# Real-Time 2.0

Treasure Data's Real-Time 2.0 provides the only solution in the market that unifies batch and real-time data at the profile level, ensuring that the experiences you deliver to your customers are the right ones and delivered at the right time.

Real-Time 2.0 allows multiple departments in your organization to communicate with current and prospective customers by unifying batch and real-time data in one CDP.

## What is Real-Time 2.0?

Real-Time 2.0 is Treasure Data's next-generation Customer Data Platform that unifies batch and real-time data processing in a single system. Unlike traditional CDPs that force you to choose between batch analytics or real-time activation, Real-Time 2.0 combines both capabilities to deliver:

- **Millisecond-latency responses** for web personalization and triggered activations
- **Complete customer profiles** that include both historical batch data and real-time behavioral data
- **Unified decisioning** across all customer touchpoints using a single platform


Real-Time 2.0 enables marketers to respond to customer behavior instantly while leveraging the full depth of historical customer data for smarter, more accurate personalization.

## Key Capabilities

Real-Time 2.0 provides three core capabilities that work together to deliver exceptional customer experiences:

### Real-Time ID Stitching

Unify customer identities in real-time across devices, channels, and sessions. When a customer visits your website or opens your mobile app, Real-Time 2.0 immediately stitches their identity with existing profiles, creating a consistent experience across all touchpoints.

**Value:** Recognize returning customers instantly and personalize their experience from the first page view.

**Implementation:** Real-Time ID Stitching is enabled through Parent Segment Configuration, where you define ID stitching keys, event tables, and attributes that power real-time identity resolution.

**Learn more:** [Setting Up Parent Segments](/products/customer-data-platform/real-time/creating-a-real-time-configuration)

### Real-Time Triggering

Respond to customer actions within minutes by triggering cross-channel activations. When a customer abandons their cart, visits a physical store, or reaches a loyalty milestone, Real-Time 2.0 can automatically trigger email, SMS, push notifications, or other activations.

**Value:** Engage customers at the exact right moment to drive conversions and build loyalty.

**Learn more:** [Real-Time Journeys](/products/customer-data-platform/journey-orchestration/realtime) | [Triggered Activations](/products/customer-data-platform/journey-orchestration/realtime/creating-a-real-time-triggered-activation)

### Real-Time Personalization

Deliver personalized web and mobile experiences in under 100 milliseconds. Real-Time 2.0 combines in-session behavior with historical customer data to power personalization engines, content management systems, and A/B testing tools.

**Value:** Provide relevant, personalized experiences that increase engagement and revenue.

**Learn more:** [Real-Time Personalization](/products/customer-data-platform/real-time/about-real-time-personalization)

![](/assets/real-time-2-0-2024-05-22.f4c3a852da2e4e84fa429ddbe84f351ef520b5cf3fbd39d26fc4956dd5bf6e4c.cf3845cb.png)

## How Events Flow

![Real-Time 2.0 Architecture Diagram](/assets/rt-2-0-architecture-diagram.2eee574433154ebc250b4d3f26e3ae824bd47866c5a899169693374b3ea683af.5b2238d2.svg)

Real-Time 2.0 uses a dual-path architecture to process every customer event simultaneously—one path for immediate decisions, one for deep analytics. Here's how a single event travels through the system:

### Event Collection

When a customer interacts with your website or mobile app, the Treasure Data SDK or Ingest API captures the event and delivers it to the **Ingest Stream**. This is the entry point where the dual-path processing begins.

### Dual-Path Processing

Every event is processed by two pipelines running in parallel:

**Realtime Pipeline**

The Realtime Decision Engine processes each event in milliseconds and performs three operations simultaneously:

- **ID Stitching** – Links the current user to their existing profile by matching anonymous IDs, cookie IDs, and known customer IDs across devices and sessions
- **Real-Time Attributes Update** – Updates in-session counters and behavioral attributes (for example, number of product views, cart value, current session duration)
- **Unify Realtime and Batch Data** – Merges the live event data with the customer's historical profile from Batch Storage to produce a complete profile for decisioning


The enriched profile is written to **Realtime Storage** and immediately available for **web personalization** and **triggered activations**.

**Batch Pipeline**

TD Workflows run on a scheduled basis (typically every few minutes to hours) to:

- Aggregate attributes across large event histories
- Calculate complex metrics and predictive scores
- Build and refresh ID relationships across the full dataset


Results are written to **Batch Storage** and are available for audience segmentation and reporting.

### How They Work Together

The "Unify Realtime and Batch Data" step in the Realtime Pipeline is where the two paths converge. When the Realtime Decision Engine processes a new event, it reads the customer's most recent batch profile—including historical purchase history, loyalty tier, and predictive scores—and combines it with the live in-session data before making a personalization or activation decision.

