SD-13.7 — Data Quality & Governance

Business Domain: BD-13 Investment Data & Reporting (Cross-cutting — data) · Applies: BOTH

Purpose

Defines and enforces data quality, ownership and stewardship across the institution’s investment data. Where the other BD-13 Service Domains produce data — masters, feeds, extractions, metrics — SD-13.7 is the capability that holds them all to a standard: it sets the quality rules, names the data owners and stewards, runs the workflows that detect and remediate issues, and maintains the lineage that makes a data point traceable to its source. It is a governance Service Domain — it does not source, curate or interpret data itself; it governs how the domains that do are run. It is the BD-13 analogue of an investment-compliance function: a control layer over the data estate.

Service Operations

Data-quality rule definition — define the quality rules that data must satisfy — completeness, validity, consistency, timeliness, uniqueness, accuracy — per data domain, and the tolerance and severity of each.
Data-quality monitoring — run the defined rules continuously across the masters, feeds and certified datasets, and produce data-quality scores and dashboards.
Data ownership and stewardship assignment — name the accountable data owner and the operational data steward for each data domain, and maintain the stewardship operating model.
Issue detection and remediation workflow — raise a data-quality issue when a rule fails, route it to the responsible steward, track it to resolution, and analyse recurring issues for root cause.
Data-lineage tracking — maintain the lineage of each significant data element — where it originated, what transformations it passed through, which datasets and reports it flows into — so a figure can be traced end to end.
Data-policy and standards governance — own the firm’s data policies (classification, retention, access, quality standards) and the change process for them.
Data-governance forum operation — operate the data-governance committee or forum where ownership disputes, policy changes and material data-quality issues are adjudicated.
Data-quality reporting — report the state of data quality and the open-issue backlog to data owners and to enterprise governance (BD-14).

Inputs and outputs

Inputs: the data produced across BD-13 (and the operational data of BD-01 to BD-12) as the population the quality rules run over; the firm’s data policies; the vendor-quality signals from SD-13.4.
Outputs: data-quality scores and dashboards; the stewardship assignments and the issue backlog; data lineage; the governed data policies — consumed by every BD-13 Service Domain (as the standard they are held to), by SD-13.8 (only quality-passing data is certified), by SD-13.10 reporting, and by SD-14.8 Internal Audit and SD-14.7 Internal Control & Assurance.

Entities

Owns: E-11 Classification Type & Value (the extensible taxonomy behind every time-varying classification — sector, geography, ESG strategy, fund stage) and E-12 Classification History (the bi-temporal record of which value applied to which subject across which interval) — the governance entities. SD-13.7 also owns the data-quality rule set, the stewardship model, the issue backlog and the lineage records as Service-Domain artefacts (an open question is whether these should themselves be first-class entities — see below). SD-13.7 does not own a business-data master; it owns the classification machinery and the metadata that governs the masters.
Consumes: every master and dataset across BD-13 as the population it monitors — E-01, E-02, E-07 (any method), E-08, E-10, PM-01, and the rest. (SD-13.7’s consumption of its own owned E-11 / E-12 is the consumption-of-owned-entity case; consumers of E-11 / E-12 across the model — the issuer-sector, instrument-classification, geography classifications — read against SD-13.7’s authoritative source.); the SD-13.4 vendor-quality signals.
Open question: whether the data-quality rule, the data-quality issue and the lineage record should be modelled as first-class entities — a governance-metadata extension to the entity model. They are currently the Service Domain’s internal artefacts.

Standards

DAMA-DMBOK (Data Management Body of Knowledge) — the reference framework for the data-management knowledge areas, including the data-quality dimensions SD-13.7 measures against.
EDM Council DCAM (Data Management Capability Assessment Model) — the capability-maturity framework used to assess and benchmark the data-governance and data-quality functions; an EDM Council standard widely adopted across the buy-side.
ISO 8000 — the data-quality standard, where the firm formalises data-quality requirements.

Open extensions

The governance-metadata entities question — data-quality rule, issue, lineage record as first-class entities.
The DCAM capability-maturity assessment as a recurring, scored operation.
The interaction with SD-14.4 Model Governance where a model consumes data the quality of which SD-13.7 governs.