Agile Analytics: A Value-Driven Approach to Business Intelligence and Data Warehousing: Delivering the Promise of Business Intelligence (Agile Software Development) [Paperback]

The Data Warehousing development arena lags behind the application development area as far as adopting project management and development techniques. Some of the common excuses to not even look at Agile adoption include: Data Warehousing is fundamentally different than application development; The first thing to get abandoned with Agile is data integrity and documentation; Agile is just an excuse to scrap architecture and planning; Agile is just another word for developer laziness. Those are just some of the excuses that I have used recently. I have seen too many 'Agile' projects that have been abandoned after twenty or more three-week iterations when the final mess that was produced was unusable and a maintenance nightmare.

After reading Agile Analytics, however, I am beginning to understand what the author means by the difference between 'doing Agile' and 'being Agile'. Agile techniques, on their own, are not a replacement for Data Warehousing methodology, but rather a complement. On the other side of the Agile fence, I have been involved in several large projects utilizing the waterfall project management strategy that suffered from inevitable scope creep; missed deadlines; missed requirements; building throw-away products that will never provide value just to meet an arbitrary deadline.

The first section of Agile Analytics is geared more to a generalized audience in that it introduces the reader to the broad spectrum of Agile literature and how it applies to Data Warehousing. The second section is geared more to the Agile team members in that it provides them with the tools and frameworks for adjusting on a daily basis to the dynamism and challenges associated with Agile techniques.

The premise that software development should have all requirements defined and well-understood before any development begins is out-dated. The 'Big Design Up Front' assumes that no challenges will arise in development, with data quality or that the consumers of the data warehouse won't come to a more advanced understanding of the business. Agile techniques require a partnership with the business and the development staff with constant and frequent feedback and continuous involvement in evolving requirements.

The emphasis on fulfilling User Stories rather than project dates and Test-Driven-Development can go along way to addressing the data warehouse planning issues. The author does an excellent job of pulling the best practises of the Agile development movement and adapting it for more data-oriented projects.

I strongly recommend this book, both as an introduction to Agile for BI/DW and as a reference for the practical tools for a day-to-day adjustment of a truly Agile project. This book can make the difference between 'doing Agile' and truly 'being Agile.' I am passing this book around my team to see how we can better provide concrete business value to our internal and external data consumers.

If your business intelligence team has discussed "going agile", this book can give you practical information to help you get there. It's refreshing to see that business intelligence and analytics professionals can adopt practices typically associated with Java, Ruby, and Objective-C developers.

The book is organized into two sections, management methods and technical methods. Most of the technical methods focus on data modeling and data integration (often referred to as Extract, Transform, and Load, or ETL). While these areas are critical to a successful business intelligence system, my role is most often focused on the presentation layer or BI toolset (such as SAP BusinessObjects). So I personally gravitated toward the first half of the book, management methods.

Ken says more than once that the whole point of agile is to "be agile", not just to "do agile". Unfortunately, "agile" can be overused as the latest management buzzword to dress up "hacking" or "unrealistic deadlines". I was actually surprised to read that agile may not improve delivery times. In the short term, delivery times may increase. But the payoff for agility is projects that more quickly respond to changing requirements and users that receive smaller functional deliveries instead of the "big bang" of the waterfall project death march.

While the book is a well-written and easy to read, I found it necessary to read slowly, chapter by chapter, and reflect on what I had read. The book would easily lend itself to a weekly BI book club, where technicians, users, and management meet weekly to discuss the book one chapter at a time. Recommended reading.

I am very pleased to submit this review. I could just say the author delivers.
Good coverage of agile topics and practices including user stories, team work, project automation and how to introduce agility to BI projects.
Even if you think you are constrained by COTS BI/DWH applications, there are great and practical pieces of advice to improve the way agile software configuration management and agile release management work.
No consulting clutter, no obscure concepts. Clear and simple. I clearly identified some of the scenarios mentioned in this book from my experience in business intelligence projects.
BTW, I don't buy Agile 100%... but I do believe agility is needed in our projects and the approach described in this book works for me.