Part A: Knowledge management
Introduction to knowledge management
Key takeaways
- Knowledge management is the deliberate and systematic coordination of an organisation's people, technology, processes and structure to add value through reuse and innovation (Dalkir's definition).
- Data are raw facts, information is data placed in context, and knowledge is information combined with experience, interpretation and judgement so that it allows action.
- Tacit knowledge is hard to articulate and moves mainly by demonstration, apprenticeship and shared experience; explicit knowledge is recorded and can be copied at almost no cost.
- Codification stores knowledge for reuse and suits repeatable work; personalisation connects people and suits novel work. Successful organisations pursue one predominantly, roughly eighty per cent, with the other in support.
- Clause 7.1.6 of KS ISO 9001 requires the organisation to determine the knowledge its processes need, maintain it and make it available, so knowledge management is a requirement of the quality standard.
Every organisation already manages knowledge; most do it by accident, and discover how badly when an experienced person leaves. This paper is about doing it deliberately, and the chapter begins with the definitions the rest of the paper relies on.
Introduction
The knowledge an organisation depends on is largely invisible. It sits in people's judgement, in relationships, in undocumented workarounds and in the accumulated sense of how things are done here. It does not appear in the accounts, it is not counted in the asset register, and it leaves the building every evening. Knowledge management is the discipline that treats it as an asset to be developed, captured, shared and applied rather than as a by-product of employment.
What knowledge management is
Knowledge management is not information management
Information management handles documents, records and data: their capture, storage, retrieval and disposal, which is the subject of Documentation and Records Management for Quality. Knowledge management is concerned with what people know and how that is developed, shared and applied. A system that stores every document an organisation ever produced and captures nothing of why decisions were taken is an information management success and a knowledge management failure.
A multidisciplinary field
- Organisational science and management theory: how structure, culture and incentives affect what people share
- Cognitive science and psychology: how individuals learn, remember, and construct understanding
- Sociology and anthropology: how communities form, how norms govern sharing, and how identity affects what people will disclose
- Library and information science: classification, taxonomy, metadata, retrieval, which supply the machinery of codification
- Computer science and artificial intelligence: repositories, search, expert systems, machine learning
- Communication studies: how meaning is transmitted and where it is lost (see Communication and Customer Relations)
- Economics: the economics of information goods, which are costly to create and nearly free to copy, and the resulting incentive problems
- Human resource management: recruitment, competence development, retention and succession
- Education and instructional design: how learning is structured and how expertise is developed
- Philosophy, specifically epistemology: what counts as knowledge at all, which is not a merely academic question when an organisation must decide whose judgement to record
- The practical consequence of the multidisciplinarity is that no single function can own knowledge management: an initiative located wholly in the ICT department produces a repository nobody uses, and one located wholly in human resources produces training with nowhere for its output to go
Types of knowledge management
Ways of classifying knowledge management
| Basis | Types | Distinguishing question |
|---|---|---|
| By knowledge form | Tacit and explicit | Can it be written down? |
| By strategy | Codification and personalisation | Is the aim to reuse knowledge or to connect people? |
| By level | Individual, group or community, organisational, inter-organisational | Whose knowledge is being managed? |
| By process | Creation, capture, sharing, application | Which part of the cycle is addressed? |
| By orientation | Technology-driven and people-driven | Is the primary investment in systems or in relationships? |
| By source | Internal knowledge and externally acquired knowledge | Was it developed here or brought in? |
Concept analysis
Concept analysis is a technique for establishing what a term actually means within a field, by examining its attributes, its cases and its boundaries. It is included in the syllabus because knowledge management is a field whose central terms are contested, and an organisation that has not settled what it means by knowledge will not agree on what to manage.
- Select the concept and state the purpose of the analysis
- Identify all uses of the concept: how the term is used in the literature and in the organisation, including uses that conflict
- Determine the defining attributes: the characteristics that appear repeatedly and without which the concept would be something else
- Construct a model case: a clear, unambiguous example possessing all the defining attributes
- Construct additional cases: a borderline case possessing some attributes, a related case that is similar but distinct, a contrary case possessing none, and where useful an invented or illegitimate case
- Identify antecedents, the conditions that must exist before the concept occurs, and consequences, what follows from it
- Define empirical referents: how the concept would be recognised or measured in practice
Applied to knowledge, the analysis yields defining attributes such as being held by a knower, being derived from information combined with experience, and enabling action. A model case is an experienced maintenance technician diagnosing a fault by sound. A related case is information, which lacks the experience attribute. A contrary case is raw data. The empirical referent is the ability to act correctly in a situation not previously encountered, which is why competence assessment is the practical measure of knowledge.
