For decades the US Air Force tracked its aircraft, fuel and spare parts around the world on hundreds of separate computer programs that didn’t talk to each other. Some of them, the Air Force said, needed “life support.”
In 2004 it started the Expeditionary Combat Support System, or ECSS, to replace them with one system. The plan was to buy, not build. Commercial software already reflected proven ways of running a supply chain, and buying it, the Air Force thought, would reduce the need for custom work. Its 2005 solicitation asked for software that was “truly ‘off the shelf’: unmodified and available to anyone.” Oracle supplied the software, and Computer Sciences Corporation (CSC) was hired to fit it into the Air Force. A 2006 Air Force strategy said the program should forgo any modifications to the commercial software.
The trouble started with what “off the shelf” meant. Oracle delivered three separate products and the tools to join them. The Air Force’s request to integrators, according to Carnegie Mellon’s Software Engineering Institute, described them as a single product that was already integrated. CSC said it had assumed it would receive one. Sorting out who would make the pieces work together took until 2009.
Then came the harder part. The commercial software assumed certain ways of ordering parts and managing maintenance. When CSC proposed changing Air Force processes to match, according to the company, the Air Force resisted and asked CSC to alter the software to fit the Air Force’s existing practices instead. Users, the Senate’s investigators found, were comfortable with the old systems and had not been told clearly why the new one mattered. Training was poor. In eight years the program had six program managers and five program executive officers.
In 2012 the Secretary of Defense cancelled ECSS. The Air Force had spent more than $1 billion without fielding any usable capability. Finishing, the investigators found, would have cost another $1 billion for about a quarter of the capability originally sought. Air Force staff went back to the old systems, and the Air Force still couldn’t say how many of them ECSS would have retired.
In its 2014 report, the Senate Permanent Subcommittee on Investigations called the program “a waste of $1.1 billion in taxpayer money, a loss of eight years of effort,” and “one of the most egregious examples of mismanagement in recent memory” at the Department of Defense. Point by point, it set the 2006 strategy against what happened. The plan was to forgo changes to the commercial software, test thoroughly, run one clear governance structure and follow change-management guidelines. By the end, the software had been customized, the integration hadn’t been properly tested, the Defense Department had directed the program to follow more than one governance model, and the change-management guidelines hadn’t been followed.