, inspiration, or revelation. All these models admit knowledge-as-data but also point towards other modes of knowing. Clearly attempts to define multiple types of knowledge are not going to succeed where defining a single type fails, and the challenge here is for professional epistemologists. The path for KM authors, as always, is to hinge off their chosen ToF and leverage from their far-from-complete knowledge of the firm and its modes of knowing. If the author’s intent is a KM system for a learning organization the result will not be the same as if it is for a mechanistic firm. The firm’s strategic choices determine which ToF is most clarifying for we know there is no single ‘one best way’. Strategic choice is always necessary and the resulting firm is unique and particular in strategically important ways. The mechanistic and learning models are only two of a larger pool of ToFs strategists can choose from. Thus the KM author’s hope is less to fully model and determine the firm’s design and operations than to gain useful practical insights into managerial practice that are not revealed by the simpler ToFs or KM notions. For instance, if the particular firm’s competitive situation rewards learning then management must pay attention to managing knowledge generation, memory, and forgetting, and not merely focus on increasing efficiency by discovering and handling the firm’s data better. Alternatively, if the competitive spoils are going to those such as the large pharma firms whose strategy is to safeguard their existing knowledge, in exploitations rather than explorations, a different ToF is implied – and this leads to a different KM. In general, an organization’s epistemological problems always match its strategic problems – they are part and parcel of each other. But, crucially for KM authors, they are on surer ground analyzing a particular firm’s strategizing than when grappling with the fundamental epistemological problems that engage professional philosophers. We are likely to have a better tacit sense of how to build and manage a ‘learning organization’, and analyze its knowledge requirements, than we have of the epistemological challenges of developing a scientific theory of organizational learning. “Making better use of the firm’s existing data” is transformed into “Managing the forms of organizational knowing necessary to bring this particular firm’s chosen strategy into being.” But while taking a ‘firm-first’ knowledge-oriented approach opens up new models of managing, it demands enough engagement with epistemology to illuminate the kinds of options available. Only then can we look at the strategic KM implications of alternative ToFs and begin to understand KM’s origins, history, and potential.

2 Some Comments on Epistemology

Epistemology is the branch of philosophy enquiring into the nature and scope of human knowledge so, given our field’s commitment to using the term ‘knowledge’, some epistemological homework is unavoidable. First, there are many epistemologies; no single one suffices if we are to avoid dogma, the claim to know for sure. We cannot avoid pluralism if we are to engage the real world, admitting our knowledge weaknesses. We advance whatever knowledge of the real world we have in hand through critique, so every epistemology calls for another from which to critique it. The epistemologies most familiar to Western writers are (a) positivist or ‘objective’ versus (b) interpretive, which some label ‘cognitive lenses’. Both circulate in our literature, differing but mutually informing each other. Second, as noted above, the ToF adopted separates KM from epistemology in toto. KM is about firms and there are several ToFs. Thus doing KM requires choices – of an epistemology (theory of knowledge) and of an ‘ontology’ – definition of the entity known, and of a ToF. The last is crucial. But understanding the differences between ToFs requires attention to the different epistemologies and ‘knowledge flows’ within them. Without addressing the full range of current epistemologies – which would include rationalism, positivism, critical realism, apriorism, constructivism, idealism, and so on – we can illuminate the KM writer’s epistemology-choosing process by distinguishing an objectivist approach from a subjectivist one (Hislop 2013). The first presumes all true knowledge is a representation of a rationally constructed and so knowable external reality – reality being the sum total of everything knowable perhaps. Reality’s nature and processes lie ‘out there’, beyond us, independent of and unaffected by our thinking and doing. The second proposes knowledge as more internal and human; what we generate within our consciousness to engage our world more effectively in pursuit of our goals and desires. The first inclines to thinking of knowledge as ‘object’ – possibly ‘intangible’ – but nonetheless separable from the conscious ‘knower’, because its ‘truth content’ is determined by the reality ‘out there’. The second inclines to knowledge as an indicator of the on-going processes of applying our consciousness (what lies ‘in here’) to our lived situation (what we experience of an ‘out there’).

The first approach is comfortable for those trained into positivist ways of thought, while the second is much less so and so strikes many as inherently radical – even absurd. The bulk of the KM literature presumes the first – for which there are many possible explanations, such as scholarly tradition, teachability, publishability, or other professional comforts in our obviously positivist era. Unfortunately there is a grave downside, for the positivist literature has a defect potentially fatal to KM’s grander endeavor; it offers no compelling justification for using the term ‘knowledge’ in lieu of well-defined science terms such as ‘theory’, ‘observation’, ‘phenomena’, or ‘discipline’ – whose truth content derives from their interaction. Those attracted to the objectivist view should turn first to the scientific method, for it surely bears on whatever we mean by knowledge and its generation. The scientific method’s content is lost when its terms are condensed into ‘knowledge’. Likewise, positivism-inclined writers who use the term knowledge but stray beyond the bounds of the scientific method are not likely to have a productive experience. Yet conversely, if they stay within those bounds, they have no place for the term ‘knowledge’. Note, for example, how little is lost from the IT-oriented literature when the term ‘knowledge’ is replaced by the term ‘data’, which is relatively easily defined and fits into science’s objectivist epistemology – data can be contrasted with theory and hypothesis.

