ipad-data-governance

IPAD Data Governance

The IPAD Framework implements data governance through the CDA model (Control, Decisions, Alignment), federated stewardship roles, six quality dimensions aligned with DAMA-DMBOK, and a four-tier data classification taxonomy enforced by the ICA security model.

CDA Governance Model

The CDA model provides the governance architecture with three interconnected pillars.

Control Pillar

Purpose: Ensure data integrity, quality, and compliance across all layers.

Key activities: data stewardship, quality enforcement, classification compliance, security governance, data contract monitoring.

Quality enforcement operates at three tiers:

1. Automated -- Rules in data fabric (Layer 3) validate at ingestion/transformation boundaries
2. Monitoring -- Continuous monitoring of 6 quality dimensions with alerting on threshold breaches
3. Manual Review -- Periodic reviews, trend analysis, remediation planning for persistent failures

Decisions Pillar

Purpose: Structure governance decision-making with clear accountability.

Level	Authority	Scope
Strategic	Data Governance Board (DGB)	Framework-wide policies, cross-domain standards
Domain	Domain Data Steward Council	Domain-specific implementation within central standards
Operational	Individual Data Stewards	Day-to-day data management decisions
Exception	DGB then Architecture Review Board	Deviations from central standards

Governance forums:

Forum	Chair	Cadence	Purpose
Data Governance Board	CDO	Monthly (90 min)	Strategic direction, cross-domain disputes, policy approval
Domain Stewardship Councils	Lead Domain Steward	Fortnightly (60 min)	Domain quality, data product lifecycle, operational improvements
Working Groups	Appointed by DGB/DSC	Weekly (during active period)	Time-bound, task-focused governance improvements
Governance Health Review	CDO	Quarterly (half day)	Comprehensive KPI review, principle compliance audit

Escalation SLAs: Domain Steward (5 days) -> Domain Council (10 days) -> DGB (15 days) -> Architecture Review Board (20 days).

Alignment Pillar

Purpose: Connect governance to IPA strategic objectives and principles.

• Outcome Mapping -- Every policy maps to IPA outcome areas (Information, Policies, Promotions, Stimulation, Relations)
• Stakeholder Alignment -- Reviewed against 5 stakeholder tiers (Leaders, Partners, Entities, Teams, Investors)
• Funnel Alignment -- Governance supports funnel stages (Outreach through Loyalty) without creating friction
• Principle Compliance -- Structured reviews against 26 IPAD architecture principles

Federated Governance Philosophy

Based on Principle P14 (Federated Governance with Central Standards):

• Central standards define the WHAT: quality thresholds, classification rules, metadata requirements, data contract specs
• Domain teams decide the HOW: implementation approach, tooling, scheduling, operational processes
• CDO role is Facilitator -- builds governance platform and defines standards, not a gatekeeper
• Graduated autonomy -- more mature domains earn greater operational independence

Six Data Quality Dimensions

All six dimensions are adopted from DAMA-DMBOK. Each is scored 0.00-1.00 with five rating bands.

1. Completeness

The degree to which required data values are present. Measures record completeness (actual vs expected records) and field completeness (populated mandatory fields vs total mandatory fields). Composite = (record x 0.5) + (field x 0.5).

2. Consistency

The degree to which data values are uniform and free from contradiction across datasets, systems, and time periods. Measures cross-source agreement (weight 0.4), referential integrity (0.4), and temporal consistency (0.2).

3. Accuracy

The degree to which data correctly represents the real-world entity or event. Measures source verification (weight 0.6) and range validation (0.4).

4. Timeliness

The degree to which data is available when needed and reflects current state. Measures freshness against SLA (weight 0.6) and normalised lag score (0.4).

5. Uniqueness

The assurance that no duplicate records exist for the same real-world entity. Measures record uniqueness (weight 0.6) and entity resolution (0.4).

6. Validity

The degree to which data conforms to business rules, formats, ranges, and referential integrity constraints. Measures format validity (weight 0.3), business rule compliance (0.4), and referential integrity (0.3).

Quality Thresholds by Maturity Level

Dimension	L1	L2	L3	L4	L5
Completeness	0.40	0.60	0.80	0.90	0.95
Consistency	0.40	0.60	0.75	0.85	0.95
Accuracy	0.50	0.65	0.80	0.90	0.95
Timeliness	0.40	0.55	0.75	0.85	0.95
Uniqueness	0.50	0.65	0.80	0.90	0.95
Validity	0.50	0.65	0.80	0.90	0.95
Composite Min	0.50	0.65	0.80	0.90	0.95

Composite formula: (Completeness + Consistency + Accuracy + Timeliness + Uniqueness + Validity) / 6

Enforcement: L1-L2 advisory only; L3 enforced at governance forums with remediation plans; L4-L5 automated enforcement with self-healing or escalation.

