I would appreciate any feedback from the community on the following scenario.
We are mastering records for individuals and organizations in a typical party-role model. Eventually, this mastered data will be used to populate a star-schema data mart containing an organization dimension and individual dimension. The sources for the mastered data as well as the dimensions are the same.
Our challenge has been where to draw the line between what data flows into the MDM and what data flows into the dimensions.
One extreme is to only include that data which is used for match and merge, but this limits the functionality of our MDM solution to only matching and merging. In this case, the dimensions and the MDM would read from the same sources.
The other extreme is to include as much detail as possible for our organizations and individuals. The drawback here is the complexity might lock us into an alignment with the dimension structure that might limit the flexibility of the MDM solution for other projects. In this case, ideally, the dimension record would read all of it's data from the MDM.
Do anyone have any suggestions for best practices in this scenario?
The second approach looks to be a better one. However, what is the drawback of it, can you please elaborate on the same.