From Data Model to Data Product

Converting inward-looking models into outward-looking products without delivery spaghetti.

Problem -> Shift -> Outcome

Problem	Strong internal models still coexist with 200+ uncontrolled outbound feeds. Spaghetti is born at delivery, not modeling.
Shift	Separate internal model (operations) from external product (exchange). Keep both, but with explicit boundaries.
Outcome	Collapse 200+ custom feeds into 5-8 canonical product contracts with versioned interfaces.

200+ -> 5-8

Feeds to Contracts

6

Non-Negotiable Delivery Layers

4 Phases

Execution Sequence

Clean Distinction

Data Model (Inward)	Data Product (Outward)
Optimises finance operations	Optimises exchange
Controls internal accuracy	Controls external stability
Focus on entities	Focus on contracts
Normalisation-oriented	Interface consistency-oriented
Invisible externally	Publicly declared

6 Non-Negotiable Delivery Layers

Canonical Identity - align keys across boundaries.
Product Contract - declare schema, semantics, SLA.
Version Gate - prevent unmanaged breaking changes.
Delivery Adapter - one pattern for SQL/API/Event/File.
Monitoring - SLA, DQ, freshness, usage visibility.
Consumer Registry - who consumes what, and since when.

4-Phase Path (Execution)

Phase 0 (Week 1-2): Inventory all outbound feeds and consumers.
Phase 1 (Month 1-3): Freeze new custom feeds and cluster interfaces.
Phase 2 (Month 3-6): Productise with contracts, adapters, sidecar rails.
Phase 3 (Month 6-12): Migrate and sunset legacy unmanaged exports.

Principle: Do not repaint the house before fixing the plumbing.

Good Cop - Success Signals

Internal models remain clean while external products stay stable.
Consumers onboard in days using reusable contracts.
Changes are versioned and impact-visible, not surprising.
Interface surface shrinks materially (target ~80% reduction).

Bad Cop - Failure Signals

Beautiful internal model but custom feeds continue to grow.
Products are renamed labels while delivery stays bespoke.
Governance exists on paper but adapters/version gates are weak.
Identity remains inconsistent across consumers and interfaces.

Problem Signals Dashboard

Total Models

87

+5 this month

Healthy

84

96.6% coverage

Breached

1

GL → Nostro Recon

Quality Avg

96.2%

↑ 0.4% from last week

Consumers

143

+8 new

Product Health

Product	Models	Health
Finance	45	98.1%
Risk	32	95.3%
Regulatory	10	94.2%

Recent Activity

Event	Model	Time
Quality Alert	Model 6	2h ago
Consumer Added	Model 5	4h ago
Schema Updated	Model 1	1d ago

Product Supply Chain Map

Product Hub:

Finance Hub - Active Models

Active Models Table

Model	Contract ID	Quality	Status	Consumers

Legacy Connections

Legacy System	Models Served	Status
Legacy Finance System	2	Active
Old Report Engine	1	Deprecated

Product Contract Catalog

Finance

Core financial data and analytics products

6 Models • 45 Consumers

Risk

Risk assessment and exposure products

4 Models • 32 Consumers

Regulatory

Regulatory reporting and compliance products

2 Models • 10 Consumers

Finance Product Detail

Finance Product Overview

The Finance product line provides core financial data, analytics, and reporting capabilities. Serving 45 consumers across 6 active models with 98.1% average quality.

Metric	Value
Total Models	6
Active Consumers	45
Avg Quality	98.1%
SLA Compliance	99.3%

Product profile coverage: Business, Data, Delivery, Consumption, Quality, Security, Reliability, and Change are declared and version-governed.

Data (Grain, Keys, Schema)

Field	Type	Required	Classification
posting_date	Date	✓	Public
legal_entity	String	✓	Public
amount_usd	Decimal	✓	Confidential
currency	String	✓	Public

Consumption Views & Consumer Models

Model Name	Contract ID	Quality	Status

Semantic Mapping

Mapping ID	Source	Target Model	Mapping Type	Status
MAP-FIN-001	GL System	CTR-FIN-001	Relational	Certified
MAP-FIN-002	GL + Manual Adjustments	CTR-FIN-002	Model-to-Model	Certified
MAP-FIN-003	P&L System	CTR-FIN-003	Relational	Certified

Delivery Profiles

Runtime	Environment	Connector	SLA Tier	Last Deploy
RT-FIN-PROD	Production	JDBC + API Gateway	Gold	2026-03-01
RT-FIN-UAT	UAT	JDBC + Stub Service	Silver	2026-02-26

Quality & Security Controls

Control	Type	Coverage	Latest Run	Status
Non-null required fields	Schema Constraint	100%	7m ago	Pass
Amount range validation	Business Rule	99.9%	7m ago	Pass
Freshness under 24h	Timeliness Test	98.6%	7m ago	Failing on CTR-FIN-006

Change (Version & Compatibility)

Release	Scope	Change Type	Approved By	Date
2026.03.0	CTR-FIN-001 to 003	Non-breaking	Data Product Council	2026-03-01
2026.02.2	CTR-FIN-006 controls	Breaking	Architecture Review Board	2026-02-18

Reliability (SLA, Recovery, Ownership)

Sidecar services manage authentication, monitoring, and data quality for the Finance product line.

