The Internal Revenue Service (IRS) has a goal that at least 80 percent of Federal tax and information returns be filed electronically by the year 2007. However, its ability to develop new product features allowing taxpayers and tax preparers to file, pay, and communicate electronically has been limited by its outdated computer systems.

To resolve this problem, the IRS is working with a private contractor, Computer Sciences Corporation (CSC), to modernize IRS business processes and technology. The IRS also established the Business Systems Modernization Office (BSMO) to assist and monitor the modernization efforts. One of the projects selected for development and deployment was the e-Services project. This project will provide the means for tax practitioners and other authorized parties to conduct business electronically with the IRS. The e-Services project will be deployed in various releases, with the first release scheduled in Fiscal Year (FY) 2002. The IRS projected that the project will provide $16 million in benefits in FY 2002.

The objective of our audit was to assess the IRS’ processes for development and implementation of the e-Services FY 2002 Release. To accomplish this objective, we evaluated the CSC’s delivery of goods and services, reviewed the project’s adherence to life cycle requirements, evaluated compliance with the Modernization Blueprint, and determined whether risks existed in the areas of project dependencies and sponsorship.

The BSMO has made progress in establishing critical processes to guide the e-Services and other modernization projects. For example, the BSMO has established a quality review process to evaluate, at the end of each major set of tasks, the products delivered by the contractor to determine whether they meet required standards. Also, the e-Services project team has begun to work more closely with the Architecture and Engineering Office to enable greater compliance with the Modernization Blueprint. In addition, project sponsorship has recently been enhanced, and business executives have become more involved in monitoring the progress of the project team.

Despite the progress that has been made, we identified several issues that have affected the success of the project thus far. Changes in requirements and setbacks in other projects on which the e-Services project is dependent resulted in delays and increased costs. Also, labor costs for IRS employees assigned to the project were not accurately tracked. In addition, we believe contracts issued to the CSC should be strengthened and project tracking measures expanded and validated. Finally, although processes to enable project success have been established, several need to be more effectively implemented to help enable the project team to meet its goals and objectives. Improvements in these areas will be needed to successfully deliver the benefits expected from e-Services as well as other modernization projects.

The IRS initially estimated in April 2000 that the e-Services project would require $3.8 million to complete its design phase by September 2000. As of March 2001, the estimated costs to complete this phase had increased to over $17 million, and the expected date of completion had been delayed until June 2001. Even if current development and deployment schedules are met, over $14 million of the $16 million in projected savings for FY 2002 will not be realized until FY 2003 or later.

Much of this increased cost and delay in project development is due to a significant change in project requirements. IRS executives were not comfortable with how e-Services integrated with other projects, with the Modernization Blueprint, and with the IRS’ overall modernization vision. Another cause for the delay was that the e-Services project was ahead of other key development efforts on which it was dependent. IRS executives decided to slow the development of the project until these other projects were further along, to reduce the risks.

We estimate that the costs for the IRS employees working on the e-Services project were understated by $1.6 million during the period October 1999 through January 2001. Project personnel indicated that it is difficult to ensure that employees, especially those who do not work full-time on the project, use the correct codes for charging their time. As a result, project costs appear to be lower than they actually are, and key measures such as return on investment cannot be accurately calculated.

The BSMO issues contracts (called task orders) to the CSC for specific products and services. Although the BSMO is improving the task order process, the task orders we reviewed did not include performance-based incentives for quality or timeliness. Also, the e-Services project team did not consistently negotiate task orders in a timely manner. Because the IRS identified various quality issues during its review of products received from the CSC, we believe that improvements in processing task orders could better protect the IRS’ interests as it works to ensure quality products at reasonable costs.

The project team uses earned value techniques to measure project progress by individual phase. Although earned value is an appropriate project tracking approach, measures that cover each individual phase are too limited to measure the overall progress of a project that covers five phases. Earned value techniques are designed to measure against a total budgeted amount, not just a current project phase. In addition, no validation of the project measures had been conducted at the time we completed our audit work. The data could be more useful and reliable if changes are made to the manner in which these measures are developed and a validation process is established.

The e-Services project team had developed a configuration management plan that followed the Enterprise Life Cycle (ELC) requirements. The plan outlines proper controls over project documentation and indicates that a project document repository would be established to ensure version control over documentation and system modules.

However, the CSC was not following the configuration management plan. The project document repository, where critical project documentation is maintained, was accessible to anyone on any of the project teams. This repository included both baselined documents and those that were still in process. There was no control over the various versions of key documents that had been accepted and signed by the IRS, and these accepted documents were not easily identified in this repository. Without this control, it will become harder to determine which documents are the official baselined documents. This could lead to project teams following the wrong set of requirements while developing the projects. During our audit, the BSMO and the CSC initiated corrective actions regarding configuration management processes.

The e-Services project team developed a risk management plan that describes the activities and processes it will follow to manage project risks. The project team was working to identify the risks and issues it faces during the project, using the correct forms to documents risks and meeting on a regular basis to discuss project risks.

However, we identified weaknesses in the implementation of the risk management process. For instance, the potential impact of identified risks on the project’s cost, schedule, or technical performance was not adequately documented. In addition, the project team did not adequately document the probable impact date, mitigation actions, measures to determine the effectiveness of mitigation actions, and the rationale for closing risks. These were contributing factors in the project delays and cost increases. The BSMO and the CSC have begun initiating corrective actions regarding risk management processes.

Managing project dependencies is critical to the success of the e-Services project and the overall modernization program. Identifying the risks associated with project dependencies is important so that all project teams and the BSMO can be aware of the criticality of any delays that may occur in completion of each project.

The e-Services project team did a good job of identifying critical dependencies on other projects; however, it was not adequately documenting the risks associated with the identified dependencies so that all project teams were aware of and agreed to the dependencies and necessary product delivery dates. Without clear and accurate documentation of project dependencies, incorrect decisions may be made at a program level and additional delays in the project can occur.

The e-Services project team was employing certain critical project management processes. The Project Manager was using a schedule, known as a Work Breakdown Structure, to manage the project team’s tasks. The schedule listed in detail the specific tasks that the project team needed to complete. Each task was identified with a specific identification number and had an assigned start date, finish date, and estimated duration. However, we found that near-term tasks were not assigned to specific team members. Identifying individuals assigned to tasks is important to ensure those individuals are available when their skills are needed. In addition, the schedule did not include any time to recover from unplanned or unforeseen events that could adversely affect the completion of critical tasks. Reserve or recovery time, while not required on every project, may be necessary if the current schedule is very aggressive or if a project team faces schedule risks, such as dependencies on other projects.

To minimize further delays and cost increases in this project, we recommend that the e-Services project team identify the critical factors the project is dependent upon and ensure BSMO executives are aware of each factor that could cause delays in future development and deployment. In addition, we recommend that the BSMO improve the tracking of IRS labor costs related to the project.

We did not provide recommendations for the other conditions identified in this report because we believe the corrective actions need to be taken at the program level rather than by the individual project teams. Because similar conditions were identified in other projects being audited, we plan to issue a separate report with recommendations that can be implemented at the program level.

Management's Response: Management's response was due on August 30, 2001. As of September 7, 2001, management had not responded to the draft report.