Navy ERP illustrates the Anna Karenina principle
Although Navy ERP isn't an unadulterated success--Rand Corp. calls it is a "qualified" one--something in particular stands out in the Pentagon-commissioned assessment of the program: When program managers implemented the material management functionality, they did it in a staged approach--with each stage installing a mixture of hard and easy problems, rather than the harder problems being pushed into the future.
This actually is in contrast to other large federal information technology programs, such as the FBI's Sentinel case management system. At the bureau, program officials decided to go for the easy wins first and ended up submerging the program under its own, increasingly-heavy weight.
Navy ERP program managers also had the benefit of prototypes, something that appears to be universally agreed on as important but not always implemented. Even in the case of Navy ERP, the prototyping may have been more accident than intention, since Navy ERP was originally meant to be the convergence of four separate enterprise resource planning systems rather than the single ERP that program managers decided in 2003 to implement.
Nonetheless, it appears that the Navy hewed to important best practices, a useful reminder that dysfunction need not be the status quo of large federal IT projects. Best practices illustrate a simple principle: Happy IT projects are all alike; every unhappy IT project is unhappy in its own way. - Dave