Service	Status	Last Check
Auth Service	Active	5m ago
Quality Monitor	Active	2m ago

Traceability (Lineage & Accountability)

Source	Mapping	Target
GL System	1:1 Direct	Model 1-5
P&L Extract	1:N Join	Model 3,5

Consumer Models (Not Custom Feeds)

ATLAS Risk Engine ▼

Model	Integration	Last Update
CTR-FIN-001	REST API	2h ago

Credit Analytics ▼

Model	Integration	Last Update
CTR-FIN-001	REST API	4h ago

ZEUS Regulatory ▼

Model	Integration	Last Update
CTR-FIN-002	REST API + CSV	1h ago

Board Reports ▼

Model	Integration	Last Update
CTR-FIN-003	Dashboard View	30m ago

Sidecar Foundation

Contract Registry ▼

Contract ID	Product	Version	Status	Lifecycle	Owner	Change Type	Expires

Quality Framework (42 Rules) ▼

Rule	Type	Status	Pass Rate
Completeness Check	Schema	Pass	100%
Data Range Validation	Range	Pass	99.9%
Freshness Check	Timing	Pass	99.5%
Uniqueness Validation	Schema	Pass	100%
Consistency Check	Logic	Pass	98.2%
Timeliness Validation	Timing	Fail	88.4%

Lineage Tracker (386 Mappings) ▼

Source Field	Transformation	Target Field
GL.posting_date	Direct Copy	Model1.posting_date
GL.amount	Currency Convert	Model1.amount_usd
GL.account + GL.entity	Concatenate	Model5.business_line
PL.revenue - PL.cost	Calculate	Model3.net_income_usd
GL.amount (30d)	Window Agg	Model5.rolling_30d_avg

SLA Monitor (13/14 In SLA) ▼

Model	Target	Actual	Status
Model 1	99.5%	99.1%	In SLA
Model 2	99.5%	98.7%	In SLA
Model 3	99.0%	97.5%	In SLA
Model 4	98.0%	97.1%	In SLA
Model 5	98.5%	95.8%	In SLA
Model 6	98.5%	88.4%	Breached

Access Control (47 Consumers) ▼

Consumer	Models	Access Level	Control Scope	Policy Ref
ATLAS Risk Engine	CTR-FIN-001	Read	Prod only, no bulk export	ACP-CTR-FIN-001
Credit Analytics	CTR-FIN-001	Read	Prod + UAT, masked PII	ACP-CTR-FIN-001
ZEUS Regulatory	CTR-FIN-002	Read	Regulatory extracts only	ACP-CTR-FIN-002
Compliance Team	CTR-FIN-002	Read	Case workflow + audit trail	ACP-CTR-FIN-002
Board Reports	CTR-FIN-003	Read	Executive aggregate only	ACP-CTR-FIN-003
CFO Dashboard	CTR-FIN-003	Read	Executive aggregate only	ACP-CTR-FIN-003

Monitoring, SLA & Reliability

League Table

Rank	Model	Contract ID	Quality Score	Completeness	Accuracy	Timeliness
1	GL → Credit Exposure	CTR-FIN-001	99.1%	100%	99.2%	98.3%
2	Balance Sheet → MAS	CTR-FIN-002	98.7%	99.5%	98.9%	98.1%
3	P&L → Board Dashboard	CTR-FIN-003	97.5%	99.0%	97.8%	96.2%
4	FTP → Liquidity	CTR-FIN-004	97.1%	98.5%	97.4%	95.9%
5	GL+P&L → Revenue	CTR-FIN-005	95.8%	97.2%	96.1%	94.3%
6	GL → Nostro Recon	CTR-FIN-006	88.4%	88.4%	92.1%	81.2%

Freeze & Rationalise Interfaces

Interface Freeze Policy

Filterable Integrations

Interface	Type	Models	Consumers	Status

Legacy Migration Board (Sunset Timelines)

Legacy Pattern	Target Contract Pattern	Sunset Date	Migration Wave	Owner	Status

Contract-first Consumer Onboarding

1

Registration

2

Model Selection

3

Policy & Purpose

4

Authentication

5

Activation

Step 1: Consumer Registration

Register your consumer application to access Archetype 2 data products.

Application Name

Team

Contact Email

Use Case Description

Requested Integration Pattern

Step 3 Preview - Policy Gate: Purpose-of-use declaration, data retention period, and required policy bundle (e.g., ACP-CTR-FIN-001) must be approved before credentials are issued.