Completing the "Big Dig": Managing the Final Stages of Boston's Central Artery/Tunnel Project (2003)

The processes for reporting and controls have changed significantly over the course of the CA/T project. During the 1990s a comprehensive project database was developed, using state-of-the-art relational database software, to enable the detailed tracking of individual contract tasks and the assembly of cost reports. Since no off-the-shelf programs were capable of handling such a mammoth project as CA/T, management consultant Bechtel/Parsons Brinckerhoff (B/PB) developed a formidable construction-information system (CIS) using a commercial software system. The CIS, which was not fully developed and available for use until about 5 years after construction began, tracks almost every aspect of the project. That includes all design and construction contracts and their funding sources, requests for information, change orders, submittals, design specifications, accounting and budgeting, forecasting, staffing, scheduling, and reporting. However, there continued to be major problems—in the flow of information to oversight groups, in the provision for adequate contingency, and in the inconsistency between project scope and cost estimation. In 2000 the MTA announced a substantial project overrun, to the apparent surprise of the FHWA oversight group. Following that event, the reporting and control processes were modified to track budget and schedule progress to a fixed project scope.

Using the CIS, the project team generates a large number of progress reports, providing what the committee believes might be an excessive amount of information (6,400 data tables, 167,000 columns) for managing the project. Much of that information lacks the strategic view needed to develop a reliable construction schedule, though this information may serve a purpose during the operation and maintenance of the highway system.

Project reports for various time intervals are regularly scheduled. A progress report on each active contract is circulated daily via email to resident engineers, project managers, and oversight agencies in order to summarize work progression, quality deficiencies, and identified needs for modifications. These daily reports are based on the detailed work logs maintained by the resident engineers.

A monthly set of reports is produced to identify exceptions (or variances) to the existing budget and schedule. In addition, a set of month-end documents for each contract is developed and reviewed by senior management and oversight agencies to inform them of issues associated with contract modifications and claims. For example, a Budget, Cost, Commitment and Forecast report is prepared for financial review by oversight agencies. In addition, a Project Management Monthly report is prepared for senior management, oversight agencies, and the public. It summarizes the status of budgets, cost exposures, schedules, safety records, and employment-

diversity records for the entire project, but it does not include detail on the status of individual contracts.

A major annual report for the project is the Cost and Schedule Update (CSU) prepared for oversight agencies and bondholders. Since 2000, the accounting firm Deloitte & Touche has conducted independent reviews of the CSUs on behalf of the Massachusetts Executive Office for Administration and Finance. The Federal Highway Administration also reviews the CSUs and prepares independent estimates of completion costs and schedules.

Cost control relies on a resource-loaded work-breakdown schedule. Each contract is divided into specific tasks with required resources and schedule milestones, and detailed information on specific tasks can be pinpointed or aggregated. As contract modifications are implemented, additional tasks may be added to the schedule and the work-breakdown structure.

The implementation of the IPO has complicated the control of expenses for the B/PB management-consultant team. MTA personnel, including the director of finance, review and approve invoices from the consultant. Annually, the staffing of the B/PB management consultant team is reviewed, with the expectation that time commitments will be reduced as the CA/T project concludes more contracts and moves toward completion.

Schedule control is maintained by monitoring work-in-progress and planning future tasks. Contractors develop detailed schedules for their own work that are aggregated into the overall project schedule, and milestones are set for individual contracts and the project as a whole. Scheduling derives from a commercial software package, Primavera™, in widespread use in the construction industry.

This software is being pushed to its limits—the current schedule for completion of the CA/T can be regarded as tight, with numerous overlaps in activities for finishing work and aggressive estimates of completion time for operational requirements—and in any case cannot compensate for problems on the work site. Indeed, the milestone date for opening I-90 was extended 81 days and the milestone date for the initial I-93 Northbound opening was extended 53 days between the forecast in CSU 8 (August 2001) and in CSU 9 (July 2002) (Deloitte & Touche, 2002).

Control of the project scope is a critical issue on a large, complicated project such as the CA/T. Since the original $2.6 billion estimate (in constant 1982 dollars), numerous additions have been made, especially for mitigation of environmental or economic impacts and to adjust for unanticipated site conditions. Currently, the senior project management reviews significant scope changes as part of the contract-modification process. As the project enters into its final stages, however, the committee expects that changes in scope should be less and less of an issue.

Project reports since 2000 have been consistent with state-of-the-practice on large construction projects, though the committee believes that the CA/T would have benefited greatly if the above-mentioned reporting procedures—especially the annual cost and schedule updates and the improved information flow to oversight agencies and the public—had been implemented

much earlier. The committee notes, however, that the public-information website appears to be a useful and effective outreach tool.

The CA/T engineering design and construction functions are separate, without a project manager responsible for the completion of distinct components. This lack of an overarching project-management organization hampers control and the transition between functions; it may well have contributed to the large number of required contract modifications.

Scheduling continues to be a problem on the CA/T project, as evidenced by the fact that milestones continue to slip. Scheduling might be improved by adding some flexibility, with tracking for “optimistic,” “most likely,” and “pessimistic” estimates. Also, wider use of incentives tied to task completion could improve schedule control.

Finding 8: Too much uncoordinated information is provided to the project’s senior management, and project reports do not lend themselves to ready analysis—particularly with regard to such concerns as the uncertainty of forecasts and the trade-off possibilities between time reductions and cost increases.

The CA/T project staff routinely generates extensive cost and schedule reports, conducts regularly scheduled meetings, and is able to provide extensive information on a myriad of project details. Special reports can be developed on request, as evidenced by several special studies developed within a day for this NRC panel. Summary information and strategic issues, however, need to be emphasized. For example, the Project Management Monthly report provides good information about the process of construction, but it is less effective for contingency management or identification of critical issues.

Recommendation 8: As the project foci shift—to managing claims, cost-recovery processes, and the transition to operation—project reports should also change to emphasize summary information and strategic issues.

Reports should focus management attention on critical management issues such as contingency management and critical-path schedules. Summary information and strategic issues need to be emphasized in the project reports.

Finding 9: The MTA has developed an aggressive and optimistic schedule for project completion. However, the project milestones are prone to slippage as events occur.

The project’s monthly reports include milestone dates and contingency dates, but they do not necessarily offer realistic completion dates. This slippage creates problems of public credibility for the CA/T management and the MTA.

Recommendation 9: Project reports should reflect the probability of meeting milestones or the probabilistic distribution of completion times, based on the results of past efforts.

Project reports include schedule milestones and the budgeted contingency milestones, but they provide little help to reviewers regarding the likelihood that these milestones will be reached on time. Additional information, such as optimistic, pessimistic, and most likely milestone-time achievements, would be helpful. These estimates could be based on the scheduling experience garnered over the course of the project.

Deloitte & Touche. 2002. “Central Artery/Tunnel Project: Project Assessment.” Prepared for Massachusetts Executive Office for Administration and Finance. September 30, 2002.