Why quality is decided at data entry, not in the dashboard — and how to measure it.
An error is paid for several times
Bad data costs once when it is entered, a second time when someone detects and fixes it, and a third time — the most expensive — when it goes undetected and a decision is made on it. The later an error is found, the higher its cost.
The six dimensions of quality
Accuracy, completeness, consistency, uniqueness, validity and timeliness: these six dimensions turn an impression ("our data isn't reliable") into an objective measure tracked over time.
Act at the source
The most cost-effective controls are those that stop errors getting in: closed value lists, enforced formats, cross-question consistency checks, double entry for critical fields.
Measure to improve
A data-quality dashboard shared with the teams who produce the data changes behaviour more surely than a memo. What gets measured gets better.
