Agile Business Intelligence
Scott Ambler, author of Refactoring Databases: Evolutionary Database Design, writing for Cutter IT Journal has written about how Agile methods can be adapted for Data Warehousing and Business Intelligence systems. Scott and his co-authors are writing for an audience who's users are suffering from large centrally designed systems.
In Scott's opinion mainstream Data Warehousing projects often fall short of user's expectations:
End users commonly complain that the delivered BI capability does not keep
pace with rapidly evolving business requirements and that the BI development team does not engage effectively with the user community. Requirements are collected at the beginning of the process, then the BI development team defines an architecture, general design, and detailed design based on the initial understanding of the requirements. The BI team often recognizes that user involvement is important through these development phases, but this involvement is very limited once the project starts.
Beyond the adoption of an Agile approach Scott has a few other key recommendations:
- Apply Agile Data techniques: Evolutionary data modelling, Database refactoring, Database regression testing, Continuous Database Integration and Reuse of existing datamodels.
- Recognize the Limits of a "Single Version of the Truth":
- Adopt SOA and Web 2.0 Strategies: consider the Data Federation and Enterprise architectural integration patterns to reduce the Total Cost of Ownership.
- Adopt a Lean Data Governance Strategy: Use business-driven project pipelines, Implement flexible architectures and Set risk-based milestones.
It is important to recognize that there can be several truths and to identify those truths, .... The driving principle should not be purity of the definition but the timely delivery of important business value.
To read more visit the free download of the Cutter IT Journal and enter the promotion code BUSINESSINTELLIGENCE.