548
Making Key Strategic Decisions
materials. Lastly, some companies use EDI to transmit their invoices and then
to receive the subsequent payments. While industries use different versions of
EDI in different ways, their goals are always the same: minimize the processing
time and lower inventory costs and overhead expenses. An industry organization
in Washington, D.C., developed and maintains a standard format that dictates
how all transactions are sent, ensuring that all companies that wish to imple-
ment EDI can be assured that all vendors’ and customers’ computers will un-
derstand each others’ transactions, without requiring any custom programming.
EDI, while still used quite extensively, has been eclipsed by electronic com-
merce, which will be discussed later in this chapter.
The 1990s has also seen the advent of virtual organizations. Virtual orga-
nizations are formed when companies join together to create products or en-
terprises that they could not have created individually. In most cases,
information technology allows companies to create these partnerships and
share information as if they were one company. Using communications and
groupware products like Lotus Notes, the partners can share information with
each other about their individual progress to ensure the best possible success.
This will be discussed further in the section on IT strategy.
DATA BASE
The following scenario depicts what information systems looked like prior to
the use of database management systems. Imagine a physical office in which
each person has his or her own file cabinet. The information in the file cabinets
belongs to the people whose desks are closest to them. They decide what infor-
mation will be in their file cabinets and how it will be organized. For example,
sales might refer to gross sales in one worker’s cabinet and net sales in an-
other’s. Yet, the discrepancy would be unimportant, because there was actually
very little sharing of data.
Database management systems assume that information is a corporate
asset to be shared by all workers in the enterprise. Database technology, there-

Exhibit 16.6 is an example of a very simple portion of a payroll applica-
tion. There are two different tables. The employee table contains data about
each of the company’s employees: name, address, marital status, number of de-
pendents, and so on. The pay table contains data about every time each of the
employees is paid: their gross payroll, social security taxes, federal withholding,
state tax, and so forth.
First, notice the common column between the two tables, the employee
number. This column enables the database management system to relate the
two tables. It allows the system, for example, to print a payroll journal that has
both the weekly payroll information from the pay table and to access the em-
ployees’ names from the employee table. Why not combine all the data into one
table? Not only would the employees’ names, social security numbers, and
other information appear multiple times, requiring the unnecessary use of data
storage, but also multiple versions of the truth might occur. If one of the em-
ployees should happen to change his name or address (if address were included
in the employee table), the database would show one name for part of the year
and another for the rest of the year. Redundant data creates opportunities for
data corruption; just because data is changed in one table, that same data is not
necessarily changed in all tables. Prudent systems design eliminates data field
duplications wherever possible.
DATE WAREHOUSE
Data warehousing attempts to reconcile and live with past applications soft-
ware, while still benefiting from today’s newer technology. As mentioned
earlier, industry is rife with older legacy systems that are currently cost pro-
hibitive to replace. Most of these older systems are mission critical operational
control systems (see Exhibit 16.3) and satisfy most of the operational needs
of the company. However, they are built on technology that cannot support
the
kinds of decision support tools that management requires. Many of these
550

Malt 567890123 S
1 6/6/80 6/2/93
6/17/96 900.000 S
6 John K Free 678901234
M
5 8/5/49 11/1/94
12/15/95 17.500 H
PAY TABLE
Number of Number of
Social
Federal
Employee
Regular Overtime
Security Medicare Withholding
Check
Number Date Hours Hours
Gross Payroll Tax Tax
Tax
Net Pay Number
1 1/7/96 40.0
4.0
391.23 24.26 5.67 101.1642534
21.52 238.62 1
2 1/7/96 40.0
0.0
750.00 46.50 10.88 193.935
41.25 457.44 2
3 1/7/96 40.0
0.0
500.00 31.00 7.25 129.29

0.0
900.00 55.80 13.05 232.722
49.50 548.93 11
6 1/14/96 40.0
0.0
700.00 43.40 10.15 181.006
38.50 426.94 12
1 1/21/96 40.0
0.0
340.20 21.09 4.93 87.968916
18.71 207.49 13
2 1/21/96 40.0
0.0
750.00 46.50 10.88 193.935
41.25 457.44 14
3 1/21/96 40.0
2.4
545.00 33.79 7.90 140.9261
29.98 332.41 15
4 1/21/96 40.0
6.7
738.49 45.79 10.71 190.9581624
40.62 450.42 16
5 1/21/96 40.0
0.0
900.00 55.80 13.05 232.722
49.50 548.93 17
6 1/21/96 40.0
5.0
831.25 51.54 12.05 214.944625

Trojan horse. Eventually, data entry moved out of the glass house to the ware-
house receiving dock to be used for inventory receipts; the order entry desk to
be used for new orders; the purchasing department to be used for purchase
orders; and, in the case of retailing, on to the sales floor for point of sale pro-
cessing. No longer were trained data-entry operators responsible for the qual-
ity of the data; others were responsible for entering data, and it was just an
ancillary part of their job, for which they were not necessarily even trained.
The control environment was breaking down, and the introduction of the
personal computer only complicated the issue. No longer was control central-
ized. While access to data could be controlled, control over the use of data and
the content of reports was lost. For example, two people could each issue a re-
port on sales, and the numbers could easily be different. Yet, both reports
could be accurate. How is this possible? Simple. One of the reports may have
been about gross sales and the other about net sales, or one may have been
based on data through Friday and the other on data through Saturday.
552
Making Key Strategic Decisions
When all programming was controlled by a small professional group, con-
trol was much easier. Because today’s spreadsheet programs are user friendly,
however, and software does not require programming knowledge, everybody is
his or her own programmer. Thus, it is difficult to control the consistency of
the information that is being distributed.
The problems only become more complicated. Now companies allow their
business partners, vendors, and even outsiders to access their computers, using
the Internet and EDI. Data is interchanged and moneys are exchanged elec-
tronically often without paper backup. While technology can prevent most
unauthorized access to data, as recent history has shown, even the U.S. De-
fense Department has not successfully prevented the best hackers from ac-
cessing its computers and wreaking havoc. What was relatively simple to
control before 1990 is now a nightmare. Accountants, systems professionals,