Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ISO/IEC 26925:2013 — Information Technology — Data Quality Model
Information technology — Data management — Data quality model
Overview of the Data Quality Model
ISO/IEC 26925:2013 establishes a comprehensive data quality model for information technology systems, providing a standardized framework for defining, measuring, and managing the quality of data throughout its lifecycle. As organizations increasingly rely on data-driven decision-making, the ability to assess and ensure data quality has become a critical business capability. This standard addresses that need by specifying a set of quality characteristics and providing guidance on how to measure and evaluate them in practical contexts.
Data quality is not an absolute concept — it is context-dependent. Data that is perfectly fit for one purpose may be completely unsuitable for another. ISO/IEC 26925 recognizes this by defining quality in terms of characteristics that can be measured against specific requirements for a given use context.
The model builds upon established concepts from ISO/IEC 25012 (data quality model for software product quality) and extends them with a broader perspective that encompasses data management processes, data governance frameworks, and organizational responsibilities. The standard defines a taxonomy of data quality characteristics organized into three main categories: inherent data quality, system-dependent data quality, and data quality management capabilities.
| Quality Category | Characteristics | Focus |
|---|---|---|
| Inherent data quality | Accuracy, completeness, consistency, credibility, currentness | The data itself, independent of the system that processes it |
| System-dependent data quality | Availability, portability, recoverability | The system’s ability to preserve and deliver data quality |
| Data quality management | Monitoring capability, enforcement capability, improvement capability | Organizational processes for managing data quality |
The three-category model is a powerful analytical tool: it helps organizations distinguish between data quality problems that originate in the data collection process (inherent), those caused by inadequate systems (system-dependent), and those resulting from weak governance processes (management capability).
Data Quality Dimensions in Detail
The inherent data quality characteristics represent the foundational attributes of data quality. Accuracy measures the degree to which data correctly represents the real-world entity or event it describes — for example, whether a customer’s address in the database matches their actual physical address. Completeness assesses whether all required data elements are present, considering both mandatory fields and optional elements that provide context. Consistency verifies that data does not contain contradictions across different records, databases, or points in time. Credibility evaluates the trustworthiness of the data source and the data collection method. Currentness (also called timeliness) measures whether data reflects the current state of the real-world entity within an acceptable time window.
System-dependent data quality characteristics address the role of the information system in preserving data quality. Availability measures the degree to which data is accessible when needed, encompassing both system uptime and data retrieval performance. Portability assesses the ease with which data can be transferred between different systems or formats without loss of quality. Recoverability evaluates the system’s ability to restore data to a correct state after a failure or corruption event, including backup and disaster recovery capabilities.
Organizations commonly invest heavily in improving inherent data quality while neglecting system-dependent aspects. Yet data availability and recoverability are equally critical — data that is accurate but unavailable during a critical business decision is functionally useless. A balanced data quality program addresses all three categories.
The data quality management capability characteristics define the organizational processes needed to sustain data quality over time. Monitoring capability refers to the ability to continuously track data quality levels against defined targets. Enforcement capability measures the organization’s ability to prevent quality degradation through validation rules, access controls, and process controls. Improvement capability assesses the effectiveness of corrective actions and root cause analysis processes for addressing data quality issues.
| Dimension | Measurement Approach | Typical Metric |
|---|---|---|
| Accuracy | Comparison against authoritative source or physical verification | Percentage of records matching the reference |
| Completeness | Ratio of populated fields to expected fields | Percentage of mandatory fields populated |
| Consistency | Cross-record and cross-system comparison | Number of contradictions per 1,000 records |
| Currentness | Time since last update versus required freshness | Percentage of records within acceptable age window |
| Availability | System uptime and query response time | 99.9% uptime, <500 ms average query time |
Implementing the Data Quality Model
Implementing ISO/IEC 26925 in practice requires organizations to translate the abstract quality model into concrete, measurable quality requirements for their specific data assets. The recommended approach begins with a data quality assessment to establish baseline quality levels across all relevant dimensions, followed by the definition of quality targets aligned with business requirements. These targets should be specific, measurable, achievable, relevant, and time-bound (SMART), and should reflect the criticality of each data asset to business operations.