Those who think of knowledge as tentative scientifically validated representations or justified beliefs about an external reality, do better using terms like theory, hypothesis, test, validity, and so on; any use of the term ‘knowledge’ is simply confusing. Note there is no knowledge-in-general; scientific knowledge is always of something specific. So to say “A has knowledge of B” is to say nothing until the statement is supported by scientifically validated and falsifiable theory and evidence about both A and B. This shows how we often talk sloppily and unscientifically of ‘knowledge’ and ‘knowing’. It seems paradoxical that many of KM’s positivist writers make comments and claims about the nature and impact of knowledge that are so obviously unscientific and un-falsifiable; especially in statements like “organizational knowledge is the key source of competitive advantage”, which is completely vacuous. In contrast, the claim that “organizational data, or organizational routines, or organizational capabilities are the source of competitive advantage” may be testable inasmuch as forms of knowing other than data, routines, and capabilities are implied even if not identified. Saying “KM is any process or practice of creating, acquiring, capturing, sharing, and using knowledge, wherever it resides, to enhance learning and performance in organizations” (Swan et al. 1999: 669) is empty language play. Absent workable definitions, the statement is purely tautological, turning on whatever the terms ‘knowledge’, ‘learning’, ‘organization’, ‘performance’, and KM are taken to mean. The bottom line is that the term ‘knowledge’ can only be used scientifically to point towards the particular body of theories and observations that comprise that science. ‘Knowledge’ is not a term within any science. This implies KM cannot be fitted into any ‘science’ of managing and one reason why the KM literature stands so obviously apart from the mainstream managerial/organizational research that presumes management can be a science. But, treating this statement as a plus that points to an opportunity, we can argue KM’s promise is to go beyond the limits of the scientific method to discuss aspects of managing that cannot be discussed within ‘management science’; most importantly, knowledge and value creation.

The subjectivist epistemological approach is no less challenging for it excises what many regard as the scientific method’s greatest strength, the deployment of objective testing to separate ‘scientific knowledge’ from mere opinion. We see another ‘knowledge paradox’. If one adopts the objectivist view, there is no justification for using the term ‘knowledge’; so ‘knowledge management’ is meaningless on that account. The scientific method dictates whatever one might mean by ‘knowledge generation’, ‘knowledge acquisition’, and ‘knowledge transfer’ and separates these terms from mere opinion. Note however the scientific method is anything but simple, and much debated by professional philosophers of science. But if one adopts a subjectivist view it is not immediately obvious there is anything meaningful or valid to be said – about anything, let alone about ‘knowledge management’. Anything goes if we have no truth criterion. Can anything in the subjectivist approach be saved and used as KM’s foundation? This chapter argues (a) yes, and (b) this is KM’s real potential – to recover from the damage done by the positivist myths about how science is done and grasp the real management work of creating firms and running them.

The Ancient Greek philosophers explored these issues thoroughly and we can learn much from their labors. Rather than anticipate modern science by thinking of ‘knowledge’ as about ‘external reality’ they explored the different ways in which they considered human beings seem to ‘know’ – by which they meant attend to the personal relationship between thinking and acting so as to act ‘knowingly’ rather than ‘mindlessly’ or against ‘proper knowing’ – which drew in the moral and ethical issues as well as matters of faith. Note the parallels to KM’s agenda of relating thinking and acting. But the Greeks’ ambitions were grander; ours are more realistic and modest. Although there are aspects of an external unchangeable ‘reality’ in their term ‘Form’ or ‘essence’, the Greek epistemologists considered many other modes of human knowing – such as techne, metis, and phronesis variously translated as ‘know how’, ‘street smarts’, and ‘situationally appropriate action’. Their list is quite long. Note too how their approach is partial, like the partial views of the seven blind men touching the elephant while none know its entirety. This is the ancient metaphor for our bounded rationality, our inability to see things ‘as they really are’. In a subjectivist epistemological paradigm ‘knowledge’ refers to the interplay of these different and contrasting modes of personal knowing. It alludes to personal experience as it looks backwards in time, and to our need to make strategic choices as it looks forwards. Our knowing remains subjective because there is no way to ‘step outside’ knowing to observe it ‘objectively’ (to reach an Archimedean ‘fulcrum’ from where everything can be seen ‘as it is’). Another way to illuminate this paradox is by asking, “Even if we were able to generate a positivistic definition of knowledge, would that be more knowledge, or meta-knowledge, or something else?” Taking up a subjectivist or knower-centered epistemological strategy the Greeks were able to (a) separate what is known ‘in here’ from reality ‘out there’, the elephant trap into which positivist epistemology falls as it conflates these and loses the knowing person, and (b) find a way of talking intelligently about knowing as an interplay of discrete and experienced modes, none comprehensive, all partial. The Greeks’ epistemological strategy lives on with those who use the explicit-tacit distinction. Positivist writers dismiss this, presuming tacit means no more than poorly expressed positivist knowledge of the real, their kind of knowledge, that is going to be restated more scientifically in due course. In contrast, those in the subjectivist camp see explicit knowledge as tentative inter-subjective discourse about a shared world, with tacit knowledge as equally tentative but shaped by the private subjective experience of living. The explicit-tacit disjunction can then be deployed to talk informingly about knowledge generation – as Boisot and Nonaka & Takeuchi did – modeled as the under-determined outcome of an interaction of these alternative types of subjective knowing. Because both are defined as bounded, they leave conceptual space for the new knowledge generated by their interaction.

