--- title: "Data Quality Dimensions | Altss Glossary" description: "Data quality dimensions are the measurable characteristics used to judge whether data is fit for a given use; a commonly used set is accuracy,…" canonical: "https://altss.com/glossary/data-quality-dimensions" --- Glossary · Evidence & data # Data Quality Dimensions Data quality dimensions are the measurable characteristics used to judge whether data is fit for a given use; a commonly used set is accuracy, completeness, consistency, timeliness, validity and uniqueness. Publisher: Altss LLCPublished 2026-01-06Content modified 2026-10-01 ALTSS-DATA-038 "Is this data good?" is too vague to act on. Dimensions split it into questions that can be measured: does the value match reality, is it there at all, does it agree with other values, is it current, is it in the right format, and is each thing recorded once? Quality is fitness for a purpose, so the same dataset can be good enough for one use and not for another. ### Formal definition The UK Government Data Quality Framework (2020) describes data quality as fitness for purpose and adopts six core dimensions as defined by the UK chapter of the Data Management Association (DAMA UK): completeness, uniqueness, consistency, timeliness, validity and accuracy. The International Organization for Standardization (ISO) and International Electrotechnical Commission (IEC) standard ISO/IEC 25012:2008 defines a data quality model of 15 characteristics viewed from an inherent and a system-dependent point of view. ## The six core dimensions | Dimension | Question | Private-markets example | | --- | --- | --- | | Accuracy | Does the value match reality? | The recorded fund size matches the size stated in the fund's final-close announcement or filing | | Completeness | Are the expected records and values present? | Every fund in scope has a vintage year; a missing value is recorded with a reason, such as unknown or not disclosed | | Consistency | Do values for the same entity contradict each other? | A fund's vintage year is not later than its final close; currencies match across fields | | Timeliness | Does the data reflect its period and is it up to date? | A role was last observed recently enough for its use (see [data freshness](https://altss.com/glossary/data-freshness)) | | Validity | Is the value in the expected range and format? | A legal entity identifier has 20 alphanumeric characters; dates are real dates | | Uniqueness | Is each entity recorded once? | One allocator is not held as three records (see [entity resolution](https://altss.com/glossary/entity-resolution)) | Definitions follow the UK Government Data Quality Framework, which takes them from DAMA UK and calls dimensions "measurable features or characteristics of data". ## Other models No single list is authoritative. **ISO/IEC 25012** defines a general data quality model for structured data held in computer systems, with 15 characteristics considered from two points of view: inherent to the data and dependent on the system that holds it. **DAMA-DMBOK** (2nd edition, revised 2024) sets out data quality dimensions in tables, and its 2024 revision added currency as a recognised dimension. Organisations choose the dimensions that matter for their use and define how each is measured. ## Measuring a dimension Each measurement needs a rule, a population, a date and, for accuracy, a reference: - **Completeness**: share of required values present, with "not applicable" separated from "unknown". - **Uniqueness**: share of entities held once; measuring it requires deciding what counts as the same entity. - **Validity**: share of values that pass format and range rules. Validity can be checked automatically; accuracy cannot. - **Accuracy**: share of a sample confirmed against a reference, usually fresh evidence from an authoritative source. The result is only as good as the reference and the sample design. Every published rate should state its numerator, denominator, population and date. ## Dimensions interact - **Accuracy and timeliness.** Many "inaccurate" values were correct when captured and have since changed. A check should record which failure it found. - **Completeness and accuracy.** Filling gaps with guesses raises completeness and lowers accuracy. Estimated values need to be labelled as estimated (see [data provenance](https://altss.com/glossary/data-provenance)). - **Uniqueness and accuracy.** Merging aggressively removes duplicates and creates false identities; merging cautiously leaves duplicates. The trade-off is measured with [precision and recall](https://altss.com/glossary/precision-and-recall). - **Validity and accuracy.** A well-formed identifier can still belong to the wrong entity. ## Quality controls Dimensions are measured after the fact; controls prevent errors. Typical controls: schema and format validation at ingestion; cross-field consistency rules; detection of conflicting sources, preserving both values until resolved; outlier checks (an AUM figure that jumps by orders of magnitude); duplicate detection before new records are created; human review queues for high-impact ambiguity; and correction workflows that keep history. Controls should be field-aware: the checks for a legal name differ from those for a contact role. ## How LPs and GPs use the dimensions In [operational due diligence](https://altss.com/glossary/operational-due-diligence) of fund administrators, data vendors and reporting providers, the dimensions turn "is the data good?" into answerable questions: how is accuracy sampled, how are duplicates prevented, how are missing values recorded, how old is the evidence, and are rates reported with denominators and dates. For benchmark datasets, quality also depends on data coverage: which funds are in the universe at all, which completeness within each record cannot fix. ## How Altss applies this (Altss methodology) Altss assesses quality claim by claim through its three-dimension evidence model rather than a single quality score: the **evidence origin** of each evidence item (REGULATORY_PUBLIC_RECORD, OFFICIAL_INSTITUTIONAL, DISCLOSED, OSINT_SOURCED or LICENSED_THIRD_PARTY), the **derivation status** of each value (OBSERVED, DERIVED or ESTIMATED) and the dated **validation status** of each claim (UNVERIFIED, CORROBORATED, RESEARCH_VALIDATED or CONFLICTING). Missing values carry a recorded reason instead of a placeholder, estimates are never presented as observed values to raise completeness, and freshness is measured per claim. No quality rates are implied here. See the [Evidence & Provenance Standard](https://altss.com/knowledge-center/frameworks/evidence-and-provenance-standard). ## Worked example ### Illustrative accuracy check on a role field In a hypothetical dataset, 200 values of a "current chief investment officer" field are rechecked against fresh primary sources. 