A short history
The development of knowledge management
| Period | Development | Significance |
|---|---|---|
| Pre-1990s | Apprenticeship, guilds, professions, corporate libraries, expert systems in the 1980s | Knowledge transfer existed as practice without a name; expert systems attempted to codify expertise and largely failed on tacit knowledge |
| Early 1990s | The term knowledge management enters management discourse; Nonaka and Takeuchi publish The Knowledge-Creating Company (1995) | Knowledge is recognised as the principal source of competitive advantage; the tacit-explicit distinction becomes central |
| Mid to late 1990s | First wave: technology-led. Intranets, document repositories, lessons-learned databases, chief knowledge officers appointed | Widespread disappointment: repositories were built and not used, because contributing had cost and no benefit for the contributor |
| Early 2000s | Second wave: people-led. Communities of practice, social capital, storytelling, after-action reviews, culture and incentives addressed | The recognition that knowledge management is a social problem with a technical component rather than the reverse |
| 2000s to 2010s | Third wave: integration. Knowledge management embedded in business processes rather than run as a separate programme; Web 2.0, wikis, blogs and social tools | Contribution becomes a by-product of work rather than an additional task |
| 2010s to date | Analytics, enterprise social networks, mobile and cloud, and artificial intelligence applied to capture, search and summarisation | New capability and new risks: automated capture at scale, and models that produce fluent output of uncertain reliability |
From physical to knowledge assets
- The proportion of firm value represented by tangible assets has fallen sharply over recent decades, while the proportion attributable to intangibles, brand, capability, data, relationships and know-how, has risen
- Knowledge assets behave differently from physical assets in ways that matter economically: they are non-rival, so use by one person does not deplete them; they are costly to create and nearly free to copy; they can increase in value with use, since a shared method improves as more people apply it; and they depreciate through obsolescence rather than through wear
- They are also uniquely fragile: they walk out of the building, they can be lost entirely with one departure, and they cannot be repossessed
- Accounting recognises them poorly: internally generated knowledge assets are generally not capitalised, so the balance sheet of a knowledge-intensive firm understates it substantially, which is why market capitalisation and book value diverge
- The management consequence is that investment in knowledge assets competes for budget against investment in physical assets and loses, because the physical asset appears on the balance sheet and the knowledge asset does not
- For a Kenyan organisation the practical form of this is turnover: a technical specialist leaving takes an asset whose replacement cost is real and whose loss is recorded nowhere
Why KM matters today
- Workforce mobility: people change employers far more frequently than a generation ago, so knowledge that resides only in individuals is knowledge in transit
- Demographic change: experienced staff retire, and in organisations with a compressed hiring history a large proportion may leave within a short period
- Complexity: work increasingly requires knowledge no single person holds, so coordination of distributed expertise becomes the constraint
- Speed: the value of knowledge decays faster, so the time between learning something and applying it determines whether it is worth anything
- Scale and dispersion: multi-site and remote working removes the corridor conversations through which knowledge previously moved
- Competition on capability rather than on assets: where competitors can buy the same equipment, the difference is what the organisation knows how to do with it
- Repetition of errors: an organisation that does not capture lessons pays for the same mistake in each site, each project and each year
- Regulation and accountability: reconstructing why a decision was taken years later requires that the reasoning was recorded, not only the decision
- Quality management: clause 7.1.6 of KS ISO 9001 requires the organisation to determine the knowledge necessary for the operation of its processes, to maintain it, and to make it available to the extent necessary, so knowledge management is a requirement of the quality standard and not merely good practice
Clause 7.1.6 is why this paper is in a quality qualification
The 2015 edition of ISO 9001 introduced organisational knowledge as a resource requirement for the first time. The organisation must determine the knowledge its processes need, maintain it, make it available, and when addressing changing needs and trends, consider its current knowledge and determine how to acquire what is additionally required. That is a knowledge management obligation stated in an auditable standard, and it is examinable in both this paper and Quality Management Systems.