Stewardship Roles

Chief Data Officer (CDO)

Senior governance leader operating as Facilitator (not gatekeeper). Chairs DGB, defines central standards, provides governance platform and training. Required from L2+ (at L1, combined with IT leadership).

Domain Data Stewards

Responsible for governance within specific domains aligned with 5 IPAD data marts (Attract, Leads, Facilitation, Expansion, Advocacy) plus cross-cutting. Own data quality, maintain data products, lead Domain Stewardship Council. Required from L2+.

Data Custodians

Technical professionals managing data infrastructure within a domain. Implement storage, pipelines, integration; configure quality checks and access controls. Required from L2+.

Data Producers

Teams or systems that create, collect, or ingest data. Accountable for quality at point of creation; comply with data contracts. Formalisation increases with maturity.

Data Consumers

Teams, systems, or stakeholders that use data products. Use data per classification and purpose; report quality issues.

RACI Summary (Key Activities)

Activity	CDO	Domain Steward	Custodian	Producer	Consumer
Define central standards	A/R	C	I	I	I
Define domain quality rules	C	A/R	C	C	I
Monitor quality scores	I	A	R	I	I
Classify data assets	C	A/R	C	C	I
Maintain catalogue entries	I	A	R	C	I
Author data contracts	C	A	R	C	C
Manage data product lifecycle	I	A	R	C	C

Data Classification (4-Tier Taxonomy)

Tier	Description	Access Controls	Encryption	Review Cadence
Public	Openly available (e.g., published FDI statistics)	Open access	Optional	N/A
Internal	IPA operational data not for public release	Role-based access	At rest + in transit	Annual
Confidential	Investor data, commercial intelligence, personal data	Need-to-know access	Mandatory (AES-256+)	Semi-annual
Restricted	National security, highly sensitive policy data	Named individuals only	Mandatory + enhanced	Quarterly

Classification is enforced at every layer boundary (NFR-SEC-001). Dark data from Confidential/Restricted sources must be discovered, classified, and subjected to privacy impact assessment before mining.

Metadata Management

• Standard: DCAT-compatible metadata catalogue (required from L3+)
• Content: Entity definitions, attribute schemas, lineage, quality scores, classification, SLAs
• Data Contracts: Authored in ODCS v3 format; specify producer obligations, consumer expectations, quality SLAs
• Lineage: Tracked end-to-end from source (Layer 2) through transformation (Layer 3) to consumption (Layer 6)

ICA Security Model

The ICA model (Integrity, Confidentiality, Accessibility) is the cross-cutting security concern spanning all eight IPAD layers.

Pillar	Purpose	Key Controls
Integrity	Ensure data not tampered with or corrupted	Hash verification, checksums, immutable audit logs, digital signatures
Confidentiality	Protect from unauthorised access	Encryption, access control, classification enforcement
Accessibility	Ensure authorised users can access when needed	Availability SLAs, disaster recovery, self-service discovery

DAMA-DMBOK Alignment

DAMA Knowledge Area	IPAD Concept	IPAD Layer
Data Governance	CDA Model (Control, Decisions, Alignment)	Layer 4
Data Architecture	8-Layer Architecture Stack	All Layers
Data Modelling & Design	Entity Model (49 entities, 5 domains)	Layer 3-4
Data Storage & Operations	Data Stores (5 data marts)	Layer 4
Data Security	ICA Security Model	Cross-cutting
Data Integration & Interoperability	Data Fabric (Layer 3), API-first	Layer 3
Reference & Master Data	MDM (P13), Entity Reference Data	Layer 3-4
Data Warehousing & BI	Data Marts, Analytics (Layer 5-6)	Layer 5-6
Metadata Management	Data Catalogue (DCAT), Lineage	Layer 3
Data Quality	6 Quality Dimensions, Quality Framework	Layer 3-4

Key alignment points: IPAD quality dimensions directly adopted from DAMA-DMBOK; stewardship roles follow DAMA RACI pattern; governance lifecycle aligns with DAMA operating model; CDO defined as Facilitator per DAMA best practice.

Governance Adoption by Maturity Level

Level	Governance Model
L1: Ad-Hoc	Minimal -- basic stewardship and classification
L2: Centralised	Structured -- central IT-led with defined standards and reporting
L3: Federated	Full CDA -- federated model operational, all three forum tiers active
L4: Cognitive	AI-augmented -- automated compliance monitoring, proactive issue detection
L5: Autonomous	Self-enforcing policies with human strategic oversight only