The main point here is to appreciate our literature encompasses two distinct epistemological projects or ‘paradigms’ – one positivist, the other subjectivist. When we do not attend to their differences and interactions KM gets mangled in mutual confusion and distaste. The vast majority of KM writers identify with the positivist project, are scientifically disposed, and focus on extending the profitable application of data-handling computer systems. But these writers have not managed to escape IT. The computer’s correlate to reality is its universe of logical statements; it defines reality as computability. Our lived reality is very different, so the computer-oriented KM writer’s principal concerns are about the relationship between the reality within the system and our social reality. There is an academic ‘trick’ here; when the writer presumes human beings and social reality are fully rational the problems of the relationship are defined away and the system’s users become part of the computer system. The universe and everything within it are defined as computable. The activities within this system are purely data-oriented; collecting, analyzing, and acting rationally on the data provided. But, as we have seen, there is no place for forms of ‘knowledge’ that stand outside science and computability. Another way to put this is to see the positivist KM project as building an all-encompassing computer system, a clockwork universe that excludes and denies all other modes of human knowing, especially of the social and personal realities we experience. If we presume people are as rational as computer systems we have no problem getting them working together, no need for the term knowledge, and no KM project to be discussed – it is all IT. There is no space for emotion, faith, or morality – a bleak inhuman world indeed. If, in contrast, we presume people are not able to meet positivism’s ‘rational man’ standard and so do not conform to this model, then we see doing KM obliges us to go beyond positivist epistemology and science. Or, to be more precise, as we admit neither people nor organizations conform to the fully rational model we begin to scope out KM’s true challenge. It lies in finding modern ways to implement the Ancient Greek’s strategy, but at the level of the organization by, for instance, making these objective and subjective approaches complementary in the interest of understanding organizational practice better – understanding how the KM writer’s chosen ToF works.

To reiterate, much of the KM literature presumes one or other epistemological approach can be adequate on its own. This is methodological naïveté for the knowledge paradoxes noted above show it is crucial to interplay the different paradigms of partial knowing. Consequently no single approach can be fully separated out or used to generate our disciplinary process. The productive interplay of objective and subjective paradigms is already familiar to most academics through the interaction of qualitative and quantitative methods of empirical research; (a) the open-ended search for suitable data categories followed by (b) statistical analysis of evidence gathered within them. Thus Popper argued the scientific method does not reveal the real’s true nature; at best it drives those within a discipline to engage in continued experimentation and peer review to police, falsify, modify, and settle on its provisional truths. Note how experimental practice binds the subjective and objective paradigms in potentially useful ways – suggesting practice itself as a third paradigm or domain of human knowing indicated by the term ‘tacit’. Knowledge, this third paradigm suggests, should be seen as the capacity for skillful practice rather than anything in mental domain. Academics point to methodology as their skillful practice. There can be no deterministic theory of method, for then there could be no growth of knowledge.

Attending to practice opens up a new category of ToFs. In place of the firm as a bundle of economic resources, a mechanical design, a conceptual model implemented, or a cranking machine, the firm is re-defined as an integrated community of skillful practices. Of course, practice is as complex a notion as knowledge, so this may be just another tautology. But we can contrast skillful practice against both mindful and mindless practice. Many presume good practice is, or should be, the mindful implementation of good theory or at least the best knowledge available. Likewise what is learned from practice is, or should be, known unambiguously, so that experiment proves decisive. Note how these assumptions excise practice from the discussion. Keying the meaning of practice off theory denies all aspects of experience that lie beyond what is known by the mind. Polanyi’s notion of tacit contests this and points toward those aspects of practice that lie beyond the mind but can be observed as skillful practice. Note also that we cannot capture or express the totality and immediacy of practice, there is always an element of “You had to be there to understand” (Tsoukas and Mylonopoulos 2004). Similarly as Hayek noted, important ‘here and now’ aspects always get left out of an analysis (Hayek 1945) – Peirce used the term ‘indexical’. Practice is indexical, a series of fully experienced instants. Explanation, on the other hand, requires language that, because it always stands on generalities, must leave some of the indexicalities of practice behind. The uniqueness of practice cannot ever be fully articulated in language, a profound epistemological issue if we want to talk about the firm as a value-generating practice.

In summary: what can KM writers gain from this kind of epistemological discussion? First, acquire an appreciation for the unfinished and maybe unfinishable epistemological work required on the term ‘knowledge’, and thereby learn not to depend on the ‘knowledge’ notion alone. Second, that epistemological choice is unavoidable in KM projects, it comes with using the term knowledge. Third, KM projects will always lie outside the full rationality of the IT realm precisely because they connect to people and organizations that are ‘boundedly rational’ rather than computer-like. Bringing human beings into the discussion brings in bounded rationality. Human knowing is utterly unlike a machine’s knowing, just as organizational knowing differs from personal knowing because organizational life is purposive in ways much of one’s own life is not. Fourth, that organizational knowing can be usefully separated into three paradigms: objective, subjective, and skilled practice (data, meaning, and skilled practice (Spender 2007)). The first two can be captured in language, the essence of practice cannot. So, fifth, inasmuch as organizations cannot be understood within any single epistemological paradigm and KM is about managing the relationships between human knowing and organizational practice, KM must always embrace all three. It follows that the resulting discussion is always non-rigorous, under-determined, inconclusive, and open to generating surprises. Sixth, and perhaps most important, the KM analyst must recognize s/he is also an in-the-world practitioner and so does well to ground the analysis where her/his practice-based intuitions are soundest. Along these lines those inclining towards positivist thinking look to theory; anything puzzling provokes theorizing. Those inclining towards experience as the source of knowledge focus on individuals – leaders, strategists, entrepreneurs – and their Will. Those prioritizing skilled practice look to the firm as lying beyond both theory and Will, seeing it as a community of collaborative practice; the key to the last being the purpose of the practice. If purpose is not clear the processes of choosing epistemology, ontology, and ToF grind to a standstill.

