ISO/IEC 27001:2022 Annex A · Control 8.11
Data Masking: A Practical Implementation Guide
Reduce exposure by replacing sensitive values while preserving necessary business or testing use.
This control concerns using data masking in accordance with access-control policy, business requirements and law.
What should the control achieve?
- Masking use cases and sensitive fields are identified.
- Methods match risk and utility.
- Keys and mapping tables are protected.
- Re-identification and bypass are tested.
Step-by-step implementation
Identify use cases
Consider support screens, analytics, testing, training, reporting and external sharing.
Choose technique
Use redaction, substitution, tokenization, perturbation or dynamic masking.
Define visibility rules
Map roles and purposes to full, partial or masked values.
Protect reversibility
Secure keys, token vaults and mapping tables separately.
Test utility and privacy
Validate business function, uniqueness, consistency and re-identification risk.
Monitor access
Log unmasking and review unusual patterns.
What this could look like in practice
Support agents see only the last four digits of payment identifiers. A small approved fraud team can temporarily unmask records with MFA, reason capture and logging. Test data uses irreversible substitution.
| Activity | Practical implementation | Evidence |
|---|---|---|
| User interface | Dynamic masking hides fields by role. | Configuration |
| Test data | Production values are substituted before export. | Masking job log |
| Analytics | Pseudonymous identifiers preserve linkage. | Design record |
| Unmasking | Approval and audit trail govern full view. | Access log |
Implementation evidence
- Masking standard
- Field inventory
- Technique decisions
- Role rules
- Key protection
- Masking job records
- Re-identification tests
- Unmasking logs
Useful metrics
- Sensitive fields masked in non-production
- Unmasking events reviewed
- Masking failures
- Re-identification findings
Common mistakes
- Using simple character replacement that is reversible.
- Keeping lookup tables beside masked data.
- Masking direct identifiers but not combinations.
- Using production data in testing by default.
- Failing to log privileged unmasking.
Questions an auditor may ask
- Where is masking required?
- Why was this technique chosen?
- How is reversibility protected?
- Show a re-identification test.
Continue through Annex A
Explore the growing library of practical guides for all 93 Annex A controls.
Open the ISO 27001 Annex A Control LibraryThis independent educational guide paraphrases the practical intent of the control and does not replace the official ISO standards, professional advice or an organization-specific risk assessment. It is not affiliated with or endorsed by ISO.