Databases Reference
In-Depth Information
the benefits of the data in the data warehouse, will continue to be sup-
ported by all such business units. The data warehouse, however, is left
holding the expenses, but none of the benefits, associated with the data.
The gravitational pull of the data in a data warehouse moves smaller
applications toward itself so they can use the data in the data warehouse.
The gravitational pull of the data in the data warehouse moves the data
warehouse toward larger business units when a larger business unit has
the clout necessary to exert control over the data warehouse. In both sce-
narios the success and growth of a d ata warehouse attracts forces that
may, if left unchecked, destroy the data warehouse.
PurPose AND roI
The solution to the gravitational pull of data is purpose and ROI. The pur-
pose of a data warehouse is that set of features and functions it intends
to deliver to the enterprise. The ROI of a data warehouse is the realized
value caused by the data warehouse when the data warehouse delivers the
features and functions of its purpose. The selection of a purpose should
be decided as early as possible in the life of a data warehouse. For exam-
ple, a data warehouse may produce Sales summary reports and Logistics
exception reports every morning while an Operational Data Store sends
low-inventory and late-shipment alerts in near real time to the PDA of
Logistics staff managers. This is when a data warehouse makes itself part
of the fabric of the enterprise. The data, reports, alerts, and analytics deliv-
ered by a data warehouse—delivered every hour, every day, sometimes in
near real time—are collectively the purpose of a data warehouse. The ROI
is benefit and value realized within the enterprise when the data ware-
house delivers the features and functions included in its purpose.
The ROI of a data warehouse presents a cost to any effort that would
change the data warehouse to fit the requirements of a ten-ton business
application. In that way, a change to the data warehouse becomes a change
to the enterprise. By delivering Sales reports every morning, Sales reports
that are immediately consumed by all levels of management, a data ware-
house inhibits any efforts that might abridge, interrupt, or delay the
delivery of those Sales reports. By delivering near real time alerts from
its operational data store to tactical managers throughout the enterprise,
near real time alerts that help those tactical managers avoid expensive
