I gave a presentation to a group of hedge fund CTOs recently. The focus was “What really matters to getting the most business value from your data?” It generated some great discussion, both because of the economic climate and because as businesses of a few hundred people the resources they have are limited. It seemed to strike a chord so I thought I’d run a series of blog posts on the same subject. Please feel free to chip in with your own experience.
Where to start? Two schools of thought: Inmon and Kimball
Every hero has their nemesis. Every Holmes has his Moriarty. In the information management world there are two big schools of thought, Bill Inmon’s and Ralph Kimball’s.
The Inmon approach tends to start from an enterprise-wide perspective and look to build a data architecture that will support any business need that may arise. It’s principled and conceptual. In contrast, the Kimball approach tends to start much smaller and start with a contained
business question. It’s pragmatic and practical.
A key architectural difference is that in Inmon’s architectures you have a central data warehouse and data marts; in Kimball’s, data marts plus an overall data model are your data warehouse.
I’ve personally worked on projects which followed both approaches and as with many things in life, each has its own merits. As a company we tend to prefer Kimball’s “dimensional modelling” approach because we’ve generally found it aligns with our company value proposition of “strong business alignment” and “short time to value”.
In the next posting we’ll consider the benefits that come from using dimensional modelling and we’ll also look at a real example of a business matrix.