Establishing Enterprise Data Standards - High-Performance Web Databases: Design, Development, and Deployment - page 18

prefaced with their datatype, all databases will have project names). It is
unlikely that this first pass will be complete. The primary challenge is to
establish a standards baseline and to begin using them. As in the “borrow”
approach, one or more pilot projects are used to validate and enhance the
data standards. In the “create” approach, it is likely that several projects
will be required to define a satisfactory set of data standards.
There is a third approach that involves combining the “borrow” and
“create” approaches. In this instance, standards organizations such as
ANSI or ISO can be used to provide a first cut at the data standards. These
are then customized as discused in the “create” approach.
Exhibit 7 shows that after the enterprise data standards have been pre-
pared and signed off by the corporate sponsor, the next task is wide
deployment. The manner in which this is done is critical for the successful
adoption of the enterprise data standards. The first step here is to ensure
that the data standards are readily available electronically or on paper. The
next step is to communicate the importance, location, and procedures for
using the data standards by both the operations and development groups.
This message must be communicated to the organization by a senior spon-
sor, such as a CIO or VP, in order to maximize corporate buy in.
COMMON PITFALLS AND TRAPS
There are several common pitfalls and traps that can befall an organiza-
tion trying to establish enterprise data standards. One of the obvious prob-
lems springs from the “enterprise” term itself. Many organizations view the
use of this term to mean “broad consensus required” or “slow adoption.”
This can lead to several problems, not the least of which is that the stan-
dards require too much attention and time to implement, or are too cum-
bersome to learn, use, or understand. Such standards are usually put aside
during tight project timeframes. Too many standards can be as bad as hav-
ing too few standards. The surest way of ensuring that standards are not
used is to spend a lot of time defining them and publishing them in a hun-
dred plus page manual. It is preferable to spend a few days conducting
research, and then having a small team publish a single digit page docu-
ment defining the suggested data standards which are then immediately
passed to project teams to test and refine in actual projects.
Another common problem involves the inaccessibility of enterprise
standards. To be effective, the standards must be readily accessible to
project teams at any time. Furthermore, the standards should also be
readily usable. They must also be easy to duplicate. A useful method of
achieving these is to use a combination of deployment strategies. A central
data repository (e.g., a Web site or a Lotus Notes database) can hold the
enterprise standards. Project teams throughout the organization should
have direct and remote access to the repositories. Reusable templates and