00-taxonomy-selection.md

Taxonomy and selection protocol

Contents

1. Classification axes
2. Method-card schema
3. Context triage
4. Scoring rubric
5. Catalog ownership and specialization
6. Methodological-density gate
7. Saturation protocol
8. Research record

Classification axes

Classify every candidate independently on five axes. Do not collapse them into one label.

Axis	Values
Object type	discipline/theory; management system; lifecycle/methodology; framework/reference model; method/technique; language/notation/schema/protocol; pattern/tactic; artifact/template; test/assurance; principle/heuristic; metric; composite/local
Primary domain	strategy/organization/product; portfolio/project/process/operations; requirements/software/systems; testing/reliability/safety/security; data/knowledge/research/evidence; decision/optimization/causality; AI/ML/agents; finance/investment/market/audit/procurement; marketing/growth/brand/commercial activation; design/experience/accessibility; futures/foresight; behavior/implementation/adoption; negotiation/mediation/conflict; participation/facilitation/deliberation; people/workforce/talent; sales/revenue/customer success; communication/reputation/content operations; innovation/R&D/technology management; journalism/verification/editorial integrity; legal/regulatory/policy/ethics/crisis; physical engineering/manufacturing/assets/human factors/health/construction; environment/circularity/energy/process engineering; supply chain/logistics/geospatial; agriculture/biological systems; learning/education; meta-engineering
Lifecycle role	explore; frame; decide; design; plan; build; verify; deploy; operate; improve; govern; retire
Stack role	backbone; specialist; representation; assurance; measurement
Maturity/source	normative standard; official framework; established practice; research method; emerging practice; proprietary method; local/composite

Secondary tags may express problem shape, artifact, sector, criticality, and scale.

For domain methods, also record jurisdiction, geography, time horizon, technology or evidence maturity, and the authoritative regulator/owner when any of these can change validity. Do not generalize a sector standard, tax/legal rule, environmental factor, clinical requirement, or spatial model outside the population and context for which it was verified.

Method-card schema

Use this schema when adding or deeply researching a method:

### Canonical name — acronym

- Type / domain / lifecycle role / stack role:
- Purpose and output:
- Use when:
- Do not use when:
- Preconditions and required capability:
- Typical procedure:
- Complements:
- Alternatives or variants:
- Failure modes and gaming risks:
- Adoption cost: low | medium | high
- Maturity: normative | established | emerging | local
- Canonical research anchors:
- Current version/status checked on:
- Evidence and unresolved questions:

For a discovery catalog, the compact fields type, purpose, use, avoid, complements, and source anchor are sufficient. Expand only shortlisted methods.

Write the boundaries as decision rules, not slogans:

• Use when should name a context signal, the minimum precondition, and the output the method will produce.
• Do not use when should name a failed assumption, an incompatible output, an authority/validity boundary, or a safer/appropriate alternative.
• State a disadvantage elsewhere only when it does not determine method fit. Do not turn every cost into a prohibition.

Context triage

Problem topology

Signal	Bias the search toward	Be cautious with
Goal and causal path are unclear	discovery, systems mapping, theory of change, experiments	detailed execution workflows
Solution is known but performance is poor	DMAIC, PDCA, SPC, process mining, TOC	broad ideation as the backbone
Requirements are volatile	iterative discovery, prototypes, dual-track, evolutionary architecture	frozen exhaustive specifications
Interfaces are stable and regulated	requirements traceability, V-model, contracts, conformance tests	informal stories as sole specification
Stateful or distributed effects exist	state machines, temporal properties, idempotency, sagas, model checking	happy-path sequence diagrams alone
Human adoption is the main risk	change, stakeholder, sociotechnical, service design methods	purely technical optimization
Evidence is contested or incomplete	systematic search, triangulation, provenance, Bayesian/ACH reasoning	single-source narratives
Consequences are severe or irreversible	hazard analysis, independent V&V, assurance cases, formal methods	“move fast” defaults
Decisions must remain adaptive	OODA, rolling-wave planning, real options, monitoring triggers	one-time deterministic plans

For physical, financial, regulatory, environmental, agricultural, spatial, health, or educational work, also extract as applicable:

• scale and maturity: concept, laboratory, pilot, demonstration, production, or operating system;
• geography, jurisdiction, target population, season, observation period, and decision horizon;
• baseline/counterfactual, measurement system, estimate class, and uncertainty source;
• competent authority, professional capability, permit, certification, or independent review required.

