PKMC Module IV cheat sheet

Explicit knowledge is codified and transferable (documents, models, registers); tacit knowledge is personal, experience-based, and hard to articulate - the two require different capture methods.

The DIKW hierarchy runs Data (raw symbols) → Information (data in context) → Knowledge (information plus experience and interpretation) → Wisdom (judgment about application).

Knowledge management in projects covers four linked processes: capture, storage/retrieval, sharing/transfer, and application/reuse; retention and archiving close the loop.

PMBOK distinguishes project information (documents, records, reports) from project knowledge (insights, expertise, lessons) - information management is a subset of knowledge management.

The knowledge life cycle: create → validate → codify → store → share → apply → refresh or retire; value is realized only at the apply stage.

Information quality dimensions to test at handover: accuracy, completeness, timeliness, consistency, relevance, and traceability to source.

Knowledge loss risk peaks at project closure and at staff turnover; identify critical knowledge and single-points-of-knowledge early in planning.

Knowledge management effort should scale with project complexity, novelty, and rate of team change - not a fixed overhead applied uniformly.

Interviews suit tacit, expert-held knowledge; structured interviews use a fixed question set, unstructured interviews follow emerging themes, semi-structured blend both.

Focus groups surface shared and conflicting views among 6-10 participants; individual interviews are better when hierarchy or politics would suppress candor.

Observation (job shadowing) captures tacit know-how that people cannot articulate; it is slower than interviews but reveals workarounds and implicit rules.

The Delphi technique uses anonymous, iterative rounds with controlled feedback to converge expert opinion; it removes dominance effects but takes multiple cycles.

Nominal group technique: silent idea generation, round-robin recording, then ranked voting - reduces groupthink compared with open discussion.

Brainstorming and affinity diagramming generate then cluster ideas; mind mapping and concept maps show relationships among knowledge items.

Document analysis mines existing records, logs, and correspondence for knowledge not otherwise accessible; pair it with validation interviews.

Retrospectives (agile) and after-action reviews (AAR) capture knowledge close to the event while memory is fresh; the four AAR questions are: what was supposed to happen, what happened, why the difference, what to improve.

Capture quality checks: triangulate at least two sources for critical knowledge, record the source and date, and have the owner validate the record before storage.

A taxonomy is a hierarchical classification; a folksonomy is user-generated tagging - combine a controlled vocabulary with free tags to balance consistency and findability.

Metadata minimum for a knowledge asset: title, owner, date created, version, source project, classification/sensitivity, and review or expiry date.

Retrieval performance is measured by precision (relevant items returned ÷ items returned) and recall (relevant items returned ÷ all relevant items); both degrade as the repository grows without curation.

Single source of truth principle: each knowledge asset has one authoritative location; copies are links or references, not duplicates.

Information architecture layers: repository (storage), taxonomy and metadata (organization), search and navigation (retrieval), permissions (access).

Version control and change history preserve auditability; superseded documents are marked, not deleted, to keep the decision trail intact.

Naming conventions and folder depth rules reduce retrieval failure; deep nesting beyond 3-4 levels measurably slows search and increases misfiling.

Search effectiveness depends on consistent indexing; unindexed or unlabeled repositories become data swamps where content exists but is never found.

Access control should follow least privilege and need-to-know; confidentiality classification drives who can store, view, and export knowledge assets.

Knowledge transfer mechanisms split into push (reports, portals, broadcasts) and pull (search, communities of practice, ask-an-expert); effective programs use both.

The SECI model: Socialization (tacit→tacit), Externalization (tacit→explicit), Combination (explicit→explicit), Internalization (explicit→tacit).

Communities of practice are voluntary, cross-project groups bound by shared domain interest - distinct from formal project teams and from temporary task forces.

Knowledge brokers and boundary spanners translate between disciplines; they reduce transfer loss across organizational and technical boundaries.

Transfer requires absorptive capacity: the receiver must have prior related knowledge and time to internalize - sending a document is not transfer.

Co-location and rich media (face-to-face, video) suit complex, ambiguous knowledge; lean media (email, documents) suit simple, codified knowledge - a media-richness fit decision.

Collaboration platforms succeed on adoption, not features; measure active contribution rate and reuse rate rather than page views.

Mentoring, job rotation, and paired working transfer tacit knowledge; documentation and wikis transfer explicit knowledge - match method to knowledge type.

Cross-cultural and cross-generational transfer needs explicit context: idioms, assumed norms, and undocumented conventions must be surfaced deliberately.

Reuse starts with a search of prior lessons, templates, and estimates during initiation and planning - before solutions are designed, not after.

Reuse candidates should be assessed for applicability: context similarity (scale, sector, technology, constraints) determines whether a lesson transfers.

Templates, checklists, estimating databases, and standard work packages are the highest-yield reusable assets because they embed lessons directly into workflow.

Knowledge application is the only stage that produces value; capture and storage without retrieval and use yield no return.

Decision logs and assumption logs preserve rationale so later teams can reuse the reasoning, not just the outcome.

Adaptation cost is real: reused assets must be tailored and revalidated; blind copy-paste of lessons across contexts is a known failure mode.

Measure reuse with indicators such as percentage of templates applied, number of prior-lesson references in plans, and reduction in repeat defects.

Lessons should be written as actionable rules (trigger condition + recommended action), not narrative history, to make them reusable.

The lessons-learned process is continuous: capture during delivery, validate at phase gates, consolidate at closure - not a single end-of-project meeting.

A lessons-learned register fields: ID, date, phase, category, description, root cause, impact, recommendation, owner, status, and reuse reference.

Root-cause analysis tools for lessons: 5 Whys, fishbone (Ishikawa) diagram, and fault tree analysis - separate symptom from cause before recording.

Archiving decisions define retention period, legal/regulatory hold, storage format, access rights, and destruction date; these are set by policy, not by the project manager alone.

Records management requires immutable final versions, audit trail, and readable formats for the retention period (avoid obsolete file formats).

Knowledge retention risk assessment: map critical knowledge to holders; a role held by one person without a documented backup is a single point of failure.

Handover packages should include decisions, assumptions, open risks, contacts, and access credentials - not just deliverables.

Post-implementation review compares actual outcomes to the business case and baseline, and feeds benefits realization and future estimates.

Destruction must be authorized and logged; retaining records past the legal retention period creates avoidable legal and storage exposure.

Governance defines decision rights for knowledge: who approves taxonomy changes, who owns each asset, who authorizes access and retention.

Typical roles: project manager (accountable for project knowledge), knowledge manager (process and repository), knowledge owner (content accuracy), contributors, and consumers.

A RACI applied to knowledge activities clarifies accountability for capture, review, publishing, and retirement of assets.

Policies needed: classification scheme, retention schedule, access control, intellectual property and confidentiality rules, and records management standard.

Culture indicators: psychological safety to admit mistakes, leadership modeling of reuse, time allocated for capture, and recognition for contribution rather than only delivery.

Blame cultures suppress lessons-learned honesty; lessons must be separated from performance appraisal to be accurate.

Knowledge metrics should be outcome-based (repeat-error rate, reuse rate, time-to-find, onboarding time) rather than activity counts like documents uploaded.

Incentives misaligned with knowledge sharing - rewarding only individual output - predict repository stagnation regardless of tooling quality.

Governance maturity progresses from ad hoc capture, to defined process, to measured and continuously improved knowledge management embedded in delivery.