Monday, January 21, 2013

Another Waterfall Disaster: Steve Denning for Forbes

Reconciling Innovation With Control: The Air Force's $1.3 Billion Lesson In Agile
What are we to make of the news that the Air Force recently canceled a six-year-old software modernization effort that had consumed $1.3 billion and produced nothing of value? Note, that’s $1.3 billion, not $1.3 million. And it’s not that the project produced less benefit than expected. It produced absolutely no benefits at all. The whole project has been canned.
The fiasco is described in the New York Times in an article by Randall Stross which notes that the Air Force’s effort began the project in 2006. It was “supposed to manage logistics using software from Oracle [ORCL].”
The Air Force awarded the $628 million contract to the Computer SciencesCorporation [CSC] to serve as lead system integrator; its job would be to “configure, deploy and conduct training and change management activities” before the launch.
The project had been “restructured” a number of times. When the Air Force realized that it would cost still another $1 billion just to achieve one-quarter of the capabilities originally planned that wouldn’t meet even minimal requirements—and that even then the system would not be ready before 2020—it gave up on the project entirely.


mantaspac said...


reading the original sources, it seems it has nothing to do with the so much hated waterfall:

"the main one [reason] was a failure to meet a basic requirement for successful implementation: having “a single accountable leader” who “has the authority and willingness to exercise the authority to enforce all necessary changes to the business required for successful fielding of the software.”"

I totally believe in Jeff Sutherland and I don't any proof to his claims. But some other readers may need some proof anyway.

Could Mr Scrum point me a link to any sw methodology that recommends no accountable authority point at all?

Jeff Sutherland said...

Well it certainly is true that Standish Group has identified the leading cause of failure of projects is not having a senior person involved and accountable. The success rate of waterfall is 16% and agile is 46% over the last decade so clearly agile projects are run badly. Most of the failures are probably due to bad or AWOL product owners. I think the benefit of agile (which is now the law for DOD) is that you usually find out something is not working before you spend a billion dollars.

The thing that is upsetting about these black hole military projects is that you could probably feed every homeless person in the U.S. at the restaurant of their choice for a year or more for the cost of a single failure.

Or consider applying the money to micro-enterprise lending in the developing world.

mantaspac said...

" I think the benefit of agile (which is now the law for DOD) is that you usually find out something is not working before you spend a billion dollars."

"Incremental" != "Agile" != "Scrum"

DoD has just embraced the best practice of 1970's. It's not yet to be called Agile (and definitely not Scrum). It's old good incremental development.

Steve said...

One of the leading credos of the agile movement is "fail early, fail often." But when agile methodologies are used without thought to this it's easy to just keep ploughing on, and the result is millions of wasted dollars.

The problem is, nobody gets paid for saving the government money, so it's much easier for large contractors to keep the money flowing in by not canceling projects.

Nick Zdunic said...

Shameful waste of money. We really need to be doing better than this. Vendors also need to assist in this matter. Doing agile contracts is a step away from adversary. Committed POs are a key - and maybe avoidance of projects without one is key as well. Still some devil in the detail around Agile but lets not have that stop us. Progressive improvement and we can get there, but being stuck in Waterfall will not help, but maybe Kanban would assist (food for thought)