I don't really understand what Business Intelligence is all about. If I start from having a corporate DB, what is it that a BI person would do? I found plenty of material on the web, but it usually is a bit too complex. I want a simple example that would make me understand what BI is all about and what would a BI person produce that is of value to my organization.
Business Intelligence is often a completely separate sect from Database Administration and Database Development. Business Intelligence, at the highest level, includes three main facets:
Reporting is the creation, deployment, and management of reports as well as the added ability for users to customize reporting dynamically.
Data integration and transformation solutions. On the very simplest level, it is the means of extracting, transforming, and loading data into a data source, from a data source (which could be anything as simple as a flat file). Integration is a mile deep, but that's the most basic functionality of it.
Online Analytical Processing (OLAP) used to design, create and manage structures that contain data aggregrated from source data stores. A catch phrase for this is data mining.
These are extremely simplified descriptions of what Business Intelligence encorporates. There is a science behind BI, as well as each of these facets individually. Database Professionals dedicate their time and careers to mastering these.
The value depends very much on the individual organisation and its requirements. Depending on the level of sophistication required, a B.I. role might fall into a few different categories:
Most B.I. staff tend to fall into one or more of these categories. The value to an organisation varies with individual circumstances. One common phenomenon that I observe is that people responsible for operational systems greatly underestimate the amount of work that actually takes place in these roles. I've seen one insurance company that had 170 staff just in the accounts department of their European operations. Most of their time was spend wrangling data extracts in spreadsheets and operating manual reconciliation and control processes.
Management Information is very often a poor cousin in the priorities during development and operation of line-of-business applications. A poorly coordinated or non-existent data architecture strategy can cost a large amount of time and money. The default behaviour is to treat systems as silos with nobody having direct authority to fix cross-system data issues. Leave this for long enough and the net effect is back office operations employing hundreds of clerical staff (often qualified finance personnel) spending most of their time doing the work of a few stored procedures.
I'm going to take a stab at this part of the question as I think others have done a good job of explaining what BI is. I work for a company with many clients and I know a great deal of information about the functions we provide for those clients.
Our applications are very data-centric; our industry is regulated by the government so compliance with federal and state laws is critical. What do our BI specialists bring to the company that makes them valuable?
If your data needs are all internal, you still may have internal clients who need this level of analysis. In this case, you are probably more concerned with the reporting aspect of data warehousing than importing data into a transactional system. But still the abilty to use the data you have been collecting to make management decisions is invaluable to most organizations.
Whether you need a BI specialist tends to revolve around how extensive your data needs are and how complex the system is. A smaller business may not have enough work for a person of this nature and may hire consultants to create the reports they need. BI specialists tend to work only at medium to large businesses.
If you are business that creates COTS software, you probably need BI specialists to be the consultants who know your product inside and out and create customized rpeoting from it for your clients.
While they are not great examples of best practice, the SQL Server sample databases would be good place to start. They include an OLTP, data warehouse and analysis services databases for a fictional organisation. Studying the differences between them should help you make sense of how OLTP (transaction) and OLAP (analytical/BI) databases differ and why.