For Oracle, Philadelphia is not the city of brotherly love these days. The fifth largest municipality in the U.S. has just killed a project to implement Oracle applications in the city's water department. Furthermore, Oracle has agreed to forgive or pay back a whopping $6.9 million in fees.
The Oracle project was put on hold in October 2005, after the city had spent $18 million with nothing to show for its time and effort. In its place, the city has launched a new project to implement Basis2, a utility billing system from Prophecy International Pty., an Oracle business partner in Australia. The new system will run on an Oracle database and will interface with Oracle E-Business Suite products that the city runs for financial applications. Oracle itself will play no role in the new project.What went wrong?
There are several hints in a Computerworld story
on the situation, concerning the root causes of failure in Philadelphia.
First, it appears that the city did not appreciate the magnitude and difficulty of managing a project of this size, at first putting it entirely under the direction of the users. In light of the fact that their IT experience was limited to a legacy application using punched cards
(!), one assumes they were ill-equipped to manage an Oracle implementation.
In contrast, the new project is being jointly managed by the users and the city's IT group. The city's new CIO, Terry Phillis says, "I know it's second-guessing, but the city suffered somewhat by not maturing the organization to care for a project like this." Under the joint-responsibility arrangement for the Basis2 project, he said "We're getting along terrifically. We're singing 'Kumbaya' -- not well, but we're singing it."
Second, the original project included quite a bit of custom software development. Custom development greatly increases the risk in an implementation, especially a large project such as this one, magnifying any deficiencies in project management. In contrast, implementation of Basis2 will require no customization. One wonders why the city didn't try harder in the beginning to find a system that better fit its requirements.
If these are the root problems, then Oracle shouldn't be blamed entirely for the failure, although a give-back of nearly $7 million suggests that Oracle made its own mistakes on the engagement. One suspects that the city had some strong terms and conditions in place for Oracle's performance. If so, at least the city did one thing right.Related postsProject management: the missing discipline