With this sense of the epistemological issues in hand we can get back to Prusak’s categories. Their connection lies in his concept of the firm, which Prusak presumed to be a “coordinated collection of capabilities”. Though he did not clarify what these terms meant, it is clear that his notions of KM and its potential are held together by his considerable personal experience of firms, of managing them, and of consulting to them. He advances no theory or data, no leader-driven hagiographic model – the way most speak of Apple Inc. as a manifestation of Steve Jobs. Likewise his later discussion of ‘proxy measures’ of KM effectiveness implied no prospect of ‘objective’ knowledge metrics, a hopeless wild-goose chase, and argued to the contrary, that organizations must develop their own proxy measures that reflect that particular organization’s practices and achievements against its goals. Prusak’s indications did little to help those managing KM projects, they were more cautions than prescriptions. He concluded: “(KM) is a movement, a reaction to technology hype. There has been too much focus on technology, too little on knowledge. KM is about how companies know what they know, how they know new things, and what can be done to evaluate and transfer new knowledge. What forms does knowledge take? Those are still open questions, and we’re still learning” (Allerton 2003: 36). Prusak is not the only writer in our field worth reading, of course; he acknowledged Nonaka and Davenport, and there are many others. Yet Prusak stands out from the pack by letting his experience and intuition speak louder than his considerable scholarly, technical, and computing expertise. In this way he reinforces this chapter’s argument that it is safer to ground KM practice on one’s intuitions about a particular business’s indexical practices, warts and all, than on ever more complex computing techniques.

In the section that follows I illustrate how to read some of the currently available ToFs for their KM lessons. The chances of a successful KM project turn on grasping the firm’s ToF or what many now call its ‘business model’. Put more brutally, if the analyst does not understand how the business works s/he is not likely to do useful KM. KM success does not begin with solving the epistemological puzzles around knowledge but with understanding the firm’s specifics and indexical knowledge processes. This seems so obvious that it is scarcely worth mentioning, yet yesteryear’s firm-first imperative is scarcely mentioned in our literature today. Equally, beginning with the firm is no panacea. Not all KM writers know that there is no general tenable ToF; they slip unwitting into the mechanistic model as if it were the only one. There is no doubt the model is common. But it has a fatal weakness. It cannot explain profit and how it arises. Economists know there in no established model of the profit-generating firm. Yet firms only prosper when they make profits and most managers presume the point of KM is to enhance profitability, which makes the returns to KM interesting. Doing KM may be mere fashion, as many authors note, but at bottom it is about enhancing profit. In which case KM authors should treat the mechanistic model with extraordinary caution for it cannot be the basis on which successful KM projects can be built. It may be easy to explain and attractive to academics, but it misses the managers’ ‘main event’. Note making a profit should not be confused with minimizing losses or waste. A firm’s KM project may well be directed towards making better use of the firm’s ‘knowledge assets’, but that is not the same as engaging the firm as a profit-generating entity. This chapter’s firm-first approach harks back to the 1960s EDP notions of ‘systems analysis’ that began by observing and then rationalizing the firm’s existing processes. From this point of view, as Prusak noted, the history of KM is a pushback against BPR and the excessively mechanistic modes of corporate computerization it helped introduce. It also reflects managers’ sense of the importance of bringing people back in; computerization must support the people remaining even as it displaces many.

3 Other ToFs

If KM is to be a ‘firm first’ practice then KM authors need better knowledge of pool of business models available for them to choose. Organization theory offers models beyond the mechanistic bureaucratic one – the firm as an organism, a brain, a network, a culture, a psychic prison, and so on, marvelously laid out in Images of Organization (Morgan 1997). The weakness of Morgan’s analysis is that it pays insufficient attention to the concept of the firm as a boundary between life within the firm and life outside it – and thus to the boundary between KM and epistemology. The boundary is clear in a bureaucracy for those engaged fill precise roles marked by precise rules, accountability, performance criteria, and so on. The knowledge implications of the bureaucratic model are relatively straightforward, so it provides KM project builders with a fairly solid foundation. But the model’s fatal flaw, as noted above, is that it cannot explain profit and value-creation; so it is more or less irrelevant to major KM projects for firms in a competitive environment. When it comes to organization theory’s other models their KM implications are vague at best. We have yet to discover the secret of life so organic models are seldom more that rhetorical devices to advance their author’s biases; brain notions likewise. The cultures we see around us tolerate significant diversity, so the notion of boundary dissolves when we think the firm a culture. Psychic prison is a memorable metaphor, but again cannot be modeled without a full model of the human psyche. Network is the metaphor of choice these days, but the notion of boundary threatens to disappear entirely unless, of course, by network we mean computer network – wherein every element has to conform to a shared notion of computability. Thus the knowledge nature of a network is clear when the network elements are rational devices, utterly obscure when we mean social network. The Internet does not generate knowledge and cannot be itself a source of profit – entrepreneurs use it to actualize business models that create value in the social world, exploiting many different kinds of value, as the contrasts between, say, the Sabre booking system, Bit coins, and Facebook illustrate.