Individuals, communities and organizations
Knowledge management at four levels
| Level | What knowledge management does | Typical mechanisms | Benefit |
|---|---|---|---|
| Individual | Helps a person do their job better and develop | Personal knowledge management, mentoring, access to expertise and repositories, reflective practice | Faster problem solving; visible expertise; career development |
| Community or group | Connects people with common practice across organisational boundaries | Communities of practice, networks, forums, after-action reviews | Problems solved once rather than repeatedly; norms and standards of practice developed |
| Organisation | Builds and retains capability, and applies it to strategy | Knowledge strategy, repositories, taxonomies, lessons-learned processes, succession planning | Reduced dependence on individuals; faster response; competitive capability |
| Inter-organisational | Shares knowledge across a value chain or sector | Supplier development, joint improvement, industry associations, standards committees | Sector capability; reduced duplicated learning |
The levels interact, and the interaction is where most knowledge management fails. An individual will not contribute to an organisational repository unless doing so serves them or their community; a community cannot form unless the organisation permits the time; and organisational capability cannot be built without both. Designing at one level alone is the commonest reason initiatives produce systems nobody uses.
What knowledge walking out of the door actually costs
A Kenyan cement plant loses its Chief Process Engineer, 22 years'
service, to a competitor. The role is filled in 4 months.
WHAT THE ACCOUNTS RECORD
Recruitment cost KES 1,900,000
Salary during handover KES 780,000
Total recorded KES 2,680,000
WHAT ACTUALLY LEFT, and none of it appears anywhere
1. Kiln stability judgement. The departing engineer could hold
the kiln stable through raw-material variation by adjusting
three parameters together in a way no procedure describes.
After departure, kiln stops rose from 4 to 11 a quarter.
Each unplanned stop costs 9 hours of production at
130 t/hour and a contribution of KES 2,900 per tonne:
9 x 130 x 2,900 = KES 3,393,000 per stop
7 additional stops a quarter = 7 x 3,393,000
= KES 23,751,000 a quarter
= KES 95,004,000 a year
2. Supplier relationships. Knew which of four limestone
suppliers could be trusted on moisture declaration without
retesting. Successor retests everything: 340 additional
tests a year at KES 4,200 = KES 1,428,000.
3. Regulatory history. Knew the reasoning behind three NEMA
licence conditions negotiated in 2016 and why the plant's
effluent design is as it is. Reconstructing it for the
licence renewal cost 6 weeks of consultancy: KES 2,100,000.
4. Undocumented workarounds. Roughly 40 small pieces of
know-how, individually trivial, collectively the difference
between the plant running and the plant being fixed.
Estimated first-year loss beyond the recorded cost:
95,004,000 + 1,428,000 + 2,100,000 = KES 98,532,000
Recorded cost as a share of actual loss:
2,680,000 / (2,680,000 + 98,532,000) = 2.6%
WHAT KNOWLEDGE MANAGEMENT WOULD HAVE COST
Structured knowledge capture in the final 3 months: exit
interviews focused on decisions rather than duties, shadowing
by the successor, recorded problem-solving sessions, and
documentation of the kiln parameter relationships.
40 hours of the departing engineer's time KES 340,000
Facilitation and documentation KES 260,000
Successor overlap, 2 weeks KES 190,000
Total = KES 790,000
Even recovering a quarter of the loss would repay it 30 times.
The reason it is not done is that the KES 790,000 is a visible
cost in this year's budget and the KES 98 million is invisible
in every year's.
Quantify the loss before proposing the system
Every knowledge management question is answered better by starting with what the absence of it costs this organisation. Departures, repeated errors across sites, time spent searching for information, and problems solved twice are all measurable, and the measurement is what obtains the resource that the initiative needs.
