Bringing Pilots to the Federal Government: An Interview with Clay Johnson
As highlighted last week with our conversation with Stan Soloway, we had some great interviews with very smart people about increasing competition and eliminating barriers to innovation (that paper is available for free download, by the way). One of the most outspoken critics of the way tech procurement is done at the federal level has been Clay Johnson, Founder of Blue State Digital and Department of Better Technology, former director of Sunlight Labs for the Sunlight Foundation, and now a fellow for the Center for American Progress.
Johnson’s Blue State Digital became somewhat legendary for delivering a new kind of digital engagement to Barack Obama’s 2008 presidential run, and it’s there that Johnson first saw the sort of model that government needed:
“The idea was to build software for what the private sector pays, to teach clients how to use that tech, and to set up educational organizations to improve skills of the customer. [The Democratic Party] wanted to have ‘savvy clients.'”
On how he became interested in procurement reform:
“Sunlight was dependent on the data coming out of government, and it was during that time that I got concerned about how the market works. Federal IT, the shape of federal IT, eventually led me to give up on open government and transparency initiative. I realized that until we fix procurement policy, open government initiatives just mean large windfalls to contractors. Once government invests in an open government program, they tend to think the job’s done. It takes longer to fix a broken implementation than it does to start a new implementation.”
Clay was one of the first presidential fellows, and worked on RFP-EZ.
“I quickly learned that being in charge of a procurement innovation is like being thrown into the deep blue ocean with weight on your legs. But we were able to ship a product in four months—RFP-EZ cost less than a quarter-million dollars, and it saved 30% across projects that ran through it. From my experience, both at Blue State Digital and inside the government, is that the state of the Democratic Party’s tech six, seven years ago was very similar to the state of the government today: You have a bunch of government contractors building large-scale IT projects that consistently fail, and people inside the customer base that don’t have the expertise to ensure they don’t fail.”
We discussed Recovery.gov, which the Sunlight Foundation bid on when Johnson was there. Sunlight thought it could build it for $1.4 million, but because the site was part of the Alliant GWAC, Sunlight was ineligible. The site ended up costing $18 million.
“It certainly doesn’t look like an $18 million website. It’s touted as way to combat waste, fraud and abuse, but not a lot of people are finding mustache-twirlers by searching on there. Open government relies on technology, and so you’re not going to build a house on a foundation that’s broken.…Fixing Recovery.gov would mean changing the scope of the project. The RFP had unnecessary things in it, like datacubing, firewalls, etc., things government didn’t need to build it, but no one was equipped with the knowledge of how modern software systems are built, or they would have asked for much simpler software.”
This led to a discussion of pilot programs, which Johnson has championed in the federal government.
“I’d say the first thing we’re going to do is three $250k pilot projects for Recovery.gov. Keep it small and restrained, three teams of six people or less to implement. Over the course of one month, we’d see which is most successful out of those pilots, and attach clear metrics as to how those pilots work. After the pilots are completed, we pick the one that works the most, allow it to grow. It’s a more agile form of contracting than giving a 4-year contract and hoping the contractor does the job right. You’re less dependent upon your initial assumptions. “
On opening up contracting to more competition:
“From a federal perspective, the number one problem is SAM.gov. It’s catastrophic. In tech, it’s the new businesses that have new products entering the market. That’s where the innovation is. Chances are that technology is not being sold first to the government, it’s being sold last to government after consumers have been using it for years, and businesses are starting to use it. If I sell to the government, if a contracting officer says register on SAM.gov, the value proposition is actually relatively low. Generating billable hours is likely far off. For a lot of people, the front door is immediately closed.
SAM.gov went from being a 14-step from to an 80-step process, government asking for information it doesn’t need up front. The federal government doesn’t need my checking account routing number before it pays me.”
On problems with small business registration :
“The small business registration issues are problematic. My wife tried to self-certify as a woman-owned business and it took her six months. Why does it take six months to prove she’s a woman?: The 8(a) program and the woman-owned, veteran-owned, should be run well, allow contractors in as quickly and cleanly as possible.
Johnson has been a strong advocate for a leaner process:
“Ask yourself why are they competing contracts that way. There’s a reason they have IDIQ contracts: running public competitions is a lot of work. Reading a lot of proposals takes a whole lot of time. Exposes them to a lot of liability, frankly it’s very easy and costs you relatively little money to protest an award. One of the success stories from RFP-EZ is that the RFP response was limited in the amount of info a vendor could provide. Businesses don’t like spending a lot of time writing proposals, because it’s not billable work. Contracting officers don’t like reviewing proposals, because now businesses have confused proposal length with probability of success.”
While we were looking for ways to improve the success rate of large IT programs, Johnson said there is no such thing.
“Large IT programs are not successful often enough. Really what we should be thinking about is how do we make IT programs smaller and lower risk. If there’s a 94% chance of failure, why continue to invest in that methodology? We should have small investments at the very beginning, small pilots and viable products.”
One way to fix things would be to bring in experts to help create those “savvy clients” within the federal government.
“I think that if you basically gave every contracting officer—for every IT project with a cost estimate greater than $1 million—you gave them a $1,000 expert budget. You say, ‘Before you write your RFP, here’s your budget, pay up to five experts from community xyz to help advise you.’ The experts are going to be associated with the RFP for accountability reasons. You might be getting somewhere. My take to the White House on procurement reform was: Let me review every RFP, I can tell you what you don’t need. I could find a sequester’s worth of money in an RFP. But they don’t have any budget to promote RFPs or get experts.”
For more on IT procurement, read our paper here.