Customer Data Platform (CDP)

The platform must be directly controllable by non-technical marketing and business users without ongoing IT or engineering intervention. This distinguishes a CDP from a raw data lake or enterprise data warehouse (EDW).

• No-Code Segmentation: Marketers build audience segments using a visual interface, not SQL queries.
• Direct Activation: Segments are syndicated to email, advertising, and personalization tools without custom API development.
• Governance Layer: While marketer-managed, the CDP enforces role-based access controls and data usage policies set by IT.

The CDP ingests first-party data from all sources and creates a golden record—a single, deduplicated profile that persists over time. This is the foundational identity spine.

• Identity Resolution: Merges anonymous and known identifiers (email, device IDs, loyalty numbers) into a canonical ID using deterministic and probabilistic matching.
• Schema Flexibility: Stores structured attributes, unstructured events, and behavioral signals without rigid pre-defined schemas.
• Historical Persistence: Maintains a complete, non-volatile timeline of customer interactions, not just a current snapshot.

A CDP is source-agnostic, designed to ingest and normalize data from any first-party channel without relying on third-party cookies or external data brokers.

• Source Types: Transactional systems (POS, e-commerce), web/mobile behavioral streams, email engagement, customer service logs, IoT telemetry, and offline CRM files.
• Ingestion Methods: Supports real-time streaming (via webhooks or Kafka), batch file uploads (CSV, Parquet), and direct API connectors to SaaS platforms.
• Privacy-Centric: Because data is first-party, the CDP architecture aligns with third-party cookie deprecation and builds a consented data moat.

The CDP must provide a native segmentation engine that allows users to define complex audience cohorts and push them to downstream tools for activation.

• Real-Time Segmentation: Evaluates streaming events against segment rules to qualify users instantly, not on a nightly batch schedule.
• Multi-Condition Logic: Supports nested boolean logic, recency-frequency-monetary (RFM) models, and predictive scoring as segmentation criteria.
• Syndication Protocols: Pushes segments to destinations via native connectors, webhooks, or SFTP, maintaining consistent suppression lists and frequency caps across channels.

A CDP is not a walled garden. It exposes its unified data and segmentation logic through robust, well-documented APIs, allowing engineering teams to build custom applications on top of the customer data layer.

• Profile API: Retrieve the full unified profile for a known customer in real-time to power personalization engines.
• Event Ingestion API: Accept custom event streams from proprietary or legacy systems not covered by native connectors.
• Webhook Destinations: Trigger external workflows and custom microservices when a user enters or exits a segment.

The CDP serves as the central enforcement point for customer consent and data governance. It must ingest consent signals and apply them universally to all downstream activations.

• Consent Ingestion: Captures granular opt-in/opt-out signals from a Consent Management Platform (CMP) or directly via Global Privacy Control (GPC) headers.
• Policy Enforcement: Automatically suppresses users from segments or destinations when consent is withdrawn, ensuring compliance with GDPR and CCPA.
• Data Lineage: Maintains an immutable audit trail of consent changes and data usage for regulatory reporting.

A centralized data structure that links all known identifiers—such as email addresses, device IDs, and usernames—to a single unified customer profile. The identity graph forms the deterministic backbone of the CDP, enabling the platform to recognize a user across sessions and channels. Without a robust identity graph, the CDP cannot perform its core function of creating a persistent, 360-degree view of the customer.

The definitive, best-version-of-the-truth customer profile created by applying survivorship rules to conflicting attributes from multiple source systems. When a CDP ingests data from a CRM, a loyalty app, and a POS system, it must arbitrate discrepancies—such as two different last names for the same canonical ID—to produce a single, trusted record that downstream engagement tools can rely on with confidence.

A method of identity resolution that relies on exact, verified matches of personally identifiable information (PII) to link user activity across devices with absolute certainty. Common deterministic keys include a hashed email key or a login credential. This approach provides high precision but limited reach, as it requires the user to authenticate. CDPs prioritize deterministic matches to anchor their identity spines.

A statistical approach to identity resolution that uses non-personal signals like IP address, browser type, and behavioral patterns to infer device ownership. Unlike deterministic matching, it assigns a confidence score rather than a definitive link. CDPs use probabilistic models to expand the reach of their identity graphs, stitching anonymous sessions to known profiles when exact identifiers are unavailable.

A secure, neutral environment where multiple parties can combine and analyze first-party data sets for identity resolution and attribution without exposing raw, user-level data to external stakeholders. CDPs often integrate with clean rooms to enable privacy-safe second-party data enrichment, allowing brands to match their customer base against a publisher's or partner's audience without ever moving or revealing the underlying PII.

A technology that captures, stores, and syndicates a user's granular privacy choices to downstream vendors. The CMP ensures that the CDP's identity resolution and tracking logic respects the specified legal basis for processing. A tight integration between the CDP and CMP is critical: if a user withdraws consent, the CDP must immediately suppress that profile from activation, enforcing the Global Privacy Control (GPC) signal.