As noted earlier, the organizational learning literature has yet to offer models substantial enough to be used as for practical KM design, though there are two notable attempts – Boisot’s Social Learning Cycle (SLC) and Nonaka & Takeuchi’s SECI model (Boisot et al. 2007; Nonaka and Takeuchi 1995). These are often misunderstood and are not widely regarded as at the core of the organizational learning (OL) literature. Both presume a similar knowledge-typology and ToF, seeing the firm as a body of different knowledge types that, when interacted, would lead to knowledge creation. Boisot extended the explicit-tacit distinction into a three-dimensional I-Space by adding (a) diffused within the firm versus undiffused and (b) abstract concepts versus concrete empirical data. The KM system pushes or rotates organizational knowledge through the I-space producing a ‘knowledge gain’. The best way to understand Boisot’s model is to see it as a ‘knowledge engine’ that transforms the ‘knowledge work’ involved in moving the firm’s existing knowledge around the firm’s SLC into new knowledge (Spender 2013c). It reverses the engineer’s notion of an ‘engine’ that transforms energy (gasoline) into work (moving vehicle) as the SLC transforms organizational knowledge work into new organizational knowledge. The managed interplay of knowledge types goes on as long as the firm has the motive energy needed and chooses to expend it in this way – versus expending that energy on exploiting what the firm ‘already knows’. The strategic implication is that if management is unaware of the different kinds of knowledge within the firm, and of the value of driving their interaction, the firm will lose knowledge and die an ‘entropic’ death. The practical KM implications of the SLC are difficult to divine, though work continues at the I-Space Institute at Wharton.

The SECI model is similar, but Nonaka and Takeuchi only extended the explicit-tacit distinction into a two dimensional space demarcated by (a) knowledge type and (b) organizational location (top management versus R&D lab). Through the motion of the SECI engine the tacit product-based knowledge being generated in the R&D lab gets transformed into explicit knowledge made available to top management’s resource allocation process where they balance knowledge generation and exploitation, and between profit and further research investment (Spender 2013a). Again the practical KM implications are difficult to divine, though work continues at Hitosubashi and elsewhere. But the bottom line is that neither organization theory nor organizational learning theory yet provide workable business models KM authors and project managers might use.

Micro economics offers a number of alternative ToFs – transactions cost theory, principal-agent theory, nexus of contracts theory, property-rights theory, and so on. It is useful to see these as the economics profession’s responses to Coase’s charge in his 1937 The Nature of the Firm that economics has no tenable ToF (Coase 1991a). Coase’s own intuitions were that the essence of the firm’s nature lay in its chosen mode of subordinating employees to managers. At first sight this sounds like naked managerial power, an echo of bureaucratic theory. But Coase was also talking about law – his insight is into a more complex situation that includes the legal apparatus behind both the firm’s labor market and its employment contracts. The latter, of course, are ‘incomplete contracts’ quite unlike the ‘spot contracts’ of equilibrium economics. Coase intuited managing through the interplay of legal ‘reality’, people’s preparedness to enter into incomplete contracts, and the exercise of firm-specific managerial power. The KM implications are to see the firm as both a bureaucratic device and a socially and historically situated legal one, the latter having KM implications corporate and labor lawyers know well, needing effective legal information systems to manage their part in the business. Some of the KM literature considers legal systems but more likely designed to support law firms and practices than the normal business firm’s corporate lawyers or human resource management departments.

In contrast to the discussion in micro economics and strategy there is little comment about transactions cost theory or principal-agent theory in the KM literature, in part because there is little attention to ToFs and their management implications; in part because micro economic theories do not strike most KM writers as enough about real firms to be worth considering. This is a mistake, obviously, but the micro economic discussion requires an unpacking that KM authors find overly challenging. It may not be obvious that micro economic theories manifest an intellectual theory-building strategy that is notably different from that used by organization theorists. The latter try to grasp the firm in toto, as an ontological entity. The micro economists are more modest, trying to theorize only an essential aspect of the firm’s nature; like the blind men, they seek part of the firm’s essence and do not pay attention to the firm as a whole. They call this ‘getting inside the black box’. Transactions cost theory begins with the firm’s ‘make or buy’ choices. Coase got to thinking about these because he spent 1932 touring the US on a scholarship, with excellent letters of introduction to senior business people, listening to what they thought important (Coase 1991b). His firm-first strategy led him to see that while business people had ideas about what their firms were, his economist colleagues were more or less in the dark, focusing on theoretical questions that grew out of nineteenth century marginalist economics and its puzzles over a firm’s ‘natural’ size. To simplify, Coase implied a firm comes into being because it can produce goods and services more cheaply than these goods and services can be acquired in markets (from other existing firms). He did not say much about how the new firm was able to turn this trick; nonetheless he saw firms required their managers to compare the costs of making against the costs of buying – revealing something essential about the knowledge-nature of firms. The KM implications are clear; the firm is a boundary that contrasts information about the firm against information about its markets.

