Conceptual Data Modeling: Resisting the Urge to go Physical

Data Modelers, Data Analysts, Application Developers or DBA's often have a tendency when developing a new data store (whether Transactional or Decision Support) to immediately start developing a Physical Data Model. However, for best results Conceptual Data Models should be developed prior to Logical Models, Dimensional Models and Physical Models. Developing a Physical Data Model before Conceptual, Logical and Dimensional (if appropriate) models can lead to less than optimal production systems that add additional complexity, require extensive maintenance and result in poor extensibility. Conceptual Data Modeling is used to identify:

• Business Entities
• Entity Subtypes
• Relationships
• Cardinality
• Optionality

Conceptual Data Modeling is a very effective means for reflecting your understanding of business requirements back to business analysts for validation. Conceptual Data Models also provides effective documentation of business subject areas – knowledge is retained in the models, not just in someone’s head.