Achieving user requirements for GIS implementation projectst
Neil Tansley
Severn Trent Water, Edgbaston Depot Waterworks Road, Edgbaston, Birmingham B16 9DD United Kingdom, Kristen Freese Stoner Associates, 1170 Harrisburg Pike Carlisle, PA 17013, USA Overview Severn Trent Water (STW) began two interrelated projects in 1998. The Records and Drawings Automation and Regionalisation (RADAR) and Underground Asset Data Management Systems (UADMS) projects have the objective of delivering a new records management process supported by a GIS-based asset information system. The project delivers a new Record Management Centre, operating a standardised, re-engineered business process, and supplies endusers with standard GIS-created records products. Severn Trent Water, Stoner Associates, and Severn Trent Systems (STS) have involved STW Records technicians and operational staff in the process of developing standardised records management work processes, a corporate asset data model, mapping standards and customised GIS applications. The initial step of the RADAR/UADMS project team was to involve Records Office staff in defining a re-engineered work flow for managing records data. This work flow forms the core component of the project and enables all other workstreams, database design, data conversion, application design and infrastructure design, to successfully meet user and process requirements. Ongoing STW involvement, through design support groups and key deliverable quality reviews, has achieved early prototyping, testing and refinement of the IT solutions. This paper discusses approaches used to date to ensure that business requirements are satisfied and to achieve business acceptance of both organisational and technology change. Project Management Approach The RADAR/UADMS project management team consists of a STW Project Manager, STW Business Change Manager, a STW Contract Manager, and two managers from Stoner and STS. The objective of this team is to deliver the project benefits to the defined timescale, budget, and quality level. The management team has adopted three key programs aimed at achieving benefits delivery. Quality Review Process The Quality Review Process was established to ensure project deliverables met pre-defined quality standards. In addition, the process develops peoples' understanding of specific project areas and ensures end-user 'buy-in'. The process identifies individuals who have the skills and experience to contribute to each project deliverable. Selected individuals are then invited to read and critique deliverables before they are approved. All reviewers are invited to a review meeting with the author to debate deliverable issues and agree modifications. An authorised management team member approves the deliverable once it has been updated to reflect the changes agreed at the meeting. Configuration Management Configuration management is the collective term that the management team uses to describe the management of project risks, issues, problems, and changes to scope. Any employee of STW or the project team is invited to raise a concern regarding a specific area of the project. Twice a month, several members of the management team meet to review and classify incoming concerns and ensure that existing risks, issues, and problems are being given appropriate attention to either mitigate or resolve them. This process ensures that potential threats to the successful delivery of benefits are carefully managed. Stakeholder Management One of the most common reasons for IT project failure is the lack of support and involvement of senior management (Whittaker, 1999). The RADAR/UADMS project is actively working to maintain management interest. The project management team meets formally with the STW senior operations managers, the RMC customers, on a regular basis. Additionally, the project management team meets with the sponsoring Director once a quarter to inform him of project progress and maintain support. A project communication manager, reporting to the management team, develops and promotes regular publications and presentations summarising project news. Each STW project team member has an assigned group of business stakeholders who they meet with regularly to provide project updates. The heart of this program involves STW business people sharing their project experiences with colleagues and the project team. Additionally, as the paper later describes, the team has set up several business support groups that provide input and timely feedback to the suppliers. The majority of the support members have full time operational or planning jobs within STW. Their involvement in the RADAR/UADMS project is on a volunteer basis only. Strong stakeholder management has been essential to make this support available. Business Process Development The core component of the RADAR/UADMS project is to change the way asset records are managed within STW today. Ten records offices, each representing different geographic areas of the company, provide mostly paper records products (displaying underground assets) to the operational and planning staff within their geographic area. Each office has traditionally supplied varying products to varying presentation standards. Map symbology, record content, and record product ranges produced vary significantly in each of the offices. The project seeks to gain efficiencies by standardizing processes and focusing the scope of work performed. The existing offices will be closed down and a new RMC will be opened, supporting the whole of STW. The operators in the RMC will use streamlined work processes to deliver standard products supported by customized GIS applications. Recognizing the impact these changes would have on the business, the STW project managers needed to take early ownership of the business process re-engineering