Business Intelligence for the DBA: Part 1

I was reading Michael Otey’s 'Getting Started with BI' Article and it compelled me to write two blog posts here on SQL Magazine to serve as a follow-up. In this post I’ll take you on a brief tour of what is Business Intelligence (BI) and why is it so valuable for companies to implement. In part 2, I will take you on a 5,000 foot view (mid level) of how traditional Microsoft BI solutions are implemented.

Most current day organizations will have some form of reporting or analytics. A "normal" company will have reports (operational usually) running directly off of various OLTP databases. These reports produce useful business measures for that one operational system. In another area of the company, similar reports (maybe created in a different technology) are built for a different operational system. And there might even be some form of a corporate intranet that hosts one or both sets of these reports for thin-client viewing. Spreadsheets continue to dominate the "normal" company as they are flung in-between meetings with various edits being made to their metrics. The CEO might even receive two different reports for the same metric yet having different results. The situation I am describing is "par for the course" at most companies still to this day...

Consistent, Reliable Information

Now can you imagine working in a company where information is consistent, reliable, and flows throughout? An organization that ensures executives are constantly aware of how their business is performing, where analysts explore free-flowing models to locate hidden trends, and where operations has constant visibility for their daily decisions wrapped with seamless collaboration? What I am describing to you is a 21^st century Intelligent Enterprise. Such a company is not reserved only for our generations to come; it is viable today as I write this blog post. It is the idea and current day enablement capabilities of the Intelligent Enterprise that makes BI ever increasingly popular.

So why is it hard to obtain this Intelligent Enterprise model in today's world? First, I could not agree more with Michael's thoughts on this matter. Secondly, I emphasize a lack of general enterprise BI expertise in our environment. So why are there not more BI experts? As Michael said, cost and scope are the big reasons why many BI projects never get going. Most BI industry experts will tell you only ~20% of companies have implemented BI solutions and thus the "normal" scenario I described above still prevails. However, looking forward, more and more rapid BI solutions are becoming available.

Rapid BI

Rapid BI is enabled by any combination of tools, technologies, and processes. Agile software development practices are beginning to be leveraged for BI solution development. The Kimball methodology of designing and delivering individual (yet interconnected) data marts enables much quicker ROI for enterprises. The cloud (see SQL Azure) will enable small data marts in the cloud. Microsoft's Fast Track Data Warehouses provide a reusable hardware reference framework to remove the burden of hardware architecture design for enterprise DW/BI solutions. Finally, any premier BI professional services firm should be more than willing to provide a pilot project for your company. BI pilot projects usually translate into the creation of a test data mart (but not always). BI pilot projects should last no longer than six weeks.

Related: Business Intelligence Explained for the DBA