Coase’s work was a huge step forward from both the marginalists’ production function ToF and the mechanistic ToF popularized by Max Weber and Scientific Management that paid no attention to what was going on in the firm’s markets. Coase was primed from his earlier managerial accounting studies, also known as cost accounting or estimating, and he appreciated its differences from financial accounting. Likewise Coase’s intuitions about subordination lie behind principal-agent theory. Here economists look at the costs that arise when a subordinate’s interests and knowledge differ from the principal’s (manager or owner). Some managers might address the differences by spending money on monitoring subordinates’ activity, and workplace surveillance is a rapidly expanding business. Other managers might spend money on performance incentives and there is a huge literature on various pay schemes’ strengths and weaknesses. The economic intuition is to seek the situation of greatest economic benefit, balancing the losses occasioned by interest and knowledge differences against the monitoring and incentive costs. Again the KM implications are clear, gather information about losses – such as ‘leakage’ in supermarkets – against the costs of security cameras, RFID tagging, workplace ‘snitches’, and so on. This is practical stuff about practical KM systems and much in evidence in real firms – yet typically ignored in the KM literature that provides no ‘practice-base’ to the author’s notion of knowledge. We can surmise firms that find internally generated losses their primary strategic challenge want their KM system designed around that – such as hedge funds who profit from ‘insider information’ are likewise exposed to such losses. A hedge fund’s bureaucratic aspects can probably be taken care of by off the shelf accounting software. Firms that find market cost information strategically crucial – such as high-speed stock traders – likewise know well what kind of KM system they need.

There are other micro economic theories – nexus of contracts, property rights, team production, etc. – but the principle involved is the same; each carries its own KM message, and this needs to be unpacked for a successful KM result (Spender 2013b). The academic intuition, of course, is that if all these theories could be combined, like taking reports from each of the seven blind men (micro economists), a ‘total’ theory of the firm will arise. As this happens we will develop a ‘total’ picture of the firm’s KM requirements. The practitioners’ answer is that this is a typically impractical academic notion, not one that reflects the firm’s strategic reality and the managerial judgment required to balance KM effort and return. In practice it probably makes better sense to disaggregate the ‘firm-first’ KM strategy and focus on those elements that offer the most significant immediate returns. There is strong empirical evidence that KM projects with modest partial objectives are more likely to succeed than broad firm-wide project to transform the business model; these are attractive to technology boosters, yet have a horrendous failure rate.

4 Managerial Judgment

The discussion above argues that building a successful KM system starts with understanding the specifics of the firm the system is to support. This is no trifling comment for there is no general KM systems design precisely because there is no universal ToF. While firms may share a great deal, every firm is unique in strategically important aspects and, in consequence, tricky to understand. It has a measure of ‘inimitability’. Second, it may not be worth trying to go much beyond the most strategically weighty aspect of firm and its KM implications. Managerial judgment is called for to tame the KM practitioners’ technophilia. But the history of KM is not simply of too much technology talk – Prusak’s point – rather it shows (a) the futility of trying to define ‘knowledge’, (b) the positivist tendency to fall back on defining it as computer-manipulated data, but even more importantly (c) the erosion of the 1960s firm-first ‘systems analysis’ aesthetic. Getting into the indexicalities or specifics of the existing firm surfaces the managerial judgments that shaped that firm’s nature as well as its structure, processes, market engagements, and culture. The analysis switches from seeing the firm as an object (defined by a ToF) and towards seeing the firm as an articulation of its managers’ subjective judgments. KM’s promise to get beyond the limits of management science lies through exploring and presenting management’s judgments about the data (facts) of the situation. In this way KM can provide management with strategic support rather being focused on data collection and analysis and tactical support.

The final part of this chapter unpacks managerial judgment and shows the kinds of KM system that might support the managers making judgments. Once again there is nothing new here, yet another example of how much of the history of KM turns out to be of forgetting earlier practical insights in the pursuit of the impractical and dystopian dream of computerizing everything. The promise of ‘real time modeling’ and ‘big data’ has led the KM field to forget what we always knew about what computers cannot do for us (Dreyfus 1992). There is an analogy here to the distinction between ‘hard’ and ‘soft’ approaches in AI (artificial intelligence). The history of KM shows the dream of ‘bringing people back in’ of developing a ‘soft’ approach has persistently failed to stem the technologically-driven advance of ‘hard’ approaches, even though that takes KM back into IT and abandons its distinctive promise as a discipline about the judgments of those managing firms. KM’s inability to deal with knowledge generation and the related concept of profit lurks menacingly in the background. Pre WW2 economists like Coase and Keynes were sympathetic to Frank Knight’s 1921 argument that firms that conformed to any fully determined model, such as the mechanical ToF, were incapable of generating profits (Knight 1965). In other words profit-generating firms had to have managerial judgments at their core. Conversely the fully determined mechanistic model is the antithesis of the businessperson’s ideal. If profit is the firm’s most fundamental purpose then managerial judgment takes both theoretical and practical precedence over analysis. KM’s fundamental promise is to translate this realization into appropriate practice and focus on supporting managerial judging.

