Seven aspects of a successful business intelligence (BI) architecture

What is required for a successful business intelligence (BI) architecture?

A successful BI architecture needs to:

• Support the various analytical and reporting styles of the business executives and other workers who are using the BI system. That may include reporting, ad hoc querying, online analytical processing (OLAP), data mining, predictive analytics, data visualization and, of

  More on business intelligence architecture

  Read about the factors to consider in planning an analytics big data architecture

  Find out what new options a BI architect can use to meet growing data demands

  Learn how faster data delivery improves BI performance and user satisfaction

  course, spreadsheet integration.
• Provide access to the level of comprehensive, consistent, clean and current data that the business needs for analysis – and is willing to pay for.
• Systematically implement and support data governance processes.
• Separate data integration from BI reporting and analysis processes.
• Implement the staging of data from systems of record through a data warehouse to data marts or cubes – i.e., a hub-and-spoke approach.
• Create a business-oriented “semantic layer” to give users a view of the information available for reporting and analytics.
• Create a “live” data and reporting dictionary that business users can use while accessing reports or as a table of contents (with definitions) for the data available to them.

You will be able to design a high-level BI architecture early in your project, but it later will need to be designed in greater detail and built out incrementally and iteratively if you are going to provide a robust BI environment.

Note: Don’t believe vendor “marketectures” that promise you an integrated and robust BI architecture simply by installing their software. Set realistic expectations, both for the business and IT. Implementing a successful BI architecture is a journey, not a one-time event.