A new model for program management

Previous "CIO Perspective" columns have focused on the importance of IT program management to support successful mission and business delivery. Five key elements were defined that must be in place for programs to have a good probability for success, and if even one of those key elements is not properly addressed, the risk of failure increases dramatically.

One of the elements is having a set of skilled and experienced employees leading the program, while another is ensuring that a set of mature management processes is defined and executed to run the program. There must be an appropriate system development life cycle, which lays out the approaches to designing, developing, testing and deploying the system. But there must also be a robust set of project management disciplines, which include scheduling, estimation, requirements, configuration and risk management processes.

Based on our experiences working for the government and now in the private sector serving federal customers, agencies face a shortage of seasoned, skilled employees in all the program management disciplines. That shortage extends beyond program managers to other key positions, including requirements and system engineering leads, development and test managers, and operations managers.

As a result, government employees are often thrust into program management roles without formal training and without the experience to understand what to do and how to do it. We compound the issue by thinking training will solve everything. Though beneficial, training cannot make up for experience.

And too often, those in program management leadership roles have nowhere to turn for help. Consequently, they often rely on contractors for support, and although there are excellent firms that can provide expertise, contractor support does not instill the institutionalization of program management approaches necessary to improve agencies' capabilities over time.

The General Services Administration's 18F and the Office of Management and Budget's U.S. Digital Service focus on bringing new development methodologies to government to help drive innovation. Those initiatives are important, but they are only one facet -- and only a subset of one of the five key elements -- of delivering successful IT programs in government. We still need a skilled cadre of program managers who can acquire, deliver and manage the solutions.

There is, however, a model that has made a meaningful difference in agencies' ability to deliver successful IT programs, and it could scale to support the entire federal government. We advocate the establishment of Program Management Centers of Excellence (PM COE) that draw on best practices from the commercial sector, capitalize on pockets of excellence throughout government, and provide help for programs on initiation and throughout their life cycles.

We must recognize that running a successful, complex IT program requires numerous disciplines, such as requirements management, systems engineering and development management, to name a few. Individuals specialize in those areas, and we want to make sure we are bringing together the right skills and best practices for each of the disciplines in a coordinated manner -- hence the plural "centers of excellence" coordinated under the banner of Program Management Centers of Excellence.

The PM COE would have three primary functions:

1. It would gather best practices into a single resource center. Primarily, COEs identify and codify methods, processes, sample work products and tools that represent best practices in a particular discipline housed in an accessible, centralized library.

For example, a Requirements Management COE would help define the appropriate methods and tools to use in defining and then managing requirements for different types of IT programs -- such as application development and the integration of commercial software. The library would hold sample artifacts and reports to share good examples of what a functional requirements document consists of and the appropriate level of requirements definition for a series of agile development sprints, for example.

The resource center would also provide a basis to update training programs across the federal government.

2. It would create a network of experts. The PM COE would seek out and identify expertise in order to share it across the federal government. It would help develop communities of interest in which practitioners from across government convene (both virtually and physically) to discuss the state of the art for their discipline, review artifacts in the resource center, discuss current issues and collaborate on approaches to resolve those issues.

The goal is to develop an online learning community that crosses agency barriers while assisting program managers and program executive leaders.

3. It would offer independent consultants for program management development and support. COE members would be available to help programs in need, whether they are just starting out or long established but struggling to succeed. With a mature PM COE model in place, the leaders of all major new programs should be required to use the PM COE as a way to ensure that they are adequately addressing all critical aspects of running a successful program. Much as GSA's 18F leads digital service development, the PM COE would lead program management practices.

Too many programs fail right out of the starting blocks, and the PM COE could be instrumental by helping programs at the beginning. In addition, leaders could avail themselves of a COE if they are struggling with a particular aspect of program execution. And representatives of the PM COE could participate in TechStat sessions to help review programs.

There are a few keys to success for this approach to work at the scale of the federal government. First, the model must rely on the expertise that lives in agencies and support it with a small, centralized governance and coordination function. There are world-class capabilities in all these disciplines in government; the key is to find them and use them for the benefit of all.

Second, the PM COE culture would be about helping programs, not enforcing compliance. Senior leaders must come together as a "coalition of the willing" to foster a strong commitment to helping programs succeed so that program managers will be comfortable seeking the help of the COEs.

Finally, success depends on senior leaders prioritizing this movement in the same manner that they have with the other leading initiatives, such as OMB's USDS. By making program management a priority, senior leaders will free experts from across government to spend a small portion of their time to help create and sustain the COEs.

This is not impossible. At the Internal Revenue Service, we established 14 COE discipline areas under a unified governance structure that made a significant difference in the execution of the agency's modernization program. Other agencies have had similar results, and we can use those programs as part of a governmentwide initiative.

The good news is that such a model is inexpensive to establish, given the potential improvement to program delivery. A COE in a particular discipline could be effectively run with two full-time people if it is possible to tap experts across government on a part-time basis for support. Together with a small team to handle the central governance and management of the PM COE, a dedicated team of 30 to 40 individuals could provide a robust PM COE.

Given the billions of dollars wasted because of program management problems and failures, such an investment in staff with the right senior-level support could provide an excellent return for our government in terms of dollar savings on programs -- and, much more important, better mission and business outcomes from federal agencies.