Administration, Governance, and Troubleshooting
Core mental model
Administration covers users, roles, and permissions (product profiles / access control), sandboxes for environment isolation, and APIs for programmatic access. Governance covers data usage policies (DULE), consent, and audit trails. Troubleshooting is systematic: isolate the layer (data → profile → segment → journey → channel → delivery).
Concepts candidates must know
- Access control: product profiles map to permissions; sandbox-level role assignments; API credentials (client ID/secret, JWT/OAuth).
- Sandboxes: isolation of datasets/schemas/journeys; reset/delete behavior; production vs development.
- Governance: data usage labels, policies, and enforcement; consent fields in XDM and how they gate channel sends.
- Monitoring: journey reports, message feedback, and alerts for failures.
- Troubleshooting: schema/identity mismatches, merge policy issues, unpublished journeys, invalid channel configs, and personalization errors.
Traps
- Assuming a user with journey permissions can also manage schemas - permissions are granular per product profile.
- Forgetting that a journey must be published to run; drafts don't execute.
- Overlooking consent and data usage policies as causes of blocked sends - governance can suppress delivery even when configuration is correct.
- Confusing sandbox isolation with data sharing - data does not automatically cross sandboxes.
Worked rule of thumb
Troubleshoot in pipeline order: (1) is data ingested and mapped to the schema? (2) does the identity graph resolve to one profile? (3) does the segment evaluate correctly? (4) is the journey published and does the profile enter? (5) is the channel configuration valid and consent satisfied? The first failing layer is your root cause.