Stockertim Dictionary · Human resources · Technology
Data quality audit
Also called data quality review, data audit
In short
A check that samples records, tests each field against agreed rules, counts the failures and assigns each fix to an owner with a date.
- Intermediate
- Process
- Checked
In detail
A data quality audit measures how good a set of data actually is. It begins with rules: a start date must exist and be before any leaving date, every employee must have one manager, a national insurance number must match the required format, and so on. The audit runs the rules over all records or a sample, counts the failures by type and estimates the error rate.
Results are reported against dimensions such as completeness, accuracy, consistency, validity, timeliness and uniqueness, the dimensions set out in government data quality guidance. Each type of failure is given an owner, a fix and a date, and the cause is investigated: an error that comes from a badly designed form will keep returning until the form changes.
Audits are especially important before a system migration or a major report, such as pay gap reporting, when errors become costly. Regular audits, monthly or quarterly, show whether quality is improving and give stewards evidence to change processes.
Worked example
An audit tests 2,000 records and finds 60 that fail at least one rule.
- Error rate: 60 / 2,000 = 0.03
- As a percentage: 0.03 × 100 = 3
Three per cent of records need fixing, each with an owner and a date.
In practice
Before moving to a new HR system, a team runs twenty rules over every employee record, reports the failures by field to each data steward and repeats the audit after each mock conversion until the error rate is low enough to go live.
Watch out
- Fixing errors without finding their cause means they return.
- An audit that checks format but not accuracy can pass wrong data that looks right.
Learn it properly
Official sources
- The Government Data Quality Framework · UK Government Data Quality Hub
- Quality statistics in government · Government Analysis Function
Facts last checked on 3 November 2026.
Education, not advice: the Stockertim Dictionary explains what terms mean. It never recommends buying or selling anything.