Enterprise Resource Planning: NASA-Style

The breadth and depth of requirements for enterprise resource planning at federal agencies and the way the federal government does its accounting are two good reasons why examples of ERP do not jump out in the sector.

By John Makulowich, Senior WriterThe breadth and depth of requirements for enterprise resource planning at federal agencies and the way the federal government does its accounting are two good reasons why examples of ERP do not jump out in the sector.For starters, there are very few agencies putting in place a complete ERP system, one which, in the words of IBM Corp., "automates and integrates business processes and provides the data stream for business analysis."When you do find one, for example, at the National Aeronautics and Space Administration, it is likely to be under a different name. In this case, it is IFM for integrated financial management. New York-based KPMG is the vendor tasked with putting the system in place through a contract worth $186 million won in 1997.The space agency's effort is the most complex ERP system in the entire federal government, according to Lee Holcomb, NASA's chief information officer. "If you went out today, looking at what we would like to do with IFM, you could not find a single vendor that could fulfill the full requirements of the federal sector," said Holcomb.It is not just the requirements in the federal sector that are at issue but the difference in the way the federal government does its accounting vs. the way it is done in the private or not-for-profit sectors, said industry and government officials. (See related ERP stories in a special report that begins on page 20.)The worldwide ERP market, which includes software for accounting, human resources, payroll, materials management and manufacturing applications, generated $14 billion in sales in 1997, according to International Data Corp., Framingham, Mass. The market was projected to grow by more than 19 percent in 1998. SAP is the market leader with a 15.6 percent share, followed by PeopleSoft (4.9 percent) and Oracle Corp. (4.8 percent) in 1997, according to IDC.The principal goal of the NASA IFM program seems simple enough: improve financial management processes throughout the agency. But putting it into practice is a whole other story.The IFM system includes the standard modules: accounting, budgeting, procurement, time and attendance, travel and an executive information system. That is phase one of the KPMG contract. In phase two, scheduled for completion by 2002, there will be integrated modules for human resource and payroll as well as one for asset management, covering the array of assets ranging from aircraft through facilities to computers at 11 NASA centers.Bob Glickert, the KPMG project manager for the IFM project, declined to talk about his company's efforts, saying, "That would require me to get about 15 approvals," he said.It was NASA's Chief Financial Officer Arnold Holz who started the Integrated Financial Management Project in March 1995. The purpose was to implement common agencywide solutions for many NASA business and administrative processes.As Holcomb pointed out, though, the popularity ERP enjoys in the private sector comes from its early deployment and successes in the manufacturing sector, for example, among chemical processing and automotive firms in delivering products and managing the product stream. In the federal sector, the focus is on services, not manufacturing."ERP is not unknown in the federal government, that is, the ability to integrate stovepipe or separate systems to optimize the overall performance of the organization," he said. In fact, he said, quite a few agencies are interested in this area."The difficulty is the number of legal ramifications in setting up such a system," Holcomb said. For example, not only must an agency interested in developing an ERP solution adhere to the federal budget structure, but it must comply with the laws on accounting in the federal sector.What's more, the agencies track things differently."We track very different things, like obligations and authority," Holcomb said. "In the private sector, you track asset values. We are also different in the use of time and attendance cards. There are laws associated with how the federal sector must acquire and legally track individual time and attendance ... As a result, there are very few companies that offer products certified to federal sector requirements."There is also the rapidity with which agencies must "rack and stack" their budgets, Holcomb said. Many federal entities must develop multiple iterations of a budget on different levels, something that can take up to 18 months to complete.While NASA's ERP system is far from complete — KPMG is still trying to make the phase one integrated system work at NASA headquarters in Washington — Holcomb has learned some lessons he willingly shares about the process."First and foremost," he said, "you must re-engineer the processes you want to automate. This is the most difficult and most important task. We had at least four different budget processes that were largely culturally driven and 11 different accounting systems, one in each center. We have done a reasonably good job for common processes in accounting, budgeting, time and attendance and travel."Also, understand your requirements well, and be willing to excise activities that are nonessential, Holcomb said. Next, understand the product you are getting, that is, what it can and cannot do. "We went into the program with the assumption that the product was further along than it really was," he said."Another lesson is that when you take a series of [commercial off-the-shelf] products and glue them together, you need to know them very well," he said. "The idea of plug and play does not work with one, two, three million lines of code; integration is difficult." Finally, be aware of interoperability issues, he said. NASA is itself a heterogeneous environment, with MacIntosh, PC and Unix applications interoperating on multiple platforms.Holcomb also cautioned against being too impatient with the process. "A recent study noted that the Industrial Revolution took about 70 years. Using that yardstick, the Information Revolution is at about year 45," he said.

Lee Holcomb