Proportionality tiers

Tier	Context	Expected stack
1	reversible, low impact, short-lived	one lightweight backbone or technique plus basic validation
2	meaningful cost or several teams	backbone, one or two specialists, explicit acceptance criteria, outcome metrics
3	material financial/operational impact	traceability, risk analysis, independent test/gate, monitoring and rollback
4	safety, rights, regulation, high irreversibility	normative requirements, independent assurance, formal or model-based analysis where suitable, audit evidence, controlled change

Scoring rubric

Score each surviving candidate from 0 to 4. State weights before seeing totals.

Criterion	0	2	4
Outcome fit	unrelated output	partially supports decision	directly produces required capability
Context fit	assumptions conflict	tailoring required	assumptions match
Verifiability	subjective only	mixed	machine- or independently verifiable
Evidence/maturity	unsupported/obsolete	plausible or limited	current authoritative basis plus implementation evidence
Composability	conflicts/duplicates	manageable overlap	cleanly fills a distinct role
Feasibility	unavailable inputs/skills	attainable with effort	usable now
Proportionality	overhead dominates	balanced	strong risk reduction per unit cost

Do not average away hard constraints. Mark mandatory compliance, safety, privacy, and legal requirements as pass/fail gates.

When uncertainty in weights can reverse the winner, show the reversal threshold or use an unweighted ordinal comparison.

Score only genuine alternatives for the same stack role and required output. Resolve backbone, specialist, representation, assurance, and measurement needs separately. A mandatory gate is pass/fail and cannot lose to a higher-scoring optional technique.

Catalog ownership and specialization

When the same family appears in multiple catalogs, retrieve its generic definition from the canonical owner and use the domain card only for tailoring:

