Canonical ID

Once assigned, the Canonical ID must never change. It survives email updates, device resets, and address changes. This permanence is what allows historical behavioral data to remain linked to the active profile, preventing data loss during identity transitions.

• Survivorship: Outlasts transient identifiers like cookies or session IDs.
• Historical Integrity: Maintains a continuous audit trail of all past interactions.
• System of Record: Acts as the foreign key in all downstream data warehouses.

During entity resolution, conflicting records are merged. The Canonical ID represents the winning record after applying survivorship rules. It is typically anchored to the highest-confidence deterministic match, such as a hashed email key or a verified phone number.

• Trust Hierarchy: Prioritizes authenticated login events over probabilistic device matches.
• Conflict Resolution: Applies business logic to select the most recent or most frequent attribute value.
• Golden Record: The Canonical ID is the primary key for the Golden Record.

The Canonical ID serves as the technical enforcement point for privacy compliance. Consent states, data deletion requests, and right-to-forget commands are bound to this single identifier, ensuring that a user's preferences cascade instantly across all linked devices and systems.

• Consent Propagation: Syndicates opt-out signals to all integrated vendors.
• Deletion Hard Key: Ensures complete data purging across all tables via a single cascade.
• Audit Token: Used to generate compliance reports for GDPR and CCPA requests.

The Canonical ID is the universal join key that links the Customer Data Platform (CDP) to the email service provider, the mobile app backend, and the CRM. It decouples personalization logic from raw, messy source identifiers.

• Abstraction Layer: Downstream systems reference the Canonical ID instead of raw emails or device IDs.
• Synchronization: Enables real-time profile updates across the martech stack.
• Vendor Neutrality: Prevents vendor lock-in by owning the identity spine internally.

In modern architectures, the Canonical ID is often a node in an Identity Graph. It is generated by resolving a complex web of edges (login events, household IPs, device fingerprints) into a single, compressed node representing the human entity.

• Entity Resolution Output: The final product of deterministic and probabilistic matching algorithms.
• Node Compression: Collapses multiple device nodes into a single master node.
• Graph Traversal: Allows querying all linked devices via a single hop from the Canonical ID.

The Canonical ID bridges the gap between real-time operational systems and batch analytical processing. It ensures that the user receiving a personalized offer in-session is the exact same entity being analyzed in the data lake.

• OLTP to OLAP: Unifies identity across row-based databases and columnar warehouses.
• Session Stitching: Links real-time clickstreams to historical purchase history.
• Attribution Accuracy: Prevents over-counting unique users in marketing reports.