Case Study III-1 • Managing a Systems Development Project at Consumer and Industrial Products, Inc. 439A few weeks after the start of the project, Watkins
obtained Lucy Robbins from a contractor firm to be her
lead analyst. Robbins had managed a maintenance area at a
medium-sized company and had also led a good-sized
development project. She could program, but her main
strength was in supervising programmers and communi-
cating with the technical specialists in IS. Watkins was
able to delegate much of the day-to-day supervision to
Robbins so that she could concentrate on the strategic
aspects of the project.
The Stradis methodology required the use of a
CASE tool, and Robbins became the CASE tool “gate-
keeper” who made sure that the critical project information
stored therein was not corrupted. She said,
We used the CASE tool to keep our logical data
dictionary, data flow diagrams, and entity/relation-
ship data models. The CASE tool keeps your data
repository, and then uses that repository to popu-
late your data flows, data stores, and entity/rela-
tionship models. It also assists in balancing the
diagrams to make sure that everything that goes
into a diagram is necessary, and everything that is
necessary goes in.Because IS had far more projects under way than it
had good people to staff them, Watkins was never able to
convince Henry Carter or Clark Mason to assign a quali-
fied CIPI person to the project full time, so she had to
staff the project with temporary employees from outside
contractors:
After I determined what resources I could get from
CIPI, I would look at the tasks the project team
had to perform and then try to find the best persons
I could that fit our needs. I took as much care hir-
ing a contractor as I would in hiring a permanent
CIPI person. I tried to get people who were
overqualified and keep them challenged by dele-
gating as much responsibility to them as they
could take. My people had to have excellent tech-
nical skills, but I was also concerned that their per-
sonalities fit in well with the team and with our
clients.Watkins hired two contractor analysts who had
skills that the team lacked. One was a very good analyst
who had experience with CIPI’s standard programming
language and database management system and had been
a liaison with the database people on several projects.
The second contractor analyst had a lot of experience in
testing.
The project got excellent part-time help from data-
base specialists in the CIPI IS department. Watkins recalls:We used IS database people to facilitate data model-
ing workshops and to do the modeling. We also used
a data analyst to find a logical attribute in the current
databases or set it up in the data dictionary if it was
new. There were also database administrators who
worked with the data modelers to translate the logi-
cal data model into physical databases that were
optimized to make sure we could get the response
time we needed.Watkins also used consultants from the IS develop-
mental methodologies group:Because my supervisor was not experienced in
development, I used people from the methodolo-
gies group to look at my project plans and see if
they were reasonable. We also used people from
this group as facilitators for meetings and to mod-
erate walkthroughs, where not being a member of
the team can be a real advantage. Also, when we
needed to have a major technical review, the
methodologies group would advise me on who
should be in attendance.Carrying out the Project
The project began in mid-June 1993 as the CIMS
Replacement Project. The Computerized Invoice Matching
System (CIMS) was an old, patched-up system that
matched invoices to computer-issued purchase orders and
receiving reports, paid those invoices where everything
agreed, and suspended payment on invoices where there
was disagreement.The Initial Study
Because of strategic changes in how the department
intended to operate in the future, a number of significant
changes to the system were necessary. The project team
concluded that it was impractical to modify the CIMS sys-
tem to include several of these important enhancements.
Therefore, Shaw and Watkins recommended that a new
system be developed instead of attempting to enhance the
existing CIMS system. They also suggested that the scope
of this system be increased to include manual purchase
orders and some transactions that did not involve purchase
orders, which effectively collapsed two of the planned
development projects into one. At its meeting on August 8,
1993, the steering group accepted this recommendation