In response to my columns on the qualifications of the data modeler and how that can help business, a reader inquired about who is responsible for validating the data model. The answer to this question depends on the type of model being validated. In this column, I will examine the responsibilities for the business data model and in a future column, I will look at the validation responsibilities for the technology data model.
The business data model, also known as the enterprise data model, conceptual data model or logical data model, describes the major information of interest from a business perspective. This model is rarely created as a single project - it is typically built incrementally as needed to support new data structures. It is the unifying model, and all of the subsequently developed data models must be consistent with this one.
The business data model is independent of specific application system needs and is best developed by the data management group with input from business subject matter experts. Validation activities for this model include:
- Completeness for use: Is all of the data needed to support the pertinent application development project modeled?
- Completeness: Are the business areas and concepts represented by the model completely addressed?
- Accuracy: Does the model correctly represent the entities, attributes and business relationships of interest? Are the optionalities and cardinalities correct?
- Standards conformance: Does the model and the supporting metadata store follow the prescribed policies and practices?
- Generalization and specialization: Does the model represent the data at the appropriate level of abstraction?
- Structure: Does the model follow the data modeling rules for the level of normalization (typically third normal form) used?
With the exception of validating aspects of standards conformance and structure, all of these questions require comparing the business needs and environment to the way they are described within the data model. Structure and standards compliance validation should be performed by representatives from the data management group who are experienced data modelers. The remaining validation, however, must clearly be performed by business representatives - but which ones and how?
The data model addresses several subject areas and crosses organizational boundaries. If a data stewardship council exists, it is very likely that subject area responsibilities have already been assigned and that its members are capable of performing the model validation. In the absence of a data stewardship council, the model must be segmented based on organizational responsibilities, with one or more representatives being responsible for each set of data.
Model validation by the business community can be performed in several ways. From the modeler's perspective, the most straightforward approach is to print a copy of the diagram and walk through it with the business representatives. This approach requires the businessperson to understand a diagram that is IT oriented, and it is appropriate when the validation is performed by the data stewardship council or people directly involved in IT projects.
To some businesspeople, the entity relationship diagram is like a foreign language. These people are likely to resist learning how to read the diagram, and a different approach for validating the model is needed. For these people, the best approach is for the data modeler to identify the portions of the model for each person to validate and to verbally interpret the information in the model and the associated business rules. For instance, if the model contains the entities and relationships in the example (see figure), the businessperson should be asked to confirm that a person is always assigned to a specific department, but that some departments may have no employees within them.
Further, since the metadata includes definitions, the definitions for "department" and "employee" should be confirmed. Additionally, the data elements included within each of the two entities should be listed, and the businessperson should be asked whether or not any significant ones have been omitted.
The business data model describes the business. While the representation of the data can be validated by representatives from data management, the content should be validated by businesspeople. The review process itself should be tailored to the reviewer and alternatives to an examination of the actual diagram may be needed if the reviewer is not familiar with the notations.
I welcome your input at my email address below. Please send me your thoughts and experiences on validation of the business data model.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access