Self-check
Distinguish data, information and knowledge, and distinguish tacit from explicit knowledge, giving a Kenyan manufacturing example of each. Explain why the tacit-explicit distinction is the central problem of knowledge management. (10 marks)
Explain the codification and personalisation strategies for knowledge management, and advise which a Kenyan engineering consultancy and a Kenyan call centre should each adopt, with reasons. (10 marks)
Show worked answer
Knowledge concepts (10 marks: 3 for the hierarchy, 3 for tacit and explicit, 4 for why it matters).
Data are raw facts and figures without context: a sequence of fill volumes, 496, 502, 488, recorded from a bottling line. They carry no meaning until placed in context.
Information is data processed and placed in context so that it answers a question: the mean fill of the last twenty bottles was 495.3 ml against a target of 500 ml.
Knowledge is information combined with experience, interpretation and judgement so that it enables action: the filler is drifting low in the way it did before the last bearing failure, so the bearing should be inspected before the shift ends. Wisdom, sometimes added as a fourth level, is the judgement to apply knowledge appropriately, including knowing when not to act.
Explicit knowledge is knowledge that has been articulated and recorded: the bottling line's operating procedure, the fill specification, the maintenance manual, the control chart. It can be stored, copied and transmitted at almost no cost.
Tacit knowledge is held in a person's head and is difficult to articulate: the maintenance technician who identifies a failing bearing by the sound the machine makes, the operator who can tell from the feel of the product that the mix is wrong, the supervisor who knows which of four suppliers can be trusted without retesting. Polanyi's formulation is that we know more than we can tell.
Why the distinction is the central problem: explicit knowledge is comparatively easy to manage, because it can be stored, indexed and retrieved by systems, and the technical problem of doing so is largely solved. But the great majority of the knowledge that determines organisational performance is tacit: judgement, skill, pattern recognition and relationships. It cannot simply be written down, because the holder is frequently unable to articulate it and often unaware they possess it. It is transferred principally through demonstration, apprenticeship, shared experience and conversation, all of which require time, proximity and willingness. It is also the knowledge that walks out of the door when a person leaves, and its loss is invisible in the accounts. Every serious difficulty in knowledge management, why repositories go unused, why exit interviews capture duties rather than judgement, why an experienced operator's departure raises defect rates, follows from the tacit nature of what actually matters. An initiative designed as though all valuable knowledge were explicit will build a document store and change nothing.
Codification and personalisation (10 marks: 4 for the strategies, 6 for the advice).
The codification strategy extracts knowledge from the person, makes it explicit, and stores it in a repository so that it can be reused many times by many people without further contact with its originator. It invests in documents, databases, taxonomies and search. Its economics rest on reuse: the cost of codifying is recovered because the same knowledge answers the same recurring problem many times.
The personalisation strategy leaves knowledge with the person and invests instead in connecting whoever needs it to whoever has it, through expertise directories, networks, communities of practice, mentoring and deliberate proximity. It invests in relationships and in the time to have conversations. Its economics rest on the value of a bespoke answer to a novel problem.
Hansen, Nohria and Tierney's finding is that successful organisations pursue one strategy predominantly and use the other in support, in roughly an eighty to twenty ratio, and that organisations pursuing both equally succeed at neither, because the two demand different systems, different incentives and different kinds of people.
Advice for the engineering consultancy: personalisation. Each client project differs in site, structure, ground conditions, client requirement and regulatory context, so the problems are novel rather than recurring and a repository of previous solutions would be consulted for reference and would not answer the question. What the consultancy needs is that the engineer facing a difficult foundation problem can identify and reach the colleague who solved a comparable one, which means an expertise directory, deliberate mixing across disciplines and projects, mentoring of junior engineers by senior ones, and time allowed for it. Codification supports this: standard details, a design checklist and a lessons-learned register for recurring pitfalls, but it is not the primary investment.
Advice for the call centre: codification. The work is high volume and highly repetitive, the same customer questions recur thousands of times, and the value of an answer lies in its being correct and consistent rather than bespoke. A well-maintained, searchable knowledge base that agents consult during the call, with defined ownership, review cycles and a route for agents to propose corrections, is exactly the right investment, and its cost is recovered across an enormous number of reuses. It also shortens the training period for new agents, which matters in a role with high turnover. Personalisation supports this: escalation to a named specialist for the small proportion of genuinely novel queries, and team debriefs on emerging issues so that the knowledge base is updated from what agents encounter.