Skip to content

The Missing Discipline: Managing Meaning

Every important thing has a discipline, except meaning

Discipline Thing being managed Typical signature artifact
ITSM / ITIL IT services Service catalog
DAMA / Data Governance Data Data policies, stewardship assignments
TOGAF Enterprise architecture Architecture models
COBIT IT governance and control Control objectives
MDM Master entities and records Golden record
Model Risk Management Analytical models Model inventory, validation report
AI Governance AI systems AI risk assessment
Semantic Management (SMF) Business meaning Semantic Contracts (Metric Contract, Definition Contract)

Why now

Business meaning used to be managed informally. Experienced analysts knew which "revenue" a report used and quietly corrected for it. That stops working when:

  • the same metric is rebuilt in five tools, each with its own logic;
  • leaders from different parts of the business use the same words for different things;
  • AI assistants answer business questions directly, with no analyst in between to catch the mismatch.

Analytics and AI are now the foundation for decisions. The meaning underneath them needs its own lifecycle, its own owners and its own checks.

Where SMF hands off to its neighbors

SMF does not take over any existing discipline. It manages the part each one leaves open.

Discipline Manages SMF manages the part it leaves open
DAMA / Data Governance Data: quality, stewardship, access What the data means, and which meaning applies where
MDM The Customer records What "Customer" means from each perspective
Model Risk Management The analytical model Whether a model's output means what the report or decision assumes
AI Governance The AI system: risk, safety, accountability Whether the AI interpreted business meaning correctly
ITSM / ITIL IT services Changes to meaning, and failures of meaning, flowing into existing change and incident processes
TOGAF Architecture How meaning is realized across systems
COBIT IT control objectives Evidence of control over meaning: ownership, approvals, checks

These lines are drawn from practice and are open to challenge. See CONTRIBUTING.md.

What SMF is and isn't

SMF is:

  • a discipline and an operating model;
  • technology-independent;
  • adoptable one practice at a time;
  • designed to work with other frameworks, together or separately.

SMF is not:

  • a catalog;
  • a semantic layer;
  • a metric store;
  • a knowledge graph or ontology;
  • a governance platform.

Those are implementation technologies for SMF practices. See works-with.md.