“We started with a Big Bang approach and put every possible requirement into the program, which made it very large and very complex,” says Elizabeth McGrath, the department’s deputy chief management officer.
This stuff is hard, even with commerical off the shelf (COTS) applications -- they still have to be configured; you still have to move your data.
Grover Dunn, the Air Force director of transformation, [now says he] ... can see just how unrealistic the project was: “We’ve never tried to change all the processes, tools and languages of all 250,000 people in our business at once, and that’s essentially what we’re about to do.”
After a lot of money and time, the Air Force has pulled the plug on their IT modernization.
So, what do we have here?
- A huge user community that has both central and local needs/wants/urgencies
- A big scope to change everything all at once
- Off-the-shelf with thousands of customizations and configuration details
- Probably a lot of legacy data (100's million of records, I imagine) to clean and move
- And, big bang
With COTS, it's hard to think of conventional agile methods to avoid the big bang, but you certainly can think of ways to do this with rolling waves, partial installs, etc
I've heard all the arguments about not paying for throw-away interfaces when you do partial installs, and I've heard all the arguments about multiple training rollouts and how they affect the organization, but experience shows they are cheaper in the long run.
Let's just hope the next time around is more sane.