This means every real-time decision benefits from the full depth of historical data, not just what happened in the current session.

### Example in Action

A returning customer opens your website. Here's what happens in under 100 milliseconds:

1. The TD SDK fires a `page_view` event to the Ingest Stream
2. The Realtime Decision Engine receives the event and matches the anonymous cookie ID to the customer's known profile via ID Stitching
3. Real-time attributes are updated: page views this session = 3, product category = "yoga"
4. The engine reads from Batch Storage: loyalty tier = Gold, last purchase = 45 days ago, predicted next best offer = "Premium Yoga Mat"
5. Unified profile is written to Realtime Storage
6. The Personalization API returns a targeted recommendation: the Premium Yoga Mat with a Gold member discount


Meanwhile, the Batch Pipeline logs the event for the next scheduled workflow run, which will update the customer's aggregate attributes and refresh their predictive scores for future sessions.

## Why Real-Time 2.0?

### Key Benefits

- **More accurate personalization:** Use complete customer profiles (batch + real-time data) for better targeting
- **Higher conversion rates:** Engage customers at the exact right time with relevant messaging
- **Unified customer experiences:** Create consistent experiences across devices, channels, and sessions
- **Increased operational efficiency:** Manage batch and real-time data in one platform instead of multiple tools
- **Greater customer loyalty:** Build trust by being communicative, accessible, and timely


### Business Impact

According to Harvard Business Review, companies with fast revenue growth generate 40% more revenue from personalized marketing tactics than slower-growth companies. When ads are based on real-time website activity:

- Click-through rates increase by 11%
- Time spent viewing products rises by 34%
- Revenue from products grows by 38%


### Unified Batch & Real-Time: The Treasure Data Advantage

Most CDPs force you to choose:

- **Batch-only CDPs** provide deep analytics but can't respond in real-time
- **Real-time-only CDPs** respond quickly but lack historical context


Real-Time 2.0 delivers both by unifying batch and real-time processing in one platform—simpler, more powerful, and less expensive than maintaining separate systems.

## Common Use Cases

### In-Session Web Personalization

Personalize website content based on current browsing behavior combined with historical purchase data, preferences, and predictive scores.

**Example:** A returning customer browsing yoga products sees personalized product recommendations and a targeted offer based on their preferred channel (in-store) and predicted next best offer.

### Abandoned Cart Recovery

Trigger personalized messages across email, SMS, and push notifications when a customer adds items to cart but doesn't complete the purchase.

**Example:** Send an email within minutes with a discount code, followed by an SMS reminder if the cart remains abandoned.

### Physical Store Visits

Use geofencing or check-in data to trigger mobile notifications when customers enter a store location.

**Example:** When a customer enters a store, send a push notification with a personalized discount on products they recently browsed online.

### Form Fill Follow-Up

Route customers to customer service or sales teams when they encounter issues during form submission or account creation.

**Example:** If a customer starts a subscription form but gets stuck, route them to a live agent via chat or phone to complete the signup.

### Loyalty Milestone Recognition

Trigger celebratory messages and special offers when customers reach loyalty tiers or milestones.

**Example:** Send a congratulatory email with an exclusive offer when a customer reaches VIP status.

## Getting Started with Real-Time 2.0

Ready to implement Real-Time 2.0? Follow these steps:

### 1. Learn the Basics

Start with the Getting Started guide to understand prerequisites, setup workflow, and configuration requirements.

- **[Getting Started Guide](/products/customer-data-platform/real-time/getting-started)**


### 2. Set Up Data Ingestion

Configure web tracking, mobile SDKs, or Ingestion API to capture customer events in real-time.

### 3. Configure Parent Segments

Set up Event Tables, Event Definitions, Attributes, and ID Stitching for your Parent Segment.

- **[Setting Up Parent Segments](/products/customer-data-platform/real-time/creating-a-real-time-configuration)**


### 4. Build Experiences

Create real-time journeys, triggered activations, and personalization rules.

- **[Real-Time Journeys](/products/customer-data-platform/journey-orchestration/realtime)**
- **[Real-Time Personalization](/products/customer-data-platform/real-time/about-real-time-personalization)**
- **[Real-Time Attributes](/products/customer-data-platform/real-time/real-time-attributes-overview)**


## Additional Resources

- [Getting Started Guide](/products/customer-data-platform/real-time/getting-started) - Prerequisites, setup workflow, and best practices
- [Performance Limitations and Expected Behavior](/products/customer-data-platform/real-time/real-time-2-0-system-performance-limitations-and-expected-behavior)
- [Observability and Logs](/products/customer-data-platform/real-time/real-time-observability-and-logs)
- [Real-Time FAQ](/products/customer-data-platform/real-time/real-time-faq)
- [Real-Time Attribute FAQs](/products/customer-data-platform/real-time/real-time-attribute-faqs)