GovTech is Not Broken
It’s time for us to stop pretending that the problems with government technology procurement are an accident.
This post originally appeared on civic.io.
When we talk about the challenges that face governments in acquiring and implementing new technology, the conversation eventually winds around to the procurement process.
That’s when things usually get ugly. “It’s broken,” they say. “It just doesn’t work.”
What most people who care about this issue fail to recognize, however, is that while the procurement process for technology may not work well for governments or prospective vendors (particularly smaller, younger companies), it is not broken.
It works exactly as it was designed to work.
Squeezing out the little guys
If you’re looking for a detailed account of what the government procurement process looks like from the perspective of a small startup, there is an interesting article in VentureBeat that gives the account of TallyVoting, a voting technology startup that is closing down after trying to penetrate the government market.
Here are some specific criticisms raised by the team behind TallyVoting.
The process moves too slowly:
Governments move slowly compared to regular industry. Compared to the tech sector, government moves at a glacial pace. Selling stuff—not just technology—to government requires sourcing leads, submitting proposals and counter-proposals, securing external certification, periods for official and public comments, and that’s just the tip of the iceberg of process.
The process is too costly and complex, particularly for smaller firms:
That long sales cycle doesn’t give a small bootstrapped startup a chance to demonstrate the potential of its concept and technology. (Unsurprisingly, the already-giant vendors with a stronghold on the market are more than able to handle this long cycle. They’re also more interested in profits than improvements and innovation.) It also doesn’t give the little startups the opportunity to show traction and product-market fit.
These are fair criticisms, but in discussing them we often make the mistake of assuming that they are unintended problems that arise through some general disfunction of government.
These issues are not accidental, and it’s time we reframe the discussion around procurement reform accordingly.
The procurement process is too costly and complex
The most under-appreciated characteristic of the government procurement process as it exists today is that it’s current design is largely intentional. Much like the federal and state income tax systems, we imbue a number of values deemed important into our procurement processes in the hopes of fostering desired outcomes.
Requirements for women and minority-owned business participation; requirements that governments utilize local vendors; requirements that governments favor vendors that have any number of different traits or qualities—no one would argue that these requirements are adopted in the hopes of fostering positive outcomes. But the price of using the procurement system as the vehicle for achieving these ends means that the process is more complex for all participants.
This is true, even if the positive outcomes are not realized, or realized to a lesser extent than hoped for. Much like with our tax system, the added complexity borne by all participants doesn’t guarantee the intended results.
Probably the most overarching value we imbue in the procurement system is risk aversion. Much of the complexity and cost of the current system (for both governments and vendors) can be attributed to the desire to reduce the risk assumed by governments when partnering with outside firms to build or acquire technology.
Requirements like proposal and performance bonds, professional liability insurance coverage and the submission of audited financial statements (sometimes several years’ worth), to name just a few, are all meant to offset the risk governments assume when they rely on a vendor.
In some respects, these requirements can be deemed beneficial to governments and indicate prudent stewardship of public resources. However, the widespread use of these and similar requirements by governments at all levels doesn’t seem to have lessened the incidence of IT project failures, and they can drive up the cost of technology projects by pricing smaller firms out of the bidding process.
I’ve heard representatives of smaller firms argue that the bond industry doesn’t have a good understanding of how to underwrite technology projects, so it can be difficult to get a performance bond as part of a bid response. The result? Many smaller firms that are well qualified to do the work may choose to simply walk away from the process. This is true even if these bond provisions don’t lower the actual risk assumed by governments as part of larger technology projects.
Again, the added complexity and cost borne by all participants can not guarantee the intended results.
A better much way for governments to hedge against the risks inherent in any large project—particularly technology projects—is to develop the internal capacity to manage and implement them successfully. Fortunately, we are starting to see this happen at the federal level with the creation of the U.S. Digital Service.
The procurement process moves too slowly
While this seems particularly acute in the procurement process, this is also a fair criticism of government in general.
We can make the same observation about the pace at which governments operate that we have about the procurement process; much of this is by design. But long procurement and budget cycles can drive away smaller firms that can not afford to invest significant resources in projects that don’t provide a return for months on end.
There are usually requirements for government RFPs to be posted publicly and widely advertised, and rules that ensure all interested parties have an equal opportunity to participate through public bid notices and open vendor meetings. All of these requirements are meant to enhance the transparency of the bidding process but they can also add to the time it takes to select a vendor for a project, and for the vendor to actually begin work.
An emphasis on transparency in the procurement process is undoubtedly a good thing. