Real-Time CDP Implementation Guide

Adobe Real-Time Customer Data Platform (RT-CDP) enables organizations to collect, unify, and activate customer data in real-time. This implementation guide walks through the key steps, best practices, and considerations for a successful RT-CDP deployment.

Understanding Real-Time CDP

Real-Time CDP is an application built on Adobe Experience Platform that provides:

• Profile Unification: Merge data from multiple sources into unified customer profiles
• Real-Time Processing: Update profiles within milliseconds of data arrival
• Audience Building: Create dynamic segments using any profile attribute
• Cross-Channel Activation: Send audiences to 100+ destinations
• Privacy Controls: Built-in consent and governance management

Implementation Phases

Phase 1: Discovery and Planning

Success starts with thorough planning:

Define Use Cases

Identify specific business outcomes you want to achieve:

• Suppress existing customers from acquisition campaigns
• Create look-alike audiences for prospecting
• Personalize website experiences based on CRM data
• Coordinate messaging across email, ads, and site
• Build predictive audiences using AI models

Data Source Inventory

Document all data sources to be integrated:

• Web analytics (Adobe Analytics, GA)
• CRM systems (Salesforce, Dynamics)
• Email platforms (Campaign, Marketo)
• E-commerce platforms
• Point-of-sale systems
• Mobile apps
• Offline data files

Identity Strategy

Determine how customers will be identified:

• What IDs are available (email, phone, CRM ID)?
• How will anonymous visitors be tracked (ECID)?
• What linking rules should apply?
• How will privacy requirements affect identity?

Phase 2: Schema Design

XDM schemas define your data structure. Good schema design is critical for long-term success.

Profile Schema

Store person-level attributes:

// Example Profile Schema Fields
personalEmail.address      // Primary email
person.name.firstName      // First name
person.name.lastName       // Last name
homeAddress.postalCode    // Location
loyaltyDetails.tier       // Loyalty status (custom)
lifetimeValue.amount      // LTV (custom)
preferredChannel          // Communication preference (custom)

Event Schema

Capture time-series behaviors:

// Example Event Schema Fields
timestamp                  // When event occurred
eventType                 // Type (pageView, purchase, etc.)
web.webPageDetails.URL    // Page viewed
productListItems[]        // Products in cart/order
commerce.order.priceTotal // Order value
marketing.campaignGroup   // Campaign attribution

Schema Best Practices

• Start with standard Adobe field groups
• Create custom field groups for organization-specific data
• Use consistent naming conventions
• Document field definitions and expected values
• Plan for future extensibility

Phase 3: Data Integration

Source Connectors

Use pre-built connectors for common systems:

• Adobe Sources: Analytics, Campaign, Audience Manager
• CRM Sources: Salesforce, Microsoft Dynamics
• Cloud Storage: S3, Azure Blob, Google Cloud Storage
• Databases: Snowflake, BigQuery, Redshift
• Marketing: Mailchimp, HubSpot

Web SDK Implementation

Collect web data using the Experience Platform Web SDK:

// Initialize Web SDK
alloy("configure", {
  "edgeConfigId": "your-datastream-id",
  "orgId": "your-org-id@AdobeOrg",
  "defaultConsent": "pending"
});

// Track page view
alloy("sendEvent", {
  "xdm": {
    "eventType": "web.webpagedetails.pageViews",
    "web": {
      "webPageDetails": {
        "URL": window.location.href,
        "name": document.title
      }
    }
  }
});

Batch Ingestion

Import historical and offline data via files:

1. Format data according to XDM schema
2. Save as Parquet (preferred), JSON, or CSV
3. Upload to designated cloud storage
4. Create dataflow to ingest on schedule

Phase 4: Profile Configuration

Identity Configuration

Set up identity namespaces and linking:

• Define custom namespaces for your IDs
• Configure namespace priority
• Set up identity graph rules
• Test identity linking behavior

Merge Policies

Configure how profiles from different sources combine:

• Dataset Precedence: Which source wins for conflicting fields
• Timestamp Ordering: Most recent value wins
• ID Stitching: How related identities merge

Computed Attributes

Create derived profile values:

• Count of purchases in last 30 days
• Days since last engagement
• Total lifetime revenue
• Most frequent product category

Phase 5: Audience Building

Segment Builder

Create audiences using the visual segment builder:

• Profile Attributes: Demographics, preferences, scores
• Events: Behaviors within time windows
• Existing Segments: Include/exclude other audiences
• AI Predictions: Propensity scores from Customer AI

Evaluation Methods

• Streaming: Real-time evaluation as profiles update
• Batch: Scheduled evaluation (daily, hourly)
• Edge: Evaluation at the edge for millisecond personalization

Example Audiences

• High-value customers who haven't purchased in 90 days
• Cart abandoners with items over $100
• New subscribers interested in specific category
• Customers with high churn propensity score

Phase 6: Activation

Destination Types

• Advertising: Google, Facebook, LinkedIn, DV360
• Email/CRM: Salesforce, Oracle, SAP
• Personalization: Adobe Target, web personalization
• Analytics: Export for analysis and reporting
• Cloud Storage: Data warehouse export

Activation Configuration

1. Connect destination (OAuth, API key, etc.)
2. Map profile fields to destination schema
3. Select audiences to activate
4. Configure scheduling (real-time or batch)
5. Apply consent policies

Data Governance

Data Labels

Apply labels to sensitive data:

• C1: Cannot be exported to third parties
• C2: Cannot be used for on-site targeting
• S1: Sensitive geographic data
• S2: Sensitive health data

Usage Policies

Create policies that restrict how labeled data can be used:

• Prevent PII from going to advertising destinations
• Require consent for email activation
• Restrict health data to specific use cases

Testing and Validation

Pre-Launch Checklist

• Verify data flowing to correct schemas
• Confirm identities linking properly
• Test segment qualification logic
• Validate destination data format
• Review governance policy enforcement
• Test consent preference handling

Ongoing Monitoring

• Monitor ingestion success rates
• Track profile merge rates
• Review audience size trends
• Audit destination delivery

Conclusion

Implementing Real-Time CDP requires careful planning across data, identity, governance, and activation dimensions. The investment pays off with unified customer views, real-time responsiveness, and coordinated cross-channel experiences. Start with focused use cases, prove value quickly, and expand from there.

Need expert guidance for your RT-CDP implementation? Contact WegWorks for proven implementation expertise.