Lesson 4 — Business Rules, DQ Rules, Metrics & Thresholds
The exam often places four related layers in one stem. Separate them mentally.
The chain
BUSINESS EXPECTATION → DIMENSION → DQ RULE → MEASURE/METRIC → THRESHOLD → STATUS/ACTION
Business Rule
Describes how data or operations should behave for useful, successful, or compliant business activity.
Data Quality Rule
Turns a relevant expectation into a testable, dimension-aligned condition.
Example: “Every U.S. customer must have ZIP code.” This is a completeness rule.
Rules can operate at: - element level; - record level; - data-set level; - cross-data-set/process level.
Managed rule metadata should include meaning, owner/steward, source/authority, dimension, scope, measurement method, threshold, implementation location, version/change history, and breach response.
Metric / Measurement
The observed conformance or exception result.
Examples: - 97.5% of required ZIP codes are populated — positive conformance result. - 2.5% violate the rule — exception result.
Threshold
The cutoff that interprets whether the metric is acceptable.
“At least 99% complete” contains an acceptance threshold of 99%.
A raw score without a threshold cannot tell you whether quality is sufficient for the intended use.
Effective metric characteristics
A useful management metric is: - measurable; - business relevant; - tied to an acceptance threshold; - accountable/stewarded; - controllable/actionable; - trendable.
If nobody can act on the metric, it is weak management evidence.
Exam trap
Do not call 97.4% conformance the rule. That is the measured result. The rule is the test; the threshold decides pass/fail.
Source: pp. 432–434, 453–456.