I am a strong proponent of modeling for all aspects of system development. One of the most important aspects of modeling, in my opinion, is to capture the “as-is” environment. Often customers or sponsors have existing systems not green fields. While most of my work has been in the systems arena, it has expanded slowly to include the enterprise architecture.
One of the first tasks we have done is to document the existing “as-is” architecture. What I am seeing when we do this, are data architectures that are sub-optimal at best. That said, you can’t make improvements to the data architecture from just your models. Unfortunately, non-technical aspects such as the governance, usage, and operational tend to be where most of these battles are fought.
The soft skills of determining, and enforcing, who controls and administers the data, storage length and archiving, data security and who controls what, are some of the biggest areas of challenge (or contention) that I have encountered. Once these and other concerns are addressed it is then possible to move forward with a “to-be” model.