What is data quality?

Data quality is a useful term — but not a sufficient diagnosis in high-consequence environments.

Useful as a label

It provides a way to describe common attributes such as accuracy, consistency, completeness and timeliness.

Weak as a diagnosis

It does not always reveal whether the problem is missing data, distorted meaning, weak controls or fragmented ownership.

What is data integrity?

Data integrity is the stronger and more practical question: can the data still be trusted across the full journey, and can that trust be evidenced rather than assumed?

The critical difference

Data quality asks whether data looks acceptable. Data integrity asks whether the data remains complete, correct and governable where decisions and controls actually depend on it.

In practical terms

Source

Data originates correctly.

Transfer

Expected records fail to arrive.

Transformation

Fields are distorted or remapped.

Aggregation

Partial populations are still used.

Decision

The output looks valid — but is not trustworthy.

Why the distinction matters in decision-critical environments

Missed transactions

Data may look clean in the target system, while a subset of expected records never arrived.

Incorrect risk decisions

Fields remain populated, but values or meaning have changed through transformation logic.

Regulatory exposure

The organisation cannot prove completeness or correctness when challenged.

Late discovery

Problems are found after outputs drift, not when integrity first breaks.

Where data quality thinking usually fails

  • Too much focus on outputs rather than the journey that produced them.
  • Terms used too broadly, masking whether the real issue is completeness, correctness or control weakness.
  • Ownership remains unclear across platforms, sources and downstream consumers.
  • Confidence in activity replaces proof of integrity.

What a data integrity approach introduces instead

Completeness controls

Prove whether all expected populations arrive where they should.

Correctness controls

Test whether meaning, mappings and transformations remain right.

Detective monitoring

Expose breaks early rather than waiting for late downstream symptoms.

Ownership clarity

Make the integrity question governable, not just observable.

Executive takeaway

Data quality improves what you can see. Data integrity determines whether what you see can be trusted.