Chapter 4 • The Data Resource 101you allow your customers to place orders via the Web. Your
organization writes programs, say, in Java and Microsoft’s ASP,
to process orders via Web pages. When a customer places an
order, your order entry program checks inventory to see if the
order can be fulfilled without delay. When your warehouse is
out of stock, you’d like to be able to check inventory in one of
your suppliers. With XML you can send over the Internet a
document with the product identifier being ordered to a Web
Service on your supplier’s computer systems; that Web Service
will respond with the description of the item, its on-hand inven-
tory level, and any costs associated with filling from the supplier’s
inventory. With this data, your e-commerce site programs can
determine what to display to your customer. XML is used in
this case to indicate what data are being exchanged by using
labels or tags (similar to HTML) that both organizations
understand. In neither case, however, does either organization
have to understand the programs, computers, or database
management systems the other is using.
Managerial Issues in Managing Data
Having considered key technical issues involved in manag-
ing data, let us now turn to managerial issues. How to plan
for data, to control data integrity, to secure access to and use
data, and to make data accessible are important to the busi-
ness manager. As with any business resource, quality sources
for data must be identified and the data acquired; enough
space must be available for data storage; obsolete data must
be identified, disposed of, or archived; and usage of data
must be accounted for, and, if appropriate, usage fees should
be charged to those utilizing the data. These are not just
issues for the IS organization—the business manager should
be equipped to deal with these issues as well.
Principles in Managing Data
Successful management of the data resource depends on
understanding certain key guidelines:
THE NEED TO MANAGE DATA IS PERMANENT Any
organization has customers or clients, whether these are other
organizations, individual consumers, or patients. Whether a
company makes to stock or to order, there are vendors or sup-
pliers, orders or reservations, products or services, and
employees. Further, irrespective of how accounting, selling,
billing, or any other management activity is performed, there
still will be data about customers, vendors, orders, products,
and employees. Data values might change, new customers
might be added, products discontinued, and employees hired
and retired, but a company will always have customers, prod-
ucts, employees, and other entities about which it needs to
keep current data. Occurrences of data are volatile, but the
existence of data is persistent and the need for excellent data
management is constant. Business processes change and so
must information systems. If the company decides to change
a sales forecasting method, programs will have to be rewrit-
ten, but customer, order, and general economic condition
data are still needed. In fact, if data are well managed, many
of the databases will remain relatively unchanged when an
organization decides to change the way it does business. At
the same time, the programs that analyze, process, and report
information might change drastically. Thus, data are funda-
mental to the business. Data remain over time and need to be
managed over time.DATA CAN EXIST AT SEVERAL LEVELS Although the
business retains vast amounts of data, there might be rela-
tively few basic classes of data on which to base most
information. One way to organize data is called the data
pyramid (as depicted in Figure 4.3). Although new data
can enter this pyramid at any level, most new data are cap-
tured at the base of the pyramid in operational databases.
These databases contain the business transaction history of
customer orders, purchases from suppliers, internal work
orders, changes to the general ledger, personnel transfers,
and other day-to-day business activities. Managerial con-
trol and strategic databases (often called data warehousing
and used to support decision making and business intelli-
gence) are typically subsets, summaries, or aggregations of
operational databases, with key external data as supple-
ments. For example, a database for sales forecasting (a
managerial function) might contain past monthly sum-
maries of sales by product family or geographical area
derived from customer orders and product data. These data
might be supplemented with external economic indi-
cators and sales force judgments to produce sales estimates
needed for production planning and scheduling.
When managerial databases are constructed from
sources other than internal, shared operational databases,
there can be significant inconsistencies. For example, the
sales organization might track customer orders in a local
database before passing these on to order entry. If they use
these figures for forecasting final sales, they might not con-
sider canceled orders, orders rejected due to insufficient cred-
it, returned goods, or sales not met because of inadequate
production capacity. These information items might not be
considered because they enter the business at other points of
contact with the customer. A well-run organization must con-
sider all the transactions that define sales level to be able to
build an accurate sales forecasting system.
Developing an understanding of the relationships
between data in various databases is a critical element of
managing the data resource. Ideally, aggregate data will be
derived from operational data, not collected separately