ICCP DGSP cheat sheet
ICCP DGSP · Keep the key ideas within reach.
Data Governance & Stewardship
Data governance is the decision-rights and accountability framework for data: it defines who decides, who is accountable, and how decisions are enforced; stewardship is the execution layer that applies those decisions day to day.
Core governance artifacts follow a hierarchy: policy (intent and constraints) → standard (mandatory, measurable requirement) → procedure (step-by-step execution) → guideline (recommended practice, non-mandatory).
A data domain is a subject-area partition of data with a single accountable owner; each domain contains data products or subject areas, each with a steward responsible for quality, definitions, and issue triage.
RACI for data: data owner = accountable for access/use decisions and protection; data steward = responsible for quality and metadata execution; data custodian/technical owner = responsible for storage, backup, and platform controls; consumer = informed.
A data governance council (steering committee) sets policy and resolves cross-domain conflicts; a data governance office (DGO) operates the program; domain councils or working groups handle domain-level decisions.
Decision rights must be explicit per decision type: definition approval, access approval, quality thresholds, issue prioritization, and exception handling can each sit with a different role - a single 'owner does everything' model fails at scale.
The operating model choices are centralized (single team decides), decentralized (business units decide), and federated/hybrid (central standards plus domain execution); federated is the common target for large enterprises.
Policy enforcement mechanisms are preventive (access controls, mandatory fields, workflow gates) or detective (quality dashboards, audits, exception reports); a control is only effective if a named role owns remediation.
Governance metrics typically include policy exception rate, issue closure time, definition coverage, stewardship coverage (domains with assigned stewards), and percentage of critical data elements under stewardship.