CDP Data Quality Discipline

Data quality typically gets 5 - 15% of CDP investment but determines 60 - 80% of the marketing outcome. This module covers the four-dimension framework and the governance that maintains quality.

1. Data quality leverage

A CDP with poor data quality produces poor marketing outcomes regardless of how sophisticated the rest of the stack. Data quality typically receives 5 - 15% of CDP investment but determines 60 - 80% of the marketing outcome.

2. Four dimensions

3. Source vs resolved

Quality must be enforced both at source (event tracking, form validation, integration accuracy) and after resolution (deduplication, normalization, merging). Source quality is cheaper than after-the-fact cleanup.

4. Event schema discipline

• Standard event names (page_viewed, product_added, purchase_completed).
• Standard property names and types.
• Versioning for schema changes.
• Documentation accessible to engineering and analytics.
• Schema validation in the data pipeline.

5. Customer record deduplication

• Exact match on email or phone.
• Fuzzy match on name + address.
• Probabilistic match with confidence threshold.
• Manual review queue for uncertain matches.
• Audit trail for merge / split actions.

6. Address normalization

USPS address standardization (Smarty Streets, Lob, Loqate, Melissa). Normalized addresses enable better matching and reduce duplicate records.

7. Email and phone validation

• Email syntax validation at collection.
• Email deliverability validation (NeverBounce, ZeroBounce, Kickbox).
• Phone format normalization (E.164).
• Phone-line-type detection.

8. Suppression and lifecycle

• Unsubscribed.
• Bounced / undeliverable.
• Spam-flagged.
• Compliance-suppressed (GDPR / CCPA opt-out).
• Deceased.
• Test / internal.

9. Data quality testing

• Continuous validation rules (dbt tests, Great Expectations, Soda).
• Alerts on anomalies.
• Dashboards for key quality metrics.
• Regression testing on schema changes.

10. Governance

Who owns data quality? Typically a data steward role works with marketing operations, analytics, and engineering. Without clear ownership, quality degrades.

11. Annual audit

1. Inventory of sources.
2. Schema compliance review.
3. Sample-based accuracy audit.
4. Duplicate-record analysis.
5. Suppression-list audit.
6. Improvement plan.

Part of the Customer Data Platforms series · RGM Training