Family	Canonical owner	Domain specialization examples
Project, programme, portfolio, PMO, project controls, benefits, and recovery	30-project-programme-portfolio-management.md	investment economics in 12; engineering lifecycle in 04/14; adoption in 22; operations transition in 03
Process improvement, SPC, generic operations control	03-process-operations-quality.md	manufacturing in 14; agriculture in 17
Reliability, hazard analysis, generic FMEA/FMECA	06-testing-reliability-safety-security.md	DFMEA/PFMEA in 14; biological hazards in 17
Data governance, semantics, quality, lineage, provenance, and knowledge management	07-data-information-knowledge.md	AI data in 09; evidence infrastructure in 31; BIM information in 36
Research design, evidence synthesis, investigation, claims, and epistemic assurance	31-research-evidence-investigation.md	causal identification in 08; journalism in 29; legal research in 13
Experiments, causal inference, optimization, uncertainty	08-decision-optimization-causality.md	field designs in 17; learning evaluation in 18; process DOE in 15
Probability, statistics, sampling, estimation, prediction, and generic statistical inference	37-probability-statistics-inference.md	causal identification in 08; evidence synthesis in 31; spatial methods in 33; process control in 03; clinical/device evidence in 35
Financial appraisal, DCF, BoE, investment gates	12-finance-market-audit-procurement.md	TEA inputs in 15; manufacturing cost/readiness evidence in 14
Market research, sizing, conjoint, WTP, commercial diligence	12-finance-market-audit-procurement.md	activation/positioning in 19; research interaction/prototypes in 20
Marketing strategy, brand, GTM, channels, lifecycle, media measurement	19-marketing-growth-brand-commercial.md	market/economics in 12; communication/experience in 20
Design research, UX, service/interaction/information design, accessibility, design systems	20-design-experience-communication.md	product strategy in 02; safety-oriented HFE in 35; learning design in 18
Legal applicability, licensing, policy, compliance	13-legal-policy-ethics.md	sector-specific environmental, health, agricultural, or construction gates
Crisis, continuity, emergency/incident management, resilience, and recovery	32-crisis-continuity-emergency-management.md	crisis communication in 27; cyber incident response in 06; service incidents in 03
Product realization and manufacturing readiness	14-physical-engineering-manufacturing.md	process scale-up in 15
Asset management, maintenance, inspection, condition monitoring, spares, and renewal	34-asset-maintenance-reliability.md	manufacturing equipment in 14; built-asset handover in 36
Human factors, ergonomics, medical devices, and clinical-device evidence	35-human-factors-health-medical-devices.md	UX in 20; product realization in 14; health regulation in 13
Hierarchical Task Analysis (HTA), Cognitive Task Analysis (CTA), and expert task/decision elicitation	35-human-factors-health-medical-devices.md	training objectives/scenarios in 18; interface/information requirements in 20; workforce application in 25
SHERPA: task-based human-error prediction and recovery	35-human-factors-health-medical-devices.md	broader hazard analysis and assurance in 06; interaction redesign in 20
GOMS family and Keystroke-Level Model (KLM): skilled interaction modeling	20-design-experience-communication.md	HTA task structure in 35; measured usability/performance evidence remains distinct
Functional Resonance Analysis Method (FRAM): performance variability and functional couplings	06-testing-reliability-safety-security.md	work-as-done evidence in 35; process interventions in 03
Construction, infrastructure, BIM, constructability, and commissioning	36-construction-infrastructure-bim.md	project controls in 30; asset operations in 34; physical engineering in 14
Urban, territorial, regional, land-use, mobility, public-space, and place-based planning	38-urban-territorial-planning.md	geospatial data/models in 33; construction/BIM in 36; participation in 24; climate resilience in 15/21/32
LCA, environmental claims, TEA, industrial processes	15-sustainability-circular-energy.md	agricultural sustainability in 17
Supply chains, inventory, logistics, reverse networks, and traceability	16-supply-chain-logistics.md	circular feedstock/recovery in 15; spatial network design in 33
Geospatial data, spatial analysis/econometrics, remote sensing, and location decisions	33-geospatial-location-analysis.md	route/facility operations in 16; causal inference in 08; investigations in 31
Strategic foresight, scenarios, backcasting, adaptive pathways	21-futures-foresight.md	strategy in 02; technology roadmaps in 28; robust decisions in 08
Behavior-change and implementation/adoption	22-behavior-change-implementation.md	organizational change in 02; learning in 18; marketing activation in 19
Negotiation, mediation, agreements, dispute systems	23-negotiation-mediation-conflict.md	grievance/workforce in 25; legal procedure in 13
Participation, facilitation, deliberation, collective decisions	24-participation-facilitation-collective-decisions.md	research participation in 31; co-design in 20; public policy in 13
Workforce, selection, performance, talent, people analytics	25-people-workforce-talent.md	organization/change in 02; learning in 18; employment law in 13
Sales, pipeline, RevOps, customer success	26-sales-revenue-customer-success.md	GTM/offer in 19; negotiation in 23; economics in 12
Communication, reputation, editorial/content operations and distribution	27-communication-reputation-content-operations.md	marketing acquisition in 19; content design in 20; journalism in 29
Innovation systems, R&D, technology maturity/transfer	28-innovation-rd-technology-management.md	product/strategy in 02; engineering readiness in 14; finance/IP in 12/13
Journalism, fact-checking, editorial verification	29-journalism-verification-editorial-integrity.md	generic evidence/OSINT in 31; legal investigation in 13; communication distribution in 27

A specialization may change preconditions, experimental unit, authority, metric, or failure modes. Do not add both generic and specialized cards to the stack as if they were separate work unless each produces a distinct artifact.

Canonical ownership and multi-category routing

Use faceted classification for discovery, but keep one canonical card for each generic method family.

