I have read virtually every Data Warehouse Design book that I could find, ranging from titles from Ralph Kimball to William Inmon to everybody in between, and I found this title to be by far the best, most useful, and clearly written book available on the subject. I have worked in many data warehouse environments large and small, and I find his advice to be imminently reasonable and highly valuable for anybody involved in a data warehouse construction effort.

The author's basic premise is that the chief purpose of a data warehouse should be to maximize the power of Customer Relationship Management. He covers all of the basic data warehouse concepts and then provides a conceptual model of a Wine Club as the basic example for the book. I find his use of the model exceptional, in that he starts with a transactional data model for this example and then proceeds to extend it into a well-formed dimensional data warehouse. This is rather exceptional, in that I have found that many data warehouse authors are so intent on proving the superiority of dimensional modeling that they often harshly criticize the underlying transactional models as "not representing the business problem." I have always found this to be an unfair proposition and support that idea that databases which are designed to support transaction processing are not necessarily good for decision support but that the transactional environment must be fully understood before formulating a dimensional solution.

The book deals extensively with one of the most critical elements in data warehouse design: the problem of how to handle time in the model. He examines many different roles that time plays in the model, including its role in the construction of dimensions.

The crux of his analytical efforts come with the introdution of his "Dot Modeling" method. This method provides a simple, but highly useful, conceptual framework for generating the logical model of the data warehouse. It also gives explicit support for handling most of the time-related issues. A major added benefit is that the basic Dot Modeling constructs are far simpler yet more complete than standard Entity/Relationship modeling diagrams, which makes it a joy to use with end users, who are able to construct complete and easy to comprehend data models without being required to learn any arcane Parent/Child terminologies or be dragged down into a discussion of the pros and cons of a Snowflaked model.

The rest of the book covers many other essentials of data warehouse construction, including physical implementation issues, an excellent discusion on the business justification for the project and its importance for success, and project management suggestions.

He also has a chapter discussing various software products. I agree wholehearedly with his contention that there is no such thing as a single CRM product. Rather clients are often faced with an array of numbingly complex and quite expensive options to satisfy different parts of the customer-focused needs of the data warehouse. He has wise counsel about the pros and cons of picking such applications, though he does not mention any vendors by name.

The book seems far less driven by ideological issues that tend to dominate other texts. I think that the author's experience working on data warehouses in the UK (some of the time with Hewlett-Packard) has left him much more focused on the practical issues faced in solving data warehouse construction efforts. He does hold a Ph.D. in the subject, so his level of abstraction may, however, be difficult for those who prefer cookbook solutions for their problems. But for those of us in the profession who prefer to apply active thought and analysis for what are ultimately unique problems for each client will find this book to be of imcomparable value in their careers.