We have the answers to your questions! - Don't miss our next open house about the data universe!

Kimball Method: What is it? How do I use it?

- Reading Time: 4 minutes
Kimball Method: What is it? How do I use it?

The volume of data generated and used by companies continues to grow exponentially. In this context, the efficient structuring and organization of this data for optimal analysis is of paramount importance.

Dimensional modeling is an inescapable solution to this challenge. It is a methodological approach to the design of data warehouses.

Ralph Kimball, a pioneer and expert in business intelligence, established standards and principles that today guide many organizations in the design and operation of their information systems.

Kimball's fundamental principles

The Kimball method is based on a series of key principles that define how data should be structured and organized to facilitate its analysis and exploitation.

These principles form the basis of dimensional modeling, providing a clear and systematic framework:

1. Tables et dimensions

  • Definition : Dimension tables contain the descriptive attributes of the data. They provide the context needed to understand and interpret the quantitative measures contained in the fact tables.
  • Main features :
    • Textual and descriptive attributes.
    • Often denormalized to optimize query performance and simplicity.
    • May contain hierarchies to facilitate analysis at different levels of granularity.

2. Tables de faits

  • Definition : Fact tables store quantitative or metric measurements that are usually the result of a transaction or event.
  • Main features:
    • Contain metrics such as sales, quantity, cost, etc.
    • Linked to dimension tables via foreign keys.
    • May include composite keys to uniquely identify a record.

3. Granularity

  • Definition : Granularity refers to the level of detail or summary of the data stored in the fact table.
  • Importance :
    • Determining granularity is crucial as it influences how data is collected, stored and analyzed.
    • It must be defined according to business needs and the questions for which the Data Warehouse is supposed to have the answers.

4. Standardization vs. denormalization

  • Normalization : The process of structuring data to reduce redundancy and improve integrity. It is often used in transactional database management systems.
  • Denormalization : The process of structuring data to improve query performance, often at the expense of redundancy. It is favored in dimensional modeling to facilitate data analysis.

Kimball’s method, with its principles of dimension and fact tables, provides a solid framework for data warehouse design.

By understanding and applying these basic principles, organizations can create information systems that are robust, flexible and optimized for analysis.

Advantages of the Kimball Method

Kimball’s dimensional modeling approach didn’t just happen to make its way into data warehousing. Its distinctive advantages make it a preferred approach for many organizations.

Image Optimal Performance for Queries Even on large volumes of data, this method facilitates quick queries. Thus, end users experience better performance with reduced wait times when generating reports or dashboards.
Image Flexibility Dimensional modeling allows adding new dimensions or facts without disrupting the existing structure. This facilitates updates and evolutions of the Data Warehouse as business needs change.
Image Ease of Understanding Clear separation between dimensions (context) and facts (measures) makes the Data Warehouse intuitive and easy to understand, even for non-specialists.
Image Scalability Dimensional structure is designed to handle increasing data volumes without compromising performance.
Image Data Consistency and Integrity Dimensional modeling, with its clear structure, promotes better data quality and consistency. Inconsistencies or anomalies are more easily identifiable.
Image Cost Reduction While initial setup may require investment, the ease of maintenance, scalability, and flexibility of the Kimball method can lead to significant long-term savings.

Kimball design process

The successful implementation of a data warehouse relies heavily on rigorous, methodical design. Kimball’s method offers a structured process to guide designers through the essential stages of this complex task.

Kimball contre Inmon

Two iconic figures dominate the world of data warehousing: Ralph Kimball and William Inmon. These two experts have each proposed distinct approaches to data warehouse modeling and design.

Principle Kimball Inmon
Image Philosophical Foundations Its approach is business process-oriented. It aims to build the data warehouse incrementally, starting with areas that bring the most value to the business. It envisions a centralized and holistic enterprise data warehouse. It advocates building a large standardized data warehouse first, followed by creating derived data marts for specific needs.
Image Architecture Favors a bottom-up approach, usually starting with the creation of data marts to address specific needs, which can then be integrated into a broader data warehouse. Favors a top-down approach, building a large enterprise data warehouse first, then deriving data marts for specific applications.
Image Modeling Dimensional Modeling: Data is organized into fact and dimension tables, as addressed in this article. 3NF Modeling (Third Normal Form): for the central data warehouse, ensuring maximum integrity and consistency.
Image Data Loading Process The ETL process is usually straightforward, feeding data from the source system directly into data marts or the warehouse. Data is first loaded into the central data warehouse, then an ELT process is used to feed derived data marts.
Image Flexibility and Consistency The Kimball method offers quicker implementation and greater flexibility to meet changing needs (may require more effort to ensure consistency across different data marts). With its holistic approach, Inmon ensures maximum data consistency across the enterprise (initial implementation may be longer and more costly).

Kimball and Inmon offer two different perspectives on the design and implementation of data warehouses. The choice between these approaches will depend on specific needs, available resources, and the company’s strategic objectives. Understanding the nuances of each method is essential to making an informed decision about which approach is best suited to a given situation.


The Kimball method, with its solid principles of dimensional modeling, offers a valuable framework for organizations seeking to optimize the efficiency, performance and flexibility of their information systems.

However, like any methodology, it is not a one-size-fits-all solution.

Companies need to carefully assess their specific needs, resources and long-term objectives to choose the approach best suited to their context.

You are not available?

Leave us your e-mail, so that we can send you your new articles when they are published!
icon newsletter


Get monthly insider insights from experts directly in your mailbox