The measurement framework specified in the standard provides flexible guidance rather than rigid prescriptions. Organizations may use automated data profiling tools to measure completeness and consistency, manual sampling and verification processes for accuracy assessment, and system monitoring tools for availability and recoverability measurement. The key principle is that measurement methods should be documented, repeatable, and subject to quality assurance themselves — consistent with the metrological principle that the measurement system must be at least as accurate as the data it measures.
A common failure mode in data quality initiatives is attempting to measure all dimensions for all data assets simultaneously. This approach overwhelms organizational capacity and dilutes focus. A more effective strategy is to prioritize high-impact data assets — such as customer records, financial data, and regulated information — and focus on the dimensions most relevant to their use, gradually expanding the scope as the data quality management capability matures.
The standard also addresses the organizational aspects of data quality management, including the assignment of data stewardship responsibilities, the establishment of data quality review boards, and the integration of data quality metrics into performance management systems. These governance mechanisms ensure that data quality is not treated as a one-time project but as an ongoing organizational capability that requires sustained investment and management attention.
Frequently Asked Questions
Q1: How does ISO/IEC 26925 relate to ISO/IEC 25012?
ISO/IEC 25012 provides a data quality model specifically within the context of software product quality, focusing on data as part of a software system. ISO/IEC 26925 extends this model to cover the broader data management lifecycle, including organizational processes, governance, and management capabilities. Organizations implementing ISO/IEC 25012 can use 26925 to expand their quality program beyond software-centric boundaries.
ISO/IEC 25012 provides a data quality model specifically within the context of software product quality, focusing on data as part of a software system. ISO/IEC 26925 extends this model to cover the broader data management lifecycle, including organizational processes, governance, and management capabilities. Organizations implementing ISO/IEC 25012 can use 26925 to expand their quality program beyond software-centric boundaries.
Q2: Can the data quality model be applied to unstructured data?
While the model was initially conceived with structured data in mind, its quality characteristics can be adapted for unstructured data such as text documents, images, and multimedia content. For unstructured data, accuracy may be assessed through content verification sampling, completeness through metadata analysis, and consistency through cross-document semantic comparison. The measurement methods differ, but the underlying quality dimensions remain applicable.
While the model was initially conceived with structured data in mind, its quality characteristics can be adapted for unstructured data such as text documents, images, and multimedia content. For unstructured data, accuracy may be assessed through content verification sampling, completeness through metadata analysis, and consistency through cross-document semantic comparison. The measurement methods differ, but the underlying quality dimensions remain applicable.
Q3: What is the relationship between data quality and data governance?
Data quality management as defined in ISO/IEC 26925 is a key component of data governance. The standard’s data quality management capability characteristics (monitoring, enforcement, improvement) directly support data governance objectives by providing measurable criteria for evaluating governance effectiveness. Data governance establishes the policies, roles, and responsibilities, while the 26925 quality model provides the measurement framework to assess whether those policies are achieving their intended outcomes.
Data quality management as defined in ISO/IEC 26925 is a key component of data governance. The standard’s data quality management capability characteristics (monitoring, enforcement, improvement) directly support data governance objectives by providing measurable criteria for evaluating governance effectiveness. Data governance establishes the policies, roles, and responsibilities, while the 26925 quality model provides the measurement framework to assess whether those policies are achieving their intended outcomes.
Q4: How often should data quality be measured?
The frequency depends on the volatility and criticality of the data. High-volatility data such as customer contact information or real-time sensor data may require daily or even continuous monitoring. Low-volatility reference data such as product classification codes may be adequately assessed quarterly or annually. The standard recommends that measurement frequency be documented in the data quality policy and adjusted based on observed quality trends and business impact assessments.
The frequency depends on the volatility and criticality of the data. High-volatility data such as customer contact information or real-time sensor data may require daily or even continuous monitoring. Low-volatility reference data such as product classification codes may be adequately assessed quarterly or annually. The standard recommends that measurement frequency be documented in the data quality policy and adjusted based on observed quality trends and business impact assessments.
