In 2005, the Air Force attempted to replace aging systems with new hardware and new ERP software from Oracle. The Air Force signed a $627 million contract with Computer Sciences Corporation to integrate the system, but 7 years and $1 billion later, it realized that it would cost another $1 billion just to implement only 25% of the planned capabilities. The Air Force’s main mistakes were failing to define its business requirements while trying to upgrade its disparate systems.
The new project, whose costs the Air Force declined to reveal, calls for a new software system that offers a complete view of the Air Force’s business operations, including its supply chain, sales and operations, maintenance repair and engineering data. But to meet the 2017 deadline, the team must comb through thousands of deficiencies across 19 legacy systems. For example, the current system does not allow the deletion of vehicle ID numbers, which means the Air Force cannot properly track and take inventory of existing vehicles.
“We think we can break this problem up into smaller subsets, field those subsets on a quicker cycle and at a lower cost, and with less risk,” says the general in charge. But the CEO of one IT consulting firm said that while he agreed that the Air Force’s approach of breaking up the project into smaller subsets is correct he wondered why this wasn’t done before the Air Force spent so much money on the previous project. “The most basic tenet of ERP implementation is know why you need the system and how you intend to make it work.”
Discussion questions:
1. Why are ERP (and other large programs) so difficult to implement?
2. What mistakes did the Air Force make in this project?