work. Initially, the project consultants led a small group of records office representatives (all from different offices) through a week long workshop to understand and re-define the existing processes. After high level processes for updating, viewing, and distributing records were agreed, STW took ownership. Since those early days on the project, STW project members and project consultants have developed much more refined processes using simulation modeling CASE tools. STW initiated a records focus group, represented by at least one member of each of the existing records offices. Presentations and quality review sessions with this group on the re-engineered business processes was a key step in winning business support for this major organizational change. This communication exercise is ongoing as both the development and enhancement of process and IT continues. This paper will later describe how the new business processes are being thoroughly tested and accepted by business users and customers before implementation. Data Conversion The UADMS/RADAR project is converting all of the STW underground water asset data from paper to digital form. Additionally, the project will translate and move electronic asset data from two existing GIS systems into the UADMS database*. Data conversion activities account for over 30% of the total project budget. It is extremely critical, not only for business acceptance reasons, but also for regulatory and safety reasons, to successfully convert data. If, for example, a few valves are inadvertently left out of a converted record, a serous risk of failure could be introduced for operational records users. Unless business confidence is gained in the converted data, users may revert back to using paper plans and the project benefits will be lost. STW has taken the approach of getting key people involved early in the conversion specification process. A team of Records Office representatives was formed to develop a company-wide map * STW sewerage underground asset data resides in approximately 40 separate GIS databases that, until recently, were maintained by local council offices. Additionally, one of the existing ten records offices maintains water asset data using GIS. Standard. The map standards group, as they were called, was business led. Within a few months, they agreed a standard set of symbology and defined the business requirements for a set of nine different records products. This was a great accomplishment since STW produces over forty different records products today, each using different symbology. STW and Stoner are building "map exhibits" from the data sets maintained within each of the ten records offices to 'prove' the converters understanding and interpretation of existing data. A map exhibit is a cut and paste of an asset on an original record that is referenced against the asset name and the new symbology post-conversion. For example, a map exhibit of a fire hydrant would identify what a fire hydrant looks like on the records today and what it will look like after conversion. Developing exhibits for each type of asset within each office builds confidence in the conversion process and sets expectations of what the converted records will look like. After the set of map exhibits were created for the first of the ten offices (Birmingham), Stoner and STW began a data conversion process prototype*. Stoner converted an area of ten maps. STW records office staff then went through the maps and highlighted problem areas. This process was repeated multiple times for the same map set until agreed acceptance criteria for positional accuracy and text placement standards was achieved. STW viewed the prototype data using GIS tools in the "Model Office". This is a ring-fenced project IT environment established to provide STW a place to examine and test data deliverables and software. The UK-based project team works in an office adjacent to the Model Office. STW staff company-wide were invited to visit the Model Office to review the prototype data. STW plans to use business people, if resources are available, to validate data deliveries in the Model Office environment. This will achieve greater business acceptance of final data deliverables. Application Design The UADMS application detailed design and development is divided into four separate stages. Each stage of application design focuses on delivering specific functionality to support business processes. The application design must incorporate the business requirements to achieve the defined business benefits. A STW application design support team (ADST) was established again with business people from outside the project. This team meets regularly with Stoner's application designers to work through 'storyboard' designs**. The objective of these meetings is to map the proposed functionality against the re-engineered processes. Consensus is sought from the business participants as to whether the functionality will support the process. A roadblock to successful discussions has been the ADST's lack of experience with GIS products. With never having seen GIS applications in use, it has been difficult for business staff * It should be noted that STW is based in Birmingham, United Kingdom. The Stoner data conversion team is based in the USA. Therefore, communication between teams was limited to occasional trips, frequent phone conversations, emails, and faxes. ** The Stoner application design and development team is also based in the USA. A Stoner UK-based employee maintains the Model Office environment on behalf of the project team. to imagine what additional capabilities can be delivered over existing records practices. In part to overcome this, early prototypes of the customised GIS applications were installed in the Model Office. The project adopted an iterative, schedule-oriented