USING THE ENTERPRISE DATA MODEL TO MAXIMIZE RETURN ON DATA WAREHOUSE INVESTMENT: PART II
by Kathy Long

The challenges in defining the data architecture for warehousing lie in achieving data integration, promoting information reuse, and providing business intelligence. The EDM will help define a current and future data architecture that meets those challenges.

1. The EDM provides an enterprise perspective

One purpose of the EDM is to discover common threads and develop a cross-functional, common definition of the entity. For example, a telecommunications corporation manufactures equipment, purchases equipment for use and resale and develops for sale network management software that manages equipment. The functional information requirements for equipment data differ depending on whether equipment is perceived from the viewpoint of the manufacturing function or the network systems management function. Is the network switch equipment or product? Does the enterprise have the same business terms and rules for equipment manufactured, purchased, and managed via network software?

The definition of information in common enterprise terms improves the context of information. The warehouse often provides global enterprise access to information that had previously been limited to a functional area. This broader audience requires that the information be defined in a context that makes sense to the reader regardless of their industry background or global location. Another important purpose is to document derivations for strategic key performance indicators. These are the business measures that executive management applies to measure performance.

2. The EDM defines strategic information needs

The enterprise data model is a 'to be' model. It is strategic. It recasts data into a model with vision that transcends biases in today's data. Strategic definitions are developed that remove data from its literal system representation and transform it to information in business language that will support growth and new initiatives. For the natural gas utility that seeks new and unique service and product offerings as it continues to evolve under FERC deregulation, the EDM plays a critical role.

Many of us are familiar with the complexities of modeling customer information. Is the customer the individual who orders the item, receives the item, pays for the item ? The roles of orderer, receiver and payer could all be different individuals. In some cases a customer can be a competitor. A supplier can be a customer. Are there universal customer rules and rules that apply only to specific customers? It is through the definition of the EDM that the natural gas utility can initiate the definition of new requirements for business party information and business party roles that will support a more dynamic business model that meets marketplace challenges.

Because the EDM is at a conceptual business level its development explores the entities and business rules most important to corporate survival and success. It is developed with key decision makers and its purpose is to represent information in a manner that will help the enterprise compete and grow. The development of the EDM provides a good understanding of business party roles and important super-type and sub-type entities which will be critical to the design of the data warehouse and dimensional models.

3. The EDM forms the foundation of the warehouse atomic data store

The data warehouse architecture most subscribed to today is designed as a multi-tiered data architecture comprised of a logically centralized, normalized, atomic data store that is the source for multiple functional data marts. In the design of the atomic data store the goal is to build a foundation of data that is a shareable, consistent, reusable atomic source of information for the dimensional data marts.

Since the atomic data store must be designed to meet cross-functional needs, it requires the EDM as a blueprint. The reference to the EDM ensures that the atomic data store is built on a forward thinking data model and a solid understanding of cross-functional entity super-types and sub-types. It ensures a more stable design of the entities that form the business dimensions in the atomic data store as the corporate business model evolves in response to market challenges. The reference to the EDM also facilitates cross-functional data integration.

4. The EDM supports planning the data warehouse release strategy

The EDM describes the integration points between primary entities, provides insight into the complexity of information, and supports the gap analysis between strategic information needs and currrent information availability. The EDM provides an improved understanding of information requirements across functional and organizational boundaries. This knowledge is input to the long term warehousing strategy and assists in determining the phased implementation of information blocks that will provide a solid return on investment.

5. The EDM initiates business data stewardship

The purpose of data stewardship is to place the accountability, control, shareability and quality management of enterprise data in the hands of the business people who define, create and access the data. The subject matter experts that contribute to the EDM serve as strategic data stewards. They are responsible for the development of data definitions and standards that support enterprise-wide use and are an important contributor to the quality management process for the data warehouse.

Summary:

The following outlines some critical success factors for the development of an Enterprise Data Model:

• The development of the model must have active executive management support.
• The definition of information must cross functional and organizational boundaries.
• From its inception, the enterprise data model must be perceived as a critical component of the information systems architecture.
• Subject matter experts must be available to participate in its development.
• The model must be actively maintained.
• The model must be developed using best practice data administration standards which includes the definition of a meta-data model and use of a CASE tool.

The EDM is more than an ideal, an academic exercise that is subscribed to in theory, but when the reality of project deadlines loom, gets tossed out of the project plan. Familiar reasons for eliminating the EDM are time and lack of subject matter expertise to contribute. There is no argument that the development of an enterprise data model is a challenging process. The internal political realities of the enterprise, extent of decentralization, and dynamics of the acquisitions and divestitures are all factors that increase its complexity. However, I think too often the EDM is bypassed because its content, purpose and application are not fully understood.

As the conceptual blueprint for the enterprise data warehouse, the EDM is indispensable. It defines the warehouse vision. It provides a strategic, horizontal view of information across the enterprise that is critical to the logical and physical design of the warehouse data architecture. The more complex the business, the more critical the EDM is to the success of the information systems architecture and specifically the data warehouse.

---

For more information, see http://www.spectrumtech.com