Given the dis-aggregation of the firm into several separable aspects, the blind men story, a different kind of managerial judgment is called for to bring the parts together into ‘integrated reasoned practice’, deciding what to do with the elephant. Implied are two ‘dimensions’ of managerial judgment and they are ‘orthogonal’; one focuses on the effort put into modeling the knowable parts of the firm, the other on synthesizing the results into practice as reasoned as possible. Clearly if the synthesis produces a fully determined model, as a fully optimized bureaucratic model is determining, then the design of the KM system to support this follows directly. But this (a) never happens in real business, and (b) denies the strategic significance and implications of the firm as a profit seeking apparatus. This can be turned around so that the KM system intentionally presents the firm as dis-aggregated, thereby focusing management’s attention on the judgments they must make to transition from thinking about the firm’s parts and to leading it into acting as a whole.

This sounds more complicated than it is for, once again, this is old stuff made un-fashionable by the technophilic trend to ‘hard’ concepts – along with using the term ‘KMS’ to designate a highly computerized KM system. Consider the Balanced Scorecard (BS) and its connection to ‘dashboard’ or KMS approaches. Academics complain the BS lacks theory and valid metrics and does not lead to an objective performance function i.e. that the BS is not a determining model (Jensen 2002; Voelpel et al. 2006). Yet it remains highly popular with managers, so academics may be misunderstanding the BS’s real value. Likewise the history of corporate portals shows the tension between the value of displaying information users find valuable and the difficulty of generating coherent models (Benbya et al. 2004; Dias 2001). The key empirical finding here is that there are two dimensions of use value (a) information quality, such as clarity, validity, and relevance and (b) the way the KMS supports collaboration between the system’s users (Urbach et al. 2010). The value of the BS is that it portrays the strategic judgments managers must make as they ‘balance’ the financial, customer, internal, and learning dimensions of the firm in the discussion that leads managers with diverging interests and knowledge towards a shared conclusion – the firm’s strategy. The BS’s four dimensions cannot be collapsed into one; the whole point is that the data available to those within the firm are not subsumed under a single objective function. Thus the BS portrays the data relevant to each interest – say finance versus marketing and production – as best it can but then presents the collaborating managers with a question of strategic judgment, to choose the balance to be acted out in the firm’s practices. Thus the BS is a management-friendly ToF. Note its history is as a management-driven pushback against the destructive dominance of financial accounting in the firm’s strategic conversation, one of the consequences of the decline of managerial accounting (Johnson and Kaplan 1987; Kaplan 2010).

The most mentioned determinant of KM system success is ‘top management support’, another way of saying top managers’ interest in the judgments they have to make. Their focus on strategic judgment needs to be appropriately balanced against the IT designer’s enthusiasm for deeper analytics and potential objective functions. The BS is just one of a number of multi-dimensioned non-determining ToFs circulating today, any one of which can be presented in dashboard KMS format. Favorite is SWOT, though others such as Key Success Factors, the BCG Matrix, and Porter 5-forces are popular (Jarzabkowski et al. 2009). These are best seen as practitioner-driven presentations of the many-dimensioned strategic judgments managers must make; they are not, as so many academics think, failed attempts to generate a one-dimensioned ToF. They are situated models, to be selected on the basis of their judged strategic relevance. The BS tends to look inward while SWOT looks outwards as well. The BCG matrix is relevant when the strategic challenge involves judging the distribution of funds between the firm’s various lines of business. Porter’s model is relevant when the strategic challenge is to protect existing rent streams. Each provides the basis for a KM dashboard. Each is a way of operationalizing a firm-first KM project. More importantly, these models show how KM project managers might get into ‘rolling their own’, developing a business model that truly captures senior management’s intentions and intuitions.

Concluding Comments

The history of KM is as fragmented as the field itself, and given the diversity of our ways of knowing – the basis of the Greeks’ KM strategy – KM is not likely to converge into a coherent or one-dimensional positivist ‘science’. Indeed the field’s pluralism is a virtue that positivistic and scientific approaches lack; it enables talk of value-creation. Clearly most writers and practitioners see KM as a technological field, a sub-field of IT with its emphasis on data. As Prusak suggested, the history here is of the gradual development of corporate IT projects from early accounting and EDP applications into the vast complex real-time enterprise models of today. These developments have been accelerated by recent advances in data collection and analysis now driving ‘clouds’, ‘analytics’, and ‘big data’, especially in travel and retailing, but increasingly important in complex operations such as electricity grid management, financial trading and derivatives, aircraft maintenance, etc. The MOOCs show promise of bringing KM into twenty-first century education. Prusak also noted globalization and, as the global economy and global corporate operations have overtaken domestic operations, the corporation’s data collection and communication challenges have scaled up immeasurably, putting pressure on their IT system designers and operations managers. Hackers and others have opened up entirely new dimensions of KM activity and concern, the bailiwick of agencies such as USCYBERCOM and the plethora of corporate computer security units. In short, the KM field’s diversity is increasing rapidly.

