Pinnell/Busch: Project Management Consultants for the Design & Construction Industry

American City & County Magazine - July 1988
By Steven Pinnell

Managing a capital improvement program is one of the most difficult tasks facing a public-works director or city or county engineer. Funds often are inadequate, staffs are too small, priorities change without notice, costs overrun and schedules slip. Yet, most accounting and reporting systems do not provide timely, useable feedback to facilitate solutions to problems. Better management procedures and software can reduce these problems.

Five elements to successful program management exist. The first step is to develop a five-year plan. Define and prioritize programs and individual projects, based on general goals and specific objectives.

The second step involves establishing an annual budget. Fund and schedule project milestones based on the five-year plan, historical costs and available resources, such as staff and equipment. Develop a program work plan, adjust schedules for staff availability, and prepare a schedule for each project.

Third, estimate the required resources, such as staff, for each activity based on previous standards, with adjustments for levels of difficulty and staff productivity. Forecast resources requirements over time and adjust project schedules to fit available staff.

Fourth, delegate responsibility. Assign each project a project engineer or manager; who must commit to a budget, schedule and staff assignments.

Finally, monitor and report on each aspect of the project monthly. Report progress on three levels: project engineer/ manager, program manager (engineering or public-works manager), and elected officials. Analyze the actual implementations vs. the planned implementations, and the impact of those changes. Then, reprioritize, reschedule and rebudget the projects as needed. Document actual cost, resource productivity and scheduling data for future planning and budgeting efforts.

The reality is actual operations never work that easily. Accounting, fiscal management, and most existing public-works software packages cannot generate a program work plan and, do not provide all the information needed for program control.

Although project-management techniques generally are understood and several good project management software packages are available, good management procedures are not always followed. Program management (the management of individual projects within a program and the overall management of the program itself) often is misunderstood, and few software packages have the required flexibility and capacity.

The first task is to document existing project- and program management procedures, and correct any discrepancies. The second task is to select the best software to plan and control the program.

For managing an individual public-works project, almost any project management software will suffice, providing it is easy to use. A capacity of only 100 activities per project is sufficient, and a hand-drawn bar chart probably will meet most engineering needs. However, a manual or simple computerized system will not work for overall program management because it does not permit tracking of total resource, requirements, cannot handle a large number of projects, and lacks the flexibility to meet the different requirements of most public-works agencies.

The third task is to prepare a project manager's procedures manual integrating the new procedures and the software. This requires an in-house or outside consultant, with the expertise and time to do the job.

The fourth and most important task is staff training. All managers, project engineers and key support staff should attend a two- or three-day seminar on management techniques, and the newly implemented project/program-management procedures and software. This assures everyone speaks the same language and understands the process.

Another important task is to assign responsibility and authority for each project to an individual project engineer or manager. Finally, implement and verify the new system is being used and then improve it.

Before selecting the software, the degree of central control of the program database (the project data files) should be determined. Also, whether to operate a single-user PC or a multi-user environment (LAN network, UNIX microcomputer or mainframe) should be considered. PC-based software is: easiest to use with a few of the more powerful packages ported to UNIX. Program- management software is available on mainframe but is expensive and not available on many computers.

Initially, central control of the software is needed on data entry, processing and reports even with project responsibility distributed to individual project engineers and managers. Training one person to use a new system is easier with others gaining familiarity and access over time. Also, less incorrect data entry or loss of data will occur if only a few people use the software directly. Software that cannot be upgraded to multi-user operation should be avoided because changing software is difficult and expensive.

Almost any size public-works programs can be handled with a 386 microcomputer with a fast, hard disk. RAM-based software can only handle the data requirements of a small agency.

The project engineer or manager can access data directly after the database is established, everyone is trained in the new procedures and adequate safeguards are installed. But management-imposed requirements, such as budgets and milestone dates, and the manager's original commitments must be maintained. This requires considerable experience and sophisticated software, and should not be attempted during the first year.

The software must be able to batch edit and copy standard networks into the program database. to avoid tedious input and reduce errors.

The standard networks should be simple, from as few as 10 to as many as 30 activities for medium-sized projects, and a maximum of 100 for large projects. One of the most frequent and serious mistakes is to track too much detail too soon. The system will bog down and be dropped. Adding detail later is better, after the system has been in operation for a year.

However, the initial network and activity coding must be designed for easy expansion, with a hierarchical structure for consistent coding between small and large projects. Eventually, network activity codes can be integrated with cost accounting.

A frequent comment by public-works managers is “but our needs are different." In some respects, his is true. Although project-management procedures are similar across all industries, technology and reporting requirements are different from one public-works agency to another.

Menus and screenforms should be customized, if more than a few people see them. But sorts, selective prints, and report formats must be customized without program modifications. The software must be able to import and export data from accounting systems, and with word-processing and database management software.

Monthly reporting is recommended, coinciding with payroll reporting so effort expended can be compared to effort planned and work accomplished.

Both tabular and graphic reports are needed, with detailed tabular reports for analysis and project level control, graphic output also is needed for project level control, as it can show what otherwise would take hours of study and years of experience to grasp. Color graphics are best for public-works managers and elected officials.

Three levels of reports are needed details For project engineers and managers, summaries for the engineering and public-works managers and highlights for elected officials. Although the focus usually is on managing design, a good software package also will track the contractor's schedule, forecast survey and testing staff requirements, and even create progress payment reports.

One example of a successful public works program-management system is King County. Wash. In addition to well-considered requirements and reasonable expectations by department managers and their consultants, a key element in the system's success was a seminar and workshop for 30 people in engineering services, plus the publics-works director, manager of engineering services and key managers from other sections.

During the training, work plans were developed; along with resource estimates and schedules for many of the current ongoing projects. Subsequently, four or five standard networks were developed covering the majority of projects.

Six months after the system's implementation, a management audit by the county auditor found substantial improvements in project planning and monitoring. There were substantial increase in communication among staff, and coordination with other divisions. Peak staffing requirements were minimized by permitting adjustment of project schedules.

Construction administration and inspection staff needs were reduced, and less, expensive, temporary summer employees were used more frequently, allowing a smaller permanent staff during winter. Temporarily transferring engineers between design teams meets short-term workload peaks. Improved reporting to top managers enables them to monitor performance more closely and respond more quickly to essential delays or acceleration. Finally, projects advertised on schedule increased from between 37 percent to 54 percent for the previous three years, to 82 percent for the first year with the new system.

The county also switched from mainframe contract payment software, which cost more than $10,00 per year, to the progress-payment module of the project-management software for a onetime cost of $2,500. In addition to saving money, the system requires less staff time, is more flexible and faster.

Engineering services sanctions and public-works departments can benefit from better project and program management procedures and software. But a significant commitment in time and money is required to make it work.