372 Part III • Acquiring Information Systems
Advantages- Highly structured, systematic process
- Careful controls on the process, including impact on related systems
- Thorough requirements definition
- Clear milestones with business management sign-offs
Disadvantages- Does not account well for evolving requirements during project
- Time-consuming (and costly) process
- Top-down commitment required
- Resistance in practice to feedback or going back to a prior phase
- Role of system user may be narrowly defined
- Fixed milestone and completion dates
FIGURE 9.7 Advantages and Disadvantages of Traditional SDLC ApproachIn the hands of competent IS specialists and knowl-
edgeable business managers, the SDLC is a systematic
process with formal steps with clear IS and user roles, for-
mal checkpoints, and techniques for analysis, design, test-
ing, and implementation. These tools and the rigorous dis-
cipline associated with an SDLC methodology help the
systems project manager produce a well-engineered sys-
tem on time and within budget.
Second, the systematic process includes controls that
assess the impacts the system under development will have
on existing systems. These controls address the concepts
of the organization as a set of systems, with carefully de-
signed interfaces between subsystems.
Third, requirements are formally and thoroughly de-
veloped. All parties understand and agree on the require-
ments. Thus, the system is less likely to be idiosyncratic,
rather it will balance the needs of all users.
Finally, the process is transparent with clear mile-
stones in which interested parties sign-off that further
funds should be allocated to continue to the next step. The
milestones also give participants a formal point to identify
the need to backtrack to a prior phase or step, if the situa-
tion suggests the need to reconsider prior design decisions.
The major disadvantages are inherent in the methodol-
ogy. First, the project’s success depends on the accurate and
complete specification of detailed requirements at the begin-
ning of the development process (Definition phase). There
are several serious problems with this dependency. For
example, many customized applications today are unique
solutions. Because the project begins with an incomplete
understanding of what this unique information system will
do, it might be necessary to try several approaches before dis-
covering the optimal one. New technologies might also be
involved, and until the capabilities of these technologies are
better understood, it might be hard to develop a firm set of
requirements or to estimate the time to perform some project
steps. With pressure to meet what were ill-conceived dead-
lines, shortcuts can be taken that affect project and system
quality. Another problem with up-front detailed requirements
specification is that today’s business environment is chang-
ing so rapidly that there can be significant differences in busi-
ness needs between the time the requirements are specified
and the time the system is installed. Although the SDLC
allows for backtracking to previous steps if necessary, this
feedback loop is often ignored in practice. Rather, the output
of one step is frozen; requested changes are recorded and
handled only during maintenance.
Note that the SDLC process also requires a total
system cost-benefit analysis based on the initial Definition
phase. The justification process can be difficult to accom-
plish using traditional approaches such as return on invest-
ment (ROI) calculations when new technologies are
involved or requirements are incomplete.
Second, the SDLC process is time consuming. In the
1980s, the typical systems project took several years, and
the pace of business change accepted this. Today, with per-
sonnel turnover, threats from new competitors, and condi-
tions changing at Internet speed, such long delivery times
are not acceptable. Third, because the SDLC process is
both lengthy and costly, strong executive sponsorship is
required. Without strong business sponsorship, business
managers and users will be reluctant to dedicate their time
to a systems project instead of working on other activities
for which they are typically measured.
Fourth, although in theory the SDLC allows for feed-
back and backtracking, there is strong resistance in practice to
doing so. At each milestone, there is a “go-no-go” decision on
whether to move forward: There may be some willingness to