Adobe Target Architect Master study materials
Adobe Target Architect Master study resources

Data Governance Conflict in Cross-Device Personalization

Adobe Target Architect Master · Put your knowledge to work in context.

All case studies

Data Governance Conflict in Cross-Device Personalization

A healthcare provider uses Adobe Target to personalize appointment scheduling messages across web and mobile app. The personalization includes condition-specific content (e.g., diabetes, cardiology) based on pages visited. The provider uses Adobe Experience Platform (AEP) to unify patient profiles and Adobe Target for delivery. The AEP dataset includes sensitive health information (PHI) such as diagnosis codes, which are tagged as 'Restricted' in AEP's data governance framework. The Target implementation uses the Adobe Experience Platform Web SDK to pass an 'industry' parameter to Target for personalization. The provider's compliance team requires that PHI not be used for marketing without explicit consent and that all data transfers be encrypted and auditable. The Target activity uses a custom audience built from AEP segments that include PHI. The provider has a consent management platform (CMP) that captures patient consent for marketing use. The CMP is integrated with AEP, but the consent status is not passed to Target. The provider wants to expand personalization to include condition-specific offers but must ensure compliance with HIPAA and internal policies. The Target implementation uses at.js 2.x with the Adobe Experience Cloud ID Service, and the mobile app uses the Adobe Experience Platform Mobile SDK.

1. What are the data governance risks in the current implementation, and how should the architect address them?

Read the worked answer

Risks: 1) PHI is used in Target audiences without verifying consent, violating HIPAA and internal policy. 2) The 'industry' parameter may inadvertently expose PHI if it maps to a diagnosis. 3) Consent status is not passed to Target, so personalization may occur without consent. 4) Data transfers may not be encrypted end-to-end if not configured properly. 5) Lack of audit trail for PHI usage in Target. Address: 1) Classify the 'industry' parameter: if it can be linked to an individual and reveals health condition, it is PHI and must be protected. Replace it with a non-PHI attribute (e.g., 'interest category' derived from consent-based data). 2) Integrate the CMP with Target: pass consent status as a parameter and use Target's consent API or Adobe Experience Platform's consent policies to gate personalization. 3) Use AEP's data governance labels to enforce usage policies: ensure that segments used in Target exclude PHI or have proper consent. 4) Encrypt all data in transit and at rest; use HTTPS and ensure Target and AEP are configured for secure data transfer. 5) Implement auditing: log all personalization decisions and data access; use AEP's audit logs. 6) Conduct a privacy impact assessment. 7) Train staff on HIPAA and data governance.

2. How can the architect enable condition-specific personalization while complying with HIPAA and consent requirements?

Read the worked answer

Enable by: 1) Obtain explicit consent for marketing use of health data; capture via CMP and pass to Target. 2) Use de-identified or aggregated data for personalization: instead of PHI, use derived non-PHI attributes (e.g., 'cardiology interest' based on consented browsing behavior) that do not reveal diagnosis. 3) Implement a consent-based audience: only include users who have opted in. Use AEP's consent policies to filter segments. 4) For users who have not consented, show generic content. 5) Use Target's experience targeting with a fallback for non-consented users. 6) Ensure that the personalization logic does not store PHI in Target: use ephemeral parameters or server-side decisions via AEP Edge Network. 7) Conduct regular audits and provide transparency to patients. 8) Work with legal to define what constitutes PHI and ensure all data flows are compliant. 9) Use Adobe Experience Platform's healthcare-specific data governance features if available.