Are procurement contests entering the IT mainstream?
Steve Kelman is hopeful that the VA's unusual choice to host a competition for an IT system will break new ground. (Stock image)
When I posted on my Facebook page a column about contests (when they’re likely to work and when they’re a bad idea) as a procurement tool, I received a comment from my friend Roger Baker, the dynamic CIO of Veterans Affairs, about a contest the VA is running where they will pay $3 million to the winner who can develop an open-source appointment scheduling system for VA hospitals.
The contest was announced in early January, and entries are due by June 13. (For blog readers unfamiliar with the idea, a procurement contest is one where the government puts out a problem it needs to solve and announces a prize for the first or the best solution to the problem. Anyone may enter – no RFP, no long proposals, etc. Experience shows that winners are typically players the government has never dealt with before.)
The VA had earlier spent over $125 million to try to develop a customized scheduling system, and it didn’t work. This was one of the systems cancelled in a bold move by Baker when he came on board as VA CIO in 2009. I’m sure I’m missing something, but scheduling sounds to me like a fairly widespread commercial business process, and it seems weird that the VA had originally thought to grow their own. If the VA is actually now able to get an application that works for $3 million, after all they’ve spent, it would be an example of the 95 percent cost savings that come up often enough to make the contest phenomenon worthy of further attention.
At any rate, we should be really paying attention to whether the contest idea works here. It has seldom (in fact, I’m trying to be cautious – I actually can’t think of a single example) been used in government to date to try to source an IT application. In the column I wrote (based on a paper by a University of Michigan business school academic), it was argued that contests are less likely to work if the product or service cannot be fully specified and if there will need to be a lot of communication between the buyer and potential contest entrants. Since IT applications typically can’t be fully specified and do require communication, VA’s approach is risky – though hardly more risky than the original decision that led to blowing $125 million on failure going the traditional route.
So this is a project that the IT community should be watching really carefully. One way or another, there are likely to be lessons learned here, which hopefully the VA will share. If this works, it’s going to open a new chapter for IT procurement in the federal government. Hats off to Roger Baker and Peter Levin, the VA’s CTO, for giving it a try!
Posted on Jan 24, 2013 at 12:09 PM