Air Force Brig. Gen. Kathryn Johnson attributed the Expeditionary Combat Support System project's extensive problems to an array of factors. (Thomas Brown / Staff)
The Air Force’s controller, Jamie Morin, admitted publicly in April that the service spent seven years and $1 billion on a logistics management system that had “negligible” capability.
But that’s not what drove Air Force leaders to finally cancel the Expeditionary Combat Support System project this month.
And it wasn’t because technical glitches forced the Air Force to repeatedly scale back expectations for the new system — from replacing 240 legacy systems, as originally planned, to just 12.
In the end, officials canceled ECSS because continuing it would have cost another $1 billion to gain a quarter of the capability it was originally supposed to have, with fielding delayed until 2020.
“It just became a financial decision,” said Brig. Gen. Kathryn Johnson, the Air Force’s director of system integration, in an interview last week.
Johnson attributed the project’s extensive problems to an array of factors.
For one thing, the Air Force handed over control over most aspects of the project — including the task of writing some of the system’s requirements — to the prime contractor, Computer Sciences Corp. of Falls Church, Va.
CSC was simply not up to the task of adapting commercial Oracle software to fulfill the many specialized tasks the system was supposed to perform, said Johnson and Robert Shofner, the Air Force’s program executive officer for business and enterprise systems.
The project became plagued with technical glitches and delays.
“It was a lot of problems,” Johnson said in the Air Force’s first public discussion of what caused the Air Force to cancel ECSS.
The system had once been touted as revolutionizing the management of parts and equipment. The ECSS’ success was a critical component in the Defense Department’s strategy for meeting a congressional deadline for having audit-worthy books by 2017.
After firing CSC in March, the Air Force was unable to keep the massive automation project alive.
Johnson, who joined the program in May, listed five major causes behind the ECSS failure, including the Air Force’s lack of a master schedule, a change in acquisition strategy and infrastructure problems that slowed the speed at which the system could share data with Air Force installations.
But she laid blame for the other two causes at the feet of CSC and software supplier Oracle. That software “may not have been as developed as we would have originally liked it to be,” Johnson said. CSC, despite running into some 2,000 design flaws in meshing Oracle’s software with the Air Force’s logistics demands, went ahead with testing anyway.
Under the military’s acquisition strategy at the time, CSC was charged with not just implementing the ECSS, but actually developing some of the requirements for how it should operate. When the contract was awarded in 2006, those requirements were “a work in progress,” Shofner said.
“Frankly, [CSC] just didn’t have the skills to go and do it,” Shofner said.
In an emailed response, CSC spokeswoman Heather Williams said only that the company “values the Air Force and its mission, and worked closely and cooperatively with its senior leadership to close out the ECSS contract.”
Efforts to reach Oracle representatives were unsuccessful.
CSC was paid about $527 million for its work on the project, including $8.2 million in contract termination fees, according to the Air Force. The price tag for the Oracle software was about $81 million.
Asked if anyone has been fired or demoted for the program’s failure, Johnson said no. “We didn’t feel it was necessary to do that,“ she added.
The ECSS was among almost a dozen “enterprise resource planning” systems undertaken by the Defense Department and the military services to modernize management of logistics, finances and other business operations. The systems are supposed to replace numerous smaller-scale legacy systems that are decades old in some cases.
But while the ECSS is the only one that has been canceled so far, a half-dozen other enterprise systems are years behind schedule and a combined $8 billion over their original budgets, the Defense Department’s inspector general said in a July report. The delays not only undercut anticipated cost savings, but also risk putting the department behind in reaching the 2017 goal to clean up its books, the IG said.
At a Senate hearing in April, Air Force controller Morin said he was “personally appalled” at how little the ECSS had delivered in terms of performance.
Lawmakers have already asked the Air Force for a briefing on the ECSS cancellation, according to an aide to Sen. James Inhofe, R-Okla., who is in line to become the top Republican on the Senate Armed Services Committee next year.
“It appears to us that no matter how much money they threw at it, it was not going to achieve what it was supposed to achieve,” said the Senate aide, speaking on condition of anonymity. “The question is, why did it take this long to figure that out?”
The program got underway in 2005 with the promise of transforming the Air Force’s aging logistics systems. When CSC won a $628 million task order the next year, the service said in a news release that the company would provide the leadership needed to “dynamically meet war fighter needs.”
But the project was soon slowed by objections from rival bidders. In a 2005 protest filed with the Government Accountability Office, software maker SAP complained unsuccessfully that Oracle’s package did not meet bid requirements for “commercial off-the-shelf software.” The next year, however, GAO ruled in the Air Force’s favor and also sided with the government on a separate challenge by IBM to the CSC contract.
As a result, work did not effectively begin until 2007, Shofner said. When the Air Force announced two years later that fielding would begin at Hanscom Air Force Base in Massachusetts, a news release said the system would “revolutionize” operations by creating a single source of logistics information accessible from any base. But test results from the next year reflected problems with hand-held scanners, data quality and almost a half-dozen other areas, according to a summary by the Air Force Operational Test and Evaluation Center.
From there, the program was hobbled by numerous delays that led to higher costs, prompting the Air Force to repeatedly restructure and scale back expectations for what the program would accomplish.
Asked if the Air Force should have pulled the plug sooner, Shofner, who also joined the program recently, called that “speculative.”
The Air Force is now seeking another approach to meeting the 2017 audibility target using the legacy systems it once sought to replace. Managers also aim to continue logistics modernization efforts, but in smaller pieces, Johnson said.
In hindsight, a critical reason for the program’s failure appears to be the decision to allow CSC a role in setting the requirements as the “lead system integrator.” That approach, which proved disastrous in acquisition programs at other agencies, has since been discarded. The Air Force decided to terminate CSC’s contract because it wanted to take back the integrator’s role, Shofner and Johnson said.
“It was what we thought was a viable strategy at the time, and clearly it was not,” Johnson said.