How Semantics Applies to Business Intelligence
Ask yourself and your data warehousing project teams this question: How will you prepare for these technologies? Focus on three areas that you’re probably already working in — your business intelligence semantic layer, business rules management, and possibly federated query definition.
Business intelligence semantic layer management
If you dig deep enough into your process and tools, you can find someone in the business or delivery team who maps key business terminology to technology requirements or implementations. People who use Business Objects Designer or Cognos Framework Manager should map key business terminology (or business metadata) to technical database tables and columns.
If your process is mature, you might even find that business analysts capture the business metadata in Microsoft Word templates, which your business intelligence tool administrators leverage. If this is the case, raise the bar on these efforts. Begin to define the semantics (another word for the business metadata) and initiate an alignment process across the silos of your business.
By performing such a process, you can begin to determine key terminology conflicts across lines of business or functional areas of your business. The sales department might refer to the dollar amount on a contract as Revenue, but the finance department refers to it as Bookings.
If finance and sales worked together to resolve this term, they might agree to use the term Booked Revenue, defined as money committed by a customer legally for the delivery of products and/or services within a given specification laid out in a contract.
Additionally, the finance and sales team might conclude that the business needs to monitor a revenue life cycle that spans sales (Booked Revenue), field operations (Deferred Revenue), and accounting (Recognized Revenue). If you want to bring the semantic world to your data warehousing environment, you start by understanding key terms, as well as their associations and rules.
Business rules management
Additionally, investigate the process around business rules — both how you define those rules and how you generate them. The process of defining business rules is very similar to the business intelligence semantic definition process.
Users define key concepts; for example, Customer means a legal entity which has conducted business by purchasing goods or services from one of our companies in the current 12-month period.
Such a definition evolves into a database query or data movement routine that flags a legal entity as a customer or not. You can move your enterprise toward the world of semantics by creating an alignment process that gathers a cross section of the business community together to agree on such terminology and rules.