Implement a Data Mart — Quickly
No matter how you decide to divide the universe of possible contents into some subset for your data mart, remember that in order to obtain maximum business value from your data mart, you must implement it quickly.
Here are the three keys to speedy implementation:
Follow an iterative, phased methodology. You spend the majority of your upfront time on the project focusing on the specific business value the end user wants and over several iterations build the solution into their vision.
Hold to a fixed time for each phase. If you set aside two weeks for your scope, for example, stick to that window. Don’t extend any phase (especially the early ones) unless the project is doomed to failure if you don’t.
Avoid scope creep at all costs. Although costly and dangerous in any project (data warehousing or otherwise), scope creep (when additional feature requests keep creeping in long past the cutoff point) can devastate a data mart effort.
By adding these last-minute features, you probably add complexity to your data mart with only marginal incremental business value (if any), so you do little other than put your project at risk.