Enterprise architects must prove their worth in reducing duplication

Enterprise architecture is vital to reducing duplication across government, but architects need to be able to measure and report on the benefits of their programs, a panel of federal officials told an audience of enterprise architects in Washington.

The General Accounting Office’s Enterprise Architecture Management Maturity Framework version 2.0 can help them do that, GAO officials said.

People know there is great potential for reducing duplication and overlap, but not everyone outside of the enterprise architecture community is convinced about what architectures have achieved, said Michael Holland, senior analyst with the Government Accountability Office. Holland spoke during a panel discussion on EA’s role in reducing duplication at 1105 Media’s Enterprise Architecture conference in Vienna, Va., Nov. 7. 1105 Media is the parent company of Government Computer News.


Related coverage:

Wanted:  Enterprise architects for data center consolidation

GAO updates framework to reflect enterprise architecture's evolution


In 2006, GAO surveyed agency managers about the benefits they expected to achieve from EA programs, such as improved alignment between IT and business, improved information sharing, better change management and infrastructure consolidation, Holland said. One common theme that ran through the responses was the ability to address duplication and overlap within agencies.

GAO agrees there are many benefits associated with EA, Holland said. Valerie Melvin, director of information management and technology resources issues at GAO, gave an overview of the report, “Opportunities to Reduce Potential Duplication in Government Programs, Save Tax Dollars, and Enhance Revenue,” released in March 2011.

In the report, EA was the only tool or methodology mentioned as an approach to reducing duplication. But the report also noted that agencies across the board are not measuring and reporting the benefits of their architecture programs. That might contribute to the skepticism about EA, Holland noted.

Every EA practitioner attending the conference knows what they are trying to do with their programs, Holland said. “The real question is, “Have you achieved those goals?” 

GAO’s EA Management Maturity Framework, Version 2.0, released last year, included a number of different practices to get at the question of what programs have achieved. The EAMMF is a seven-stage assessment framework that helps gauge the state of an architecture program and develop improvement plans.

There are elements in the framework aimed at defining and measuring EA program expectations. For example, stage one ensures that the purpose of the architecture is very clearly stated.

“You can do all sorts of different things, develop different models,” Holland said. So, “whatever direction you head in needs to be bounded by the purpose of the program”. Plus, there has to be general acceptance throughout the organization about the purpose of the program.

Also included in the early stages of the EAMMF is a performance and accountability framework so people know who are the key stakeholders, their roles and responsibility and how they will be held accountable for the results they are suppose to help the program achieve.

The framework also addresses budget concerns. EA practitioners should be able to justify what their program is going to achieve, to make it clear from the beginning what are the expected benefits so they can be measured against those benefits.

In some cases, they can set up a self-funding architecture program where, if given a certain amount of money, they can show how much can be saved. “That will help ensure the program gets the appropriate level of resources,” Holland said.

Architects have to ensure EA results and outcomes are reported and measured, Holland advised. “Make sure when measuring the results they are tied directly to the purpose of the program.” Those results might include identifying and addressing duplication and overlap across the agency, he noted.

“Finally, take a step back and look at how you are measuring results,” Holland said. It is possible enterprise architects might not get the result they expected because they have the wrong metrics or the metrics are measured wrong. 

Meanwhile, the Homeland Security Department’s enterprise architecture group is closely working with other partners such as the strategic planning and budget offices and governance office to reduce duplication and overlapping programs within the department, said Chris Chilbert, DHS’ chief enterprise architect.

The Defense Department’s move toward a business enterprise architecture based on semantic technology will help DOD effectively address the duplication of systems, said Dennis Wisnosky, chief technical officer with DOD’s Office of the Deputy Chief Management Officer.

Semantic technology provides a structure that lets information be described in a way that captures what it is, what it means and what it's related to in a machine-readable format.

“DOD has about 5,300 registered systems,” Wisnosky said. “We are the biggest, and so have failed in the biggest way in eliminating duplication.” As a result, DOD has stopped looking at systems, picking winners and losers. Instead, the focus is on looking inside the systems to see what they do.

Take for example, the Defense Travel System, which has 21 services, only seven of which are travel-related. Why should the travel system have its own personnel or payment system? Why should each system have its own identity management system? When architecture is designed there needs to be a single enterprise identity management and access control system that all services can use.

“It is the only way [we’re] going to make this small enough to be managed,” Wisnosky said. “It all begins with understanding the data; that is what the semantic technology push is all about.”

Reader Comments

Tue, Nov 8, 2011 Old Hand

For more than 20 years, the GAO has been pushing the failed idea that creating "enterprise architectures" must be done before the government can get its IT house in order. Now, at this conference, GAO speakers have the nerve to tell government managers with a straight face that the reason all their efforts to build and use EAs have never produced anything of value is because they're not "measuring the benefits of EA" properly? Give me a break. At some point, senior executives need to stand up to the GAO on this point and start working on getting the Congress to understand that "enterprise architectures" (which the GAO succeeded in getting the Congress to legislate requirements for in the misguidedClinger-Cohen Act of 1996) are never going to work. Instead, we get appeasing charlatans like Dennis Wisnosky at DOD, who is hoping people won't notice the self-evident contradictions involved (notwithstanding the magic powers of "semantic technology") when he calls for the creation of single enterprise-wide systems as the only way to make things "small enough to be managed."

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above