COMPEL Glossary / data-quality
Data Quality
Data quality is the degree to which data meets requirements for accuracy, completeness, consistency, timeliness, validity, and uniqueness.
What this means in practice
For AI, data quality demands are more stringent than for traditional analytics because AI models have no contextual judgment -- they learn whatever patterns the data contains, including patterns introduced by quality defects. A 10% data quality deficit can produce a 30-50% degradation in model performance due to the multiplicative relationship between data quality and AI outcomes. Industry surveys consistently identify data quality as the primary reason for AI project failure. In the COMPEL framework, data quality is assessed during the Calibrate stage, data quality SLAs between producers and consumers are established during Organize, and automated quality monitoring with alerting is a key capability at maturity Level 3 and above.
Why it matters
Data quality is the single most cited reason for AI project failure across industry surveys. A 10% data quality deficit can produce 30-50% degradation in model performance due to the multiplicative relationship between data quality and AI outcomes. Organizations that treat data quality as someone else's problem consistently produce AI systems that are unreliable, biased, or outright harmful, wasting investment and eroding stakeholder trust.
How COMPEL uses it
Data quality is assessed during Calibrate as a foundational capability. During Organize, data quality SLAs between data producers and AI consumers are established. Automated quality monitoring with alerting is a key capability at maturity Level 3 and above. The Evaluate stage measures quality trends, and the Learn stage captures lessons about quality management to improve practices in subsequent COMPEL cycles.
Related articles in the Body of Knowledge
Related Terms
Other glossary terms mentioned in this entry's definition and context.