Successfully integrating multiple utilities and supporting technologies
A Multi-Phased Approch
The Feasibility Study
The process to implement GIS at JUD was a phased approach that began with a feasibility study
from February to August 1993. Through the study, the business case for undertaking GIS was laid
out, including how business is done without GIS, what specific areas can be improved through
the technology, and what benefits can be expected. All of these elements needed to be quantified
as precisely as possible.
Preparing a feasibility study is really an exercise in grass-rootssupport building for a GIS project.
It involves interviewing potential key user groups, management, executives, and information
system personnel. Education is also a critical objective for the feasibility study in that key users
and management must understand what GIS is and how it can affect their jobs and the
corporation as a whole.
Prior to the interview process, JUD conducted a company-wide GIS technology and project
education presentation. After the education session, key users in engineering, operations,
construction/maintenance, as well as “peripheral user groups” including customer service
representatives and accounting personnel were interviewed. Through the interviews and a series
of workshops, the GIS project team learned what each user group’s job functions are, explored
how GIS might affect that work, and examined areas where GIS can improve work processes. The
interview process results in a set of high-level requirements that form the basis for an
implementation plan.
The feasibility study at JUD concluded that the project appeared justified—at least on paper. The
feasibility study generated management support and funding for the next phase of the project
implementation: the Pilot Phase.
The Pilot Phase
The importance of the Pilot Phase comes into focus when one realizes that roughly 20 percent of
a project’s overall budget may be earmarked for pilot development, as was the case at JUD. This
type of expenditure was justified in that the Pilot Project was designed to test the validity of many
of the assumptions laid out in the feasibility study. JUD carefully designed the Pilot study to
accomplish the following objectives:
- Validate, on a small scale, feasibility study assumptions considered of primaty importance
including:
- Implementation costs based on bids
- Hard benefit assumptions
- Strategic benefits
- Implementation plan
- Develop and refine JUD’S data preparation (scrub) procedures
- Refine data conversion procedures
- Refine the database design by demonstrating functionality to JUD users, management,
executives, and city/county personnel
After management approval, JUD initiated the Pilot Project in October 1993. JUD’S multi-disciplined
GIS Project Team focused on the objectives and made a concerted effort not to
expand or change the carefully planned scope of work. The Pilot project culminated in three
days of extensive GIS demonstrations in December 1994 after 15 months of Pilot development.
Prior to and immediately after the demonstrations, the JUD Project Team studied the results of the
Pilot project by performing time and motion studies for productivity benefits determination,
revising the cost5enefit analysis using actual costs, bids, and revised benefit measurements and
estimates, and interviewing executives and management to determine whether they felt GIS could
help JUD achieve the predetermined objectives.
After careful study, the Pilot GIS was found capable of achieving key project objectives such as:
- Improving labor productivity to a higher degree than envisioned and estimated in the
feasibility study
- Improving asset utilization and decision support mechanisms
- Maintaining utility system reliability
- Helping meet regulato~ requirements
- Demonstrating the abi Iity to interface and integrate GIS with the AS/400 customer
information and work management systemswithout the duplication of data
- Demonstrating PC access to GIS, AS/400, and PC applications
- Demonstrating the effective consolidation of JUD’S 30-plus manual conversion sources
- Demonstrating basic AM/FM functionality such as land and facility data maintenance,
plotting, data @cry, and system navigation along with a number of prototype custom
functionality such as work design and gas network isolation
Full ln@meWWl .
Full Implementation
JUD executives approved full implementation in January 1995, 16 months after the Pilot was
initiated. The implementation plan for full implementation, revised from the Feasibility study
during the Pilot Project, initially focused on the following tasks:
- Enhancing/revising core functionality based on Pilot findings and user input
- Data conversion
- GIS user training - primarily data maintenance and system access/query
- Data acceptance testing, data validation, and quality control
Data conversion was seen as the critical path to completing the entire project, therefore,
considerable emphasis was placed on these processes. The conversion strategy involved
converting the low to moderate growth areas first, minimizing the work order posting required
after a specific area of data was accepted. Once the accelerated data preparation, data
conversion, and data acceptance testing processes were working wel 1,JUD’S project focus
changed to the fo Ilowing tasks:
- More GIS user training - primarily data maintenance and system access/query
- Refining custom application and interface specifications
- Enhancing the Oracle Gateway to the AS/400
- Maintaining two recordkeeping systems during conversion (GIS for accepted areas and pre-GIS
systemsfor areas not converted)
By mid 1996, conversion was roughly 40 percent complete, and JUD’S GIS data maintenance
software and processes were beginning to come to fruition. JUD finished posting the work orders
and began replacing existing manual systemswith GIS in those areas where data had been
accepted. The process of refining the application specifications included involving JUD
Table1: GIS Productivity Improvement
* plotting standard maps on GIS can be done as a batch process which does not require the full POSt-Pilot time.