1. Classify the object type independently from its domain. A method, notation, standard, artifact, assurance mechanism, and metric may coexist in one stack without being alternatives.
2. Identify the primary decision or output. Place the generic card in the catalog whose methods own that output and its selection logic.
3. Check preconditions and failure modes. If a domain changes the unit of analysis, authority, evidence, measurement system, or validation, create a linked specialization in the domain catalog.
4. Add secondary routing aliases when the method is useful elsewhere but its generic procedure and boundaries remain unchanged. Do not duplicate the full card.
5. Treat synonyms and branded variants as aliases unless they change procedure, assumptions, outputs, governance, or evidence. A changed variant receives its own card and an explicit relationship to the canonical one.
6. Apply the subtraction and interface tests: every copied or specialized entry must produce a distinct consumed output, and the link must state what is inherited versus changed.

This is a lightweight application of faceted classification plus Situational Method Engineering: facets improve retrieval; canonical ownership prevents drift; situational tailoring handles real domain differences. Use SKOS relations only if the taxonomy becomes machine-maintained or needs interoperable broader/narrower/related links.

Methodological-density gate

Create or expand a domain catalog only when the area contains reusable ways to act and choose, not merely subject matter to retrieve. Apply this gate before broad domain research.

Criterion	0	1	2
Procedural corpus	mostly facts, rules, sources, or tools	a few reusable procedures	several distinct method families with steps and outputs
Selection logic	no meaningful alternatives	informal context differences	explicit triggers, preconditions, and do not use boundaries
Artifacts and gates	information summaries only	some templates/checklists	decisions, models, plans, tests, controls, or acceptance gates
Validation	authority/source lookup only	descriptive metrics	independent review, experiment, conformance, outcome, or failure tests
Canonical anchors	vendors/listicles dominate	mixed practitioner sources	standards, official owners, seminal research, or mature professional guidance
Distinct ownership	duplicates an existing catalog	useful specialization	coherent decision family with clear interfaces to adjacent catalogs

Interpret the proposed score as an architecture heuristic, not a scientific measurement:

• 9–12: a separate reference catalog is usually justified;
• 6–8: add a section or specialization to an existing owner;
• 0–5: treat it as subject matter, source registry, legal/data overlay, or tool/platform detail—not as a methodological domain.

Reject a proposed domain when most search results would necessarily be laws, rates, registries, datasets, vendors, platform features, or current facts. Tax, legislation, case law, and sector data may constrain a method stack, but do not become catalogs merely because they are important. Add a sector catalog only when the sector changes the procedure, unit of analysis, authority, evidence, validation, or failure modes materially.

Saturation protocol

Candidate saturation

Scale discovery to the proportionality tier:

Tier	Minimum discovery depth
1	primary domain; canonical definition/status; one focused expansion only if fit is unclear
2	primary plus one adjacent domain when warranted; problem/failure and artifact queries; one no-change expansion
3	primary plus up to two adjacent domains; all four query families; citation/version chasing; two successive no-change expansions
4	Tier 3 plus jurisdiction/authority, independent limitation/effectiveness evidence, and explicit adversarial search for invalidating conditions

The four query families are problem/failure, desired artifact, lifecycle phase, and known candidate alternatives. Stop when the required successive expansions add no candidate that changes the top stack, hard constraints, or material risk controls. Expand beyond the initial catalog cap only when this test reveals a specific uncovered capability or gate.

Record a newly found synonym as an alias, not a new candidate. Record a variant separately only when it changes procedure, assumptions, outputs, governance, or evidence.

Evidence saturation

For a selected method, seek:

• a canonical definition or normative specification;
• current version/status from the owner;
• implementation guidance or a worked example;
• limitations, failure conditions, or empirical evaluation;
• one independent source when a high-stakes claim depends on efficacy.

Stop when new sources repeat known claims without changing implementation, boundary conditions, confidence, or risk.

Research record

Maintain this compact ledger:

Candidate	Search anchors	Type/status	Included or rejected	Decisive evidence	Version checked	Open question

Separate three claim states:

• Verified: supported by an authoritative or appropriately strong source.
• Rejected: contradicted, obsolete, out of scope, or excluded by a hard constraint.
• Possible: relevant but not yet supported strongly enough for selection.