Chapter 13: Design Principles: Where to Put the Intelligence

---

You have built your universe, made it more robust with advanced objects and hierarchies, and used Supervisor to grant users access to the universe. You are off to a good start. As your universe evolves, you will face a number of choices about where to put the intelligence, specifically the business intelligence. This chapter focuses on the alternatives and the pros and cons of each alternative. My goal is to help you understand the cost/benefit implications of the choices as you deploy BusinessObjects.

What Is Intelligence?

Intelligence is information with a business context. QUANTITY may be a physical column in a table. Add a time period such as month, then a context to the time period such as order month, and multiply the column by a price, and you arrive at Sales Revenue, something with meaning and value to business users.

Users will rarely want to analyze straight columns of data. If they did, the transaction or Enterprise Resource Planning (ERP) system would meet their needs just fine. To provide a business context, the raw data must be combined with other information, perhaps cleansed, transformed, and aggregated. Many transformations may be critical to the project’s success and known to programmers and Extract Transform and Load (ETL) experts; however, they mean nothing to a business user. If a customer code is 306 in one system and 0306 in another system, the business person really doesn’t care. The business person only cares and knows that this 306/0306 customer is Mrs. Whitwell. Transformations to make the data consistent are necessary to build the data warehouse or mart but are a given to the business user.

If the business user wants to do a promotion for newly married customers, then perhaps classifying this customer under a grouping such as Newlyweds would be a form of intelligence. At first glance, you may assume that this customer grouping should exist in a dimension table. A database is certainly one place to put the intelligence, but the customer grouping could also go in the universe or a user’s report. Following are a few more examples of items that I would consider to be more than just straight calculations or transformations; they provide business intelligence:

• Measures that include time periods, such as Sales Year To Date or Days Late

• Variance analysis that compares the difference between two numbers, such as Current Year Sales versus Last Year Sales, Percentage of On-time Shipments

• Ratios, such as Market Share, Patient Visits per Diagnosis, Gross Margin

• Dimension groupings, such as customer age, income, product size, color

---