Do you have any best practices or tips for choosing and evaluating data modeling tools? We’re not quite sure where to start with the whole process.
The most important step in choosing a data modeling tool is to identify your requirements. You should break them out by must-haves vs. nice-to-haves, then rank the nice-to-have requirements in order of importance. Be sure to take all the different points of view within your organization into account when choosing a tool -- it’s a choice you’ll probably have to live with for years to come.
All of the available data modeling tools can do basic modeling functions (i.e., create physical and logical data models, reverse/forward engineering, etc.). Some things to consider beyond that include: team-based modeling capabilities, versioning, customization (e.g., an extensive number of ways to tailor diagrams), the model repository, the ability to develop conceptual data models, integration with an enterprise metadata repository, and data rationalization (for maintaining the lineage of model objects between the different model levels -- conceptual, logical and physical).
You also should identify all of the groups that currently use a modeling tool (if any) and find out what they like about their existing tool and what they would like to see in a new one. Forming a committee to identify requirements, interview vendors and make the selection is another good idea because it is such an important decision. A data governance council, if in place, would be the ideal group to make the final approval.
As it is such an important decision, bringing in an outside consultant can provide an objective view as well as expertise and assistance in choosing the right modeling tool. Please feel free to contact me at firstname.lastname@example.org for more information about the process of choosing a tool or for assistance with your assessment.
This was first published in June 2010