Agile Development in Data Warehousing

Traditional data warehouse projects follow a waterfall development model in which the project goes through distinct phases such as requirements gathering, design, development, testing, deployment, and stabilization. However, both business requirements and technology are complex in nature and the waterfall model can take six to nine months to fully implement a solution; by then business as well as technology has often changed considerably. The result is disappointed stakeholders and frustrated development teams. Agile development implements projects in an iterative fashion. Also known as the sixty percent solution, the agile approach seeks to deliver more than half of the user requirements in the initial release, with refinements coming in a series of subsequent releases which are scheduled at regular intervals. An agile data warehousing approach greatly increases the likelihood of successful implementation on time and within budget. This article discusses agile development methodologies in data warehousing and business intelligence, implications of the agile methodology, managing changes in data warehouses given frequent change in business intelligence (BI) requirements, and demonstrates the impact of agility on the business.