170 are confirmed (**85%**). Of the 30 that fail, 18 (9% of the sample) were correct when captured but the person has since changed (a timeliness failure), and 12 (6%) were wrong when captured (an accuracy failure at extraction or source). The headline 85% hides two problems with different fixes: rechecking fast-changing fields more often, and fixing the extraction or source. Examples are illustrative; figures are not market data. ## Not the same as - [Data Freshness](https://altss.com/glossary/data-freshness): Data freshness measures the age of the evidence for a value and informs the timeliness dimension; data quality is the full set of dimensions judged against a use. - [Entity Resolution](https://altss.com/glossary/entity-resolution): Entity resolution decides which records describe the same entity; the uniqueness dimension measures how well that has been done. - [Source Reliability](https://altss.com/glossary/source-reliability): Source reliability grades how dependable a source has proven to be; data quality dimensions describe the data held, whatever its sources. ## Common mistakes - Publishing one overall quality score that hides which dimension is failing. - Measuring accuracy without an independent reference. - Counting placeholders or "unknown" as complete. - Treating a valid format as proof of accuracy. - Removing duplicates by merging on similarity, which creates false identities. - Reporting rates without numerator, denominator, population and date. ## Edge cases - A field that does not apply to an entity (a fund's vintage year for an operating company) is neither complete nor incomplete; it is not applicable. - Two distinct entities with identical names are not duplicates; forcing uniqueness on names merges them. - Values that differ because they use different definitions (regulatory vs marketing AUM) are different attributes, not an inconsistency. - Accuracy of a time-varying value is checked against its as-of date, not against today. ## Questions ### What are the six dimensions of data quality? In the DAMA UK set adopted by the UK Government Data Quality Framework: accuracy, completeness, consistency, timeliness, validity and uniqueness. Other models, such as ISO/IEC 25012, use more characteristics. ## External standards | Standard | Relation | Note | | --- | --- | --- | | UK Government Data Quality Framework (2020), DAMA UK dimensions (Data quality dimensions - how to measure your data quality) | equivalent | | | ISO/IEC 25012:2008 (Data quality model (15 characteristics)) | broader | Inherent and system-dependent characteristics. | | DAMA-DMBOK 2nd edition revised (2024) (Data Quality chapter (2024 revision notes)) | related | | ## Sources - [The Government Data Quality Framework](https://www.gov.uk/government/publications/the-government-data-quality-framework/the-government-data-quality-framework). Government Data Quality Hub, UK Government (GOV.UK), Published 3 December 2020. Status: Current (checked 2026-10-01). Data quality dimensions - how to measure your data quality — supports: Fitness for purpose; six core dimensions and their definitions, attributed to DAMA UK - [ISO/IEC 25012:2008 Software engineering - SQuaRE - Data quality model](https://www.iso.org/standard/35736.html). ISO/IEC JTC 1/SC 7, ISO/IEC, 2008 edition. Status: Published (paywalled); listed as current on IEC webstore 2026-10-01 (checked 2026-10-01). Abstract (IEC webstore listing of ISO/IEC 25012:2008) — supports: General data quality model for structured data in computer systems; 15 characteristics considered from inherent and system-dependent points of view - [DAMA-DMBOK: Data Management Body of Knowledge, 2nd Edition Revised](https://dama.org/learning-resources/dama-data-management-body-of-knowledge-dmbok/). DAMA International, Technics Publications, 2nd edition revised (maintenance release), 2024. Status: Current; DMBOK 3.0 project under way since 2025 (checked 2026-10-01). dama.org revision notes for the 2024 release (Data Quality chapter) — supports: Data quality dimension tables refined; currency added as a recognised dimension in 2024 - [Introducing the Legal Entity Identifier (LEI)](https://www.gleif.org/en/about-lei/introducing-the-legal-entity-identifier-lei). GLEIF, Global Legal Entity Identifier Foundation, ISO 17442 identifier; web page accessed 2026-10-01. Status: Current (checked 2026-10-01). Introducing the LEI page — supports: LEI is a 20-character alphanumeric code (validity example) ## Related terms 6 terms - [Data Freshness](https://altss.com/glossary/data-freshness) - [Entity Resolution](https://altss.com/glossary/entity-resolution) - [Precision and Recall](https://altss.com/glossary/precision-and-recall) - [Data Provenance](https://altss.com/glossary/data-provenance) - [Source Reliability](https://altss.com/glossary/source-reliability) - [Operational Due Diligence](https://altss.com/glossary/operational-due-diligence) ## Concept record Concept ID ALTSS-DATA-038 Classification Evidence & data Topics Private markets data & OSINT Version 2.0.0 Last reviewed 2026-10-01 Structured data [JSON](https://altss.com/reference/concepts/data-quality-dimensions.json) ## Canonical URL https://altss.com/glossary/data-quality-dimensions