Chapter 4 • The Data Resource 109
to have easy access to the data they need in their jobs.
Because security is so important and can be dysfunc-
tional if managed improperly, security should be con-
sidered when databases and application systems are
originally built and not developed as an afterthought.
8.Account for Use Because there is considerable
cost to capture, maintain, and report data, these costs
must be identified and an accounting system devel-
oped to report them. Further, an organization might
choose to distribute the costs to appropriate respon-
sibility centers. Two conditions make accounting for
the use of data especially difficult as compared to
other information resources. First, frequently the
organizational unit responsible for acquiring data is
not the primary user of the data. Second, usage is
shared because data are not consumed from usage.
The operating system and database management
systems can capture the actual costs of computer
disk storage and computer processing time. The real
issue is to develop a fair charging scheme that
promotes good management of data but does not
deter beneficial use. Because the value of data is so
elusive, the linkage of readily identifiable costs to
value is difficult. At a minimum, the costs for data
storage and processing and who uses which data can
be determined. Of course, how to charge to recover
these costs is a separate and more difficult issue.
9.Recover/Restore and Upgrade When an asset
becomes old or damaged, it is often renovated and put
back into operation. When an employee’s skills
become obsolete because of new technology or meth-
ods, the employee is trained for the new environment.
The same process is true with organizational data.
When a database is damaged because of some hard-
ware or software malfunction, procedures must be in
place to restore the database to a clean and uncontam-
inated condition. Usually, periodic backup copies of
the database will be made and an electronic log will
be kept of updates to the database so the restoration
can happen quickly. The business manager must
anticipate what needs to be done in the business when
a database is not accessible because of a recovery or
upgrading that temporarily takes the database out of
action. In addition, the business manager must be able
to determine what wrong actions or decisions might
have been taken from the bad data and correct them
before they cause excess costs or other problems for
the business. For example, if an inventory file has
been inaccurately changed and inventory replenish-
ment orders have been written, an inventory control
manager should immediately analyze whether work,
purchase, or expedited orders should be recalled.
10.Determine Retention and Dispose Business man-
agers must decide, on legal and other grounds, how
Good, Clean Data
The costs of dirty data and poorly managed data are staggering. According to a survey by
PricewaterhouseCoopers, poor data management costs global organizations more than $1.4 billion per
year. A study conducted by the Data Warehouse Institute found that erroneous mailings cost U.S. busi-
ness about $611 billion in 2002.
Data quality is a critical issue that is slowly coming to the forefront in businesses today. Two key
reasons for poor data quality are the rush to install new systems and the failure to take an integrated
view of the organization’s data. In the rush to install a new system, many organizations fail to look crit-
ically at their existing data. Often, these new systems support a single function, and while the data
might get cleaned up for that function, the source of the errors is often overlooked.
There are several critical steps in a data quality initiative:
- Establish processes to prevent errors.
a.Assign a person in each area to verify the data’s quality.
b.Structure databases so that fields are consistent across areas. - Clean up existing errors.
- Verify that third-party data suppliers provide clean data.
- Focus on the critical data.
- Don’t rely too heavily on cleanup tools; focus on prevention instead.
[Based on Betts, 2001; Turek, 2003]