Data quality in plain language
Data quality describes whether data is sufficiently fit for its intended business or operational use.
AskAQA principle: Data can be technically valid and still be business-wrong.
Why data quality matters
Poor data can create incorrect reports, failed integrations, bad decisions, customer impact, compliance problems, automation failures, and misleading AI output.
Quality depends on intended use
The same dataset may be good enough for exploratory analysis but unacceptable for financial reporting, safety decisions, customer billing, or regulatory submissions.
Core data quality questions
- Is it correct?
- Is required data present?
- Does it agree across systems?
- Does it follow allowed rules?
- Are duplicates controlled?
- Is it current enough?
- Are relationships intact?
- Can its origin be traced?
Where data quality problems originate
| Stage | Example issue |
|---|---|
| Source | Missing or incorrect source values |
| Ingestion | Rows dropped or file parsed incorrectly |
| Transformation | Wrong join, formula, or mapping |
| Storage | Duplicate, stale, or corrupted records |
| Semantic layer | Incorrect measure or relationship |
| Dashboard | Wrong filter, aggregation, or date context |
Data quality needs ownership
QA can validate and expose data problems, but business definitions, source ownership, stewardship, transformation ownership, and consumer accountability must also be clear.