Conformed Dimensions

Conformed means to be in agreement; be similar

Conformed Dimensions would mean, dimensions that are in agreement with many data marts in order to evolve an enterprise data warehouse. A conformed dimension is a dimension, which is standard across all data marts. Its quite a challenge to have one thing meeting everyone's needs. However, once done can have great benefits.

Advantages of Conformed Dimensions:
1. Independent data marts become a part of fully integrated data warehouse.
2 Deliver a consistent view across the business.
3. Data should never be defined for a specific function or department. Good data is one, which is widely shareable and conformed dimensions help in doing that.

Conformed dimensions should be defined at the most granular level so that each record in these tables corresponds to a single record in the fact able.

Points to consider while standardizing dimensions:
1. A set of dimensions (customers, products or geography) can be similar for Business Unit1 and Business Unit 2, but their hierarchical arrangements can be different.
2. Hierarchies may change over a period of time, business unit re-org or product re-categorization.

Having conformed dimensions, with above points considered:
1. The first thing is to arrive at a basic set of dimensions across as many data marts/business units as possible.
2. Compile a combined, de-duplicated, complete set of members for each dimension across all data marts.
3. Compile the linkages between members with each dimension by hierarchy.
4. All hierarchies are part of the enterprise but there are certain hierarchies, which belong to specific data mart/business unit. This would require an entity, which would store the information of hierarchy and the owner.

Given below is a diagram, which shows how product dimension can be made as a conformed dimension.



Effective date and End date can be added to the linkage details, if a hierarchy needs to be changed over a period of time.