Just a brief blog this time around, folks. As I have been settling into a new job, I have run across an often-repeated common set of practices in implemented a business intelligence solution for your business. Many, many times I have encountered an IT shop implementing a BI tool that has promised to uncover hidden treasures in vast amounts of mined gold and to improve sales, satisfaction and product marketing numbers immediately.

Ok, in the Microsoft world as well as any other BI software toolset, these are promises that can be fulfilled. The operative word being "can". The tools are enablers, simplifiers and just that: tools. To successfully make strategic advantages for your business users, you must implement your BI infrastructure in a systematic process that includes fundamental data modelling and planning initiatives.

I am going to make it a goal of mine to outline those throughout the next coming months in this blog. What I will do today is to start with the beginning: requirements gathering.

I am going to make an assumption that you have already chosen Microsoft as your BI vendor and are going to begin with a database in SQL Server 2005 and utilize PerformancePoint Server 2007 to implement charts, reports and scorecards. This is good. Now go and interview the business users. Ask questions such as:

1. How do you measure your business today (KPIs - product performance, sales targets, budgets, etc.)
2. Ask about common business entities - products, people (organizational mapping), sales figures, budgets. Begin the process of looking for "dimensions".
3. Ask questions about what are the difficulties in finding out sales performance, goals, customer segmentation, etc. Look for where you will begin to map "facts".
4. Look for the relationships: How do sales relate to quotas and salespeople. What do the reports look like today.
5. Ask about time, frequency and geography. This will help to map out slicers such as "by region", "by quarter", etc. as well as to determine the update frequency you will need for ETL and cube refresh. Remember that SQL Server Analysis Services is an OLAP engine and we will likely use to MOLAP to refresh the cube on a schedule.

Use this as part of the analysis to formulate a set of requirements that will generate your specifications. We'll walk through this process in this blog coming up soon and I will show you how to trace all the way from analysis to actual reports and scorecards implemented in these tools.

In the meantime, search out my previous MSDN blogs and you will see some examples of leveraging a BI COE and data modelling techniques that I explain in my blog that can be of help to you as you map out your BI deployment strategy. We'll work on this together!!

C'Ya - Mark