Data Vault Modelling

In June 2019, Neil Strange delivered a webinar on Data Vault Data Modelling, sharing practical strategies and foundational concepts that every data professional should know.

Here’s a summary of the key insights and actionable takeaways from that session.

What Is Data Vault Data Modelling?

Data Vault is a methodology for building scalable, flexible, and business-focused data warehouses.

According to the Data Vault Alliance, it’s “a system of business intelligence containing the necessary components needed to accomplish enterprise vision in data warehousing and information delivery.”

The goal is to create a practical data warehouse solution that recognizes the realities of source systems while breaking free from their limitations.

The Core Challenge

• Achieve a practical, business-driven data warehouse

• Recognize and manage the complexities of source systems

• Serve data to business users using clear business concepts

A crucial warning from Data Vault inventor Dan Linstedt: If your Data Vault isn’t about the business, it’s time to rethink your approach.

Enterprise Scope and Business Focus

Effective Data Vault modelling requires:

• Enterprise-wide scope

• A strong business focus

• Coping with source system realities

• Top-down and bottom-up data modelling

• Glossaries, business terms, and source system mapping

• Synthesis of business and technical perspectives

The Building Blocks of Data Vault

Data Vault uses three standard building blocks, each with a specific role:

1. Hubs

• Represent business entities (e.g., Customer, Booking)

• Store unique business keys and metadata (when and where loaded)

2. Links

• Connect Hubs, recording relationships or transactions

• Contain foreign keys and metadata

3. Satellites

• Attach to Hubs or Links

• Store descriptive data, history, and metadata (including effectivity dates)

• Enable full historical tracking (“what did we know when?”)

Standardizing these components enables repeatable ELT patterns and simplifies data integration.

Advanced Data Vault Concepts

• Transactional Links: Capture high-volume, immutable transactional data

• Effectivity Satellites: Track the validity or deletion of Hubs or Links

• Reference Data: Behaves like a Hub for consistent reference management

Practical Example: From Source Table to Data Vault

The webinar walked through unpicking a complex source table (e.g., Booking) and mapping its fields to Hubs, Links, and Satellites. This process involves:

• Identifying business entities and relationships

• Creating Hubs for core entities (Customer, Booking, Brand, Product, etc.)

• Building Links to represent relationships (e.g., Customer-Booking)

• Designing Satellites for descriptive and historical data

• Using hash keys and hashdiffs for efficient change detection and historical tracking

Each table type follows a standard load pattern, making automation and scalability achievable.

Key Takeaways for Data Vault Success

• Always align your Data Vault with business needs and concepts

• Use standard building blocks for consistency and automation

• Combine top-down business analysis with bottom-up source mapping

• Leverage metadata, hash keys, and effectivity tracking for robust historical data

• Build glossaries and business data dictionaries to support clarity and governance

Conclusion

Data Vault data modelling is essential for organizations aiming to build agile, scalable, and business-aligned data warehouses. By following best practices and leveraging standard patterns, you can ensure your data warehouse delivers value, flexibility, and insight for years to come.