Many note the term ‘knowledge’ came into wider use in the 1990s as KM was popularized as tools and techniques for finding and managing the firm’s under-managed intangible ‘knowledge assets’. The term seemed to offer KM practitioners more scope than ‘information’ because it suggested different states or kinds of knowing. The explicit-tacit and individual-organizational distinctions became popular and let KM writers reach out from economic notions of corporate asset towards others like culture, skill, and ‘best practices’. Most presumed KM’s specific task was to help managements identify the firm’s tacit knowledge, bring it to light, gather it up in KM systems, share it, and deliver it to where its value was highest. Such KM seemed to have the potential to bring both tangible and intangible assets into the processes of managing the firm’s bundle of assets as rationally as possible, reinforcing the idea of KM as a subset of both IT and accounting. Likewise KM was seen as a set of techniques for improving the utilization and sharing of these assets, a perennial management anxiety about wasting resources on ‘reinventing wheels’. One KM variation, contrary to ‘knowledge sharing’, focused on protecting these newly defined assets through better identification and use of intellectual property rights legislation, patents, copyrights, use agreements, etc. A different slant, influential in the human resource management community, was to categorize intangible assets as modes of ‘intellectual capital’ – human, structural, and relational, perhaps – widely taken up in the literature though most of the metrics and models that have emerged over the last 20 years are barely practical and lack epistemological foundations. Nonetheless the ideas have impacted political projects to extend corporate accounting for intangibles to the point of provoking regulations about supplementing the firm’s balance sheet with ‘intellectual capital statements’.

These various projects and advances are best seen as knowledge-oriented applications within the broader class of IT applications now extended to include natural language processing, semantic web management, data mining, knowledge repositories, corporate intranets, and so on. The background assumption is that KM can flesh out a field of rational analysis within which all types of knowledge eventually sum into the firm’s ‘total’ knowledge capital and expose it to rational management. In due course the early enthusiasm around these KM/IT applications waned as they collided with the reality of poor metrics, high KM project failure rate, low user satisfaction, and a deepening appreciation of the limits to technology’s impact on the firm’s practices and bottom line. Even though some have been arguing for the last couple of decades that the KM party is over the KM industry still offers many jobs, conferences, excitements, and practitioners’ journals (Rao 2005).

The history of KM reads very differently for those sensing a distinctive agenda outside IT. Prusak’s third notion of the knowledge-centric firm is indicative though he did no more than suggest its history by pointing to the ideas of Arrow, Hayek, Machlup, and Coase. He did not elaborate. Yet these economists, along with philosophers like Ryle and Polanyi, were twentieth century torchbearers of a widespread academic (and political) movement deeply critical of nineteenth century marginalism, equilibrium analysis, and rational man theorizing. These authors’ epistemological sophistication, contrasting with their positivistic colleagues’ lack, converged with an effort to rethink the notion of profit and answer Coase’s questions about the nature of the firm, clarified by Knight’s linking profit and value-creation to the managerial judgment required to deal with economic uncertainty. Making something non-tautological and workable out of these ideas requires careful attention to the problems around the term ‘knowledge’. As the vast majority of KM authors and the companion industry of ‘KM practitioners’ busied themselves with new IT possibilities, a minority of writers struggled with the differences between types of knowledge and the implications for a radically different kind of KM – KM+ perhaps. The result would be theories of the firm that stand apart – epistemologically – from the mechanistic models presumed by KM authors and system designers of positivist disposition. This group’s work continues though the KM mainstream, especially the KM consultants, pays it little attention.

The story of KM’s origins, history, and development would be pretty bleak if there was no more to be said. The KM that succeeds is simply IT re-labeled, KM/IT. The contrasting promise of KM+ as a distinctive academic field engages epistemology’s profoundest questions. This chapter suggests a third history – of chances missed and earlier insights abandoned under the cultural and institutional pressure to institutionalize rational choice thinking. We have missed how real world managers have been meeting the wickedest KM challenges all along – for they had to if their firms were to survive. Understanding these challenges begins with admitting bounded rationality, the irrelevance of rational choice thinking to understanding life’s practice, especially within real firms. The bad news is that bounded rationality means the KM/IT project can never succeed to the point of making managers’ work unnecessary, for firms comprise people as well as knowledge, however this is defined, and people and computers do not share compatible views of the world. Paradoxically the ‘hard’ KM program is ultimately irrelevant to using KM/IT to best effect. But the good news, which the KM+ field knows but the KM/IT authors ignore, is that profit itself arises from the uncertainties that bounded rationality generates. The judgment managers then contribute is the font of all profit, yet managerial judgment has lain outside the KM discourse since it arose in the 1990s. While many use the term ‘judgment’ it is seldom clarified, distinguished from rational decision-making, or its managerial and epistemological implications explored (Foss and Klein 2012; Spender 1989, 2014). Judgment has also largely departed from business education since the 1970s. Prior to that time ‘systems analysts’ would base their projects on the firm’s current practices which, of course, instantiated or indexed management’s judgments about how to engage the specific firm’s uncertainties. KM’s missed chances arose as its new group of authors engaged in the unwinnable battles with epistemology that distracted them away from the ‘firm-first’ practice of earlier times.

Today KM has the opportunity to re-energize and re-shape its history by drawing in the work of a wide group of authors who already see firms as multi-dimensional complexes of activity, less of epistemology-defined knowledge types but of the firm-defined contrasting practices (Spender 1995). There are many such models in use today as strategy tools and each can be the basis of a KM dashboard (Spender 2014). The result would significantly advance the ‘soft’ KM program, using its techniques to support management, respecting the value of their judgments, rather than a disrespectful ‘hard’ program to ‘out-model’ them with computer-based algorithms.