A Framework for Classifying Data - High-Performance Web Databases: Design, Development, and Deployment - page 67

Exhibit 5-1. Representative data in the four quadrants.

Internal External

Administrative • Organization Chart

• Employee Particulars
• Compensation
• Office Location

• Customer
• Account
• Transaction
• Product

Directional • Strategy

• Policy
• Mission
• Goals

• Marketplace
• Technology
• Competition
• Economic Forecast

The focus in the internal administrative quadrant is an introspective
look at the organization. Typical data found in this quadrant includes the
organization chart, employee particulars, compensation, and office loca-
tion. In the internal directional quadrant, the interest centers on such
issues as the organization's strategy, policy, mission, and goals. The focus
of the organization's external administrative quadrants is on the custom-
ers, accounts, transactions, and product. This quadrant is the most famil-
iar to modelers because most data and most systems fit here. The focus of
the external directional quadrant is the marketplace, technology, competi-
tion, and the economic forecast; these are broad subjects of interest to the
organization as a whole.

An interesting feature of the matrix showing internal, external, adminis-
trative, and directional data is the tendency toward mutual exclusivity of
data within the quadrant. Most data seems to fit naturally into one quad-
rant, and no others. This concept is discussed further in the next section.

Mutual Exclusivity

The tendency toward mutual exclusivity of data within the matrix is
reflected in the data models that represent the data in each quadrant. In
general, the data found in data models for each quadrant applies only to
that quadrant-rarely does an organization's data model successfully
include data from more than one quadrant. The most successful modeling
efforts seem to be those that do not mix data from the different quadrants.

There is good reason for not mixing data from different quadrants in the
data model. The goal of most modeling efforts is some form of systematiza-
tion. From systematization there results payback to the organization,
either through the leveraging of technology, or through the economies of
consolidation that can be achieved through integration of common data
and processing. Without some payback, however it is measured and real-
ized, systematization makes no sense. A method of measuring payback is
described in the next section.