software development method whereby Stoner delivers new software to the model office on a monthly basis. The ADST meets regularly to review the newly delivered functionality and provide feedback to the design and development team*. The advantage of the chosen design and development methodology is that the end-users have something real to use and test. They are empowered to contribute to the design by taking a "hands-on" role during the design cycle. Ultimately, it reduces the risk of developing an application that fails to provide the tools required. Because limited detailed design work takes place prior to the start of development, the project team has learned that short, simple design memos addressing detailed functionality issues save time and effort in later stages. Once a design concept has been agreed between STW and Stoner, the application designers document the details to ensure that the proposed functionality meets the business needs. Database Design It is fair to say that the project team underestimated the amount of effort, time and business resources required to build an underground asset logical data model (LDM). Strategically, STW decided that the database design must have the flexibility to grow over time. The data modeling activity required looking beyond the RADAR/UADMS project. The model needed to be compatible with other business processes that will one day take advantage of GIS. As a result the RADAR/UADMS data modeling work laid the groundwork for development of a corporate data model. Completing the data model and database design is a critical activity that needed to be completed before data conversion or application development work could begin in earnest. After a few false starts where a project consultant attempted to develop a data model in isolation, the project management team agreed to pull a team of people offline to complete the work. A dedicated database design support team (DDST) was selected, made up primarily of project team members and representatives from the STW IT strategy department. After a few months of dedicated work, the DDST successfully developed a UADMS LDM. The final version documents over 100 entities and 1000 attributes, with relevant valid values! The critical success factors were moving the DDST to a central, off-line working environment**, interfacing with other ongoing projects within STW, using the skills of experienced data modeling consultants, and having the contribution and support of the IT strategy department. A risk acknowledged with the new data model was that very few end-users were involved with the development. Additionally, a data model is not a document an average person can pick up and easily interpret. To obtain business buy-in, the DDST focused its efforts on visiting various * STW plans to change the ADST membership for each stage of application development. ADST members are selected based on their business experience. ** The Stoner team leader made multiple 2-3 week long trips to the UK to be onsite with the DDST. work areas in the business and presenting relevant data model relationships. The distribution of thick documents and complicated diagrams were successfully avoided. During each presentation, the DDST team aimed to use examples relevant to the audience. The initial physical database was evaluated and tested in the Model Office at the same time as the data conversion prototype work. Similar to the iterative method used to agree the data conversion standards, the physical database went through a few iterations with minor changes. Comprehensive business input early in the design process proved beneficial. Business Pilot Prior to the company-wide implementation of the RMC and re-engineered business processes, the project team will run a two to three month pilot program designed to trial the processes using electronic data, software, and technical infrastructure. A business pilot program will be run for each of the four application stages. At the time that this paper was written, pilot planning for the first stage was well underway. Objectives of the pilot program include:
Conclusion At the completion of the RADAR/UADMS project, Severn Trent Water will operate records management from one central location using re-engineered business process that deliver standard GIS-created records products. Recognizing that the key to success is based on the business acceptance and buy-in to both the organizational and technology change, the project team is taking great steps to involve end-users and key stakeholders. Records technicians and operational staff have contributed to the design of the new business processes and supporting software, the creation of the map standards, records product layout, and the development of an underground asset logical data model. The project team has sought and encouraged business involvement in key deliverable quality review sessions and design support teams. This is evidence to the team's adoption and implementation of an iterative design methodology. For most arms of the project, the STW, Stoner and STS teams have started with a basic understanding of the business requirements, proposed a solution and then sought business review and recommendation. After modification and further development, the business is asked again to provide input. The cycle continues until the agreed quality criteria are satisfied. This approach ensures end products that are well matched to the business requirements and perhaps, even more important, a staff of project champions each whom, through their 'voluntary' contributions, have some form of project ownership. References Whittaker, B., 1999, "What went wrong? Unsuccessful information technology projects": Information Management & Computer Security, 7/1, p. 23-29. | ||
| © GISdevelopment.net. All rights reserved. |