Page 79 -

64 MADSEN AND KAUTZ
which was made up of three to four people with one full-time developer who did most of the actual
hands-on work. All of the project team members had long formal educations in computer science
and from zero to fifteen years of practical experience, with the full-time developer being a newly
graduated master’s student with no prior industry experience. For this reason, the developer was
formally supported by a university-employed academic supervisor and an industry supervisor, who
was part of the company management and the daily leader of the market research department.
From the outset, the application was envisioned as an RDR to be implemented based on a com-
mercial Web content management system, which would support the market research department’s
internal work practices (i.e., the report production process) and external sale to customers through
storage and online analysis and reporting of data at a high granularity. Before project initiation,
the contingency approach Multiview/WISDM (Vidgen et al., 2002) was chosen as the formalized
method and used to inform the construction of a situation-specific method outlined in the form
of a detailed project plan, which was included in the original TCS project proposal and formally
approved by the TCS program. The development process was planned as a prototype driven ap-
proach where two of the department’s core products, paper-based market reports on the bottled
water and watercooler markets, would be used as the point of departure for implementing the first
working prototype. As the application was expected to be based on a Web content management
system, the development approach was planned with an emphasis on the Web-based front-end,
organizational change and implementation.
However, when the project commenced, the two paper-based market reports quickly led the
project team to discover that the amount of data and data relationships was extensive and an
initial market scan of available content management systems revealed that they were not sophis-
ticated enough to support that kind of data complexity. Thus, a few months into the project, the
project team decided that the RDR application would be custom-made and a software tool was
adopted to support the now more technically complex development task. Furthermore, the RDR
application was not developed with an equal eye to internal process improvement and external
sale as planned, but with an emphasis on the internal report production process. This was due to a
performed requirements analysis, which showed that the current report production process could
be greatly improved by automation of mundane tasks (e.g., data collection, data entry, and report
formatting). The project team and company management therefore decided to give priority to the
internal process. Eight months into the project, company management also decided that the RDR
application should be able to produce what they perceived as the market research department’s
main product, namely, paper-based reports. In contrast, the project team had expected all data
analysis and reporting to be online. The technical complexity of the system was ever-increasing,
and after ten months the project team decided that the planned three-layer architecture would have
to be expanded with an additional layer of prespecified data queries and report-formatting scripts.
The first working prototype was released for use by market researchers after approximately a year
and a half, but work to enhance and extend functionality continued throughout and well beyond
the project’s two-year time frame. The vision remained constant throughout the project, namely to
“create an enterprise repository for [the market research company’s] research data” (TCS project
proposal), but the practicalities of what (product) and how (process) to implement were continu-
ously reconceived and scoped.

CROSS-CASE COMPARISON OF THE EMERGENT METHODS

On first inspection the Multimedia and Web cases seem very similar. Both cases concerned tech-
nically complex information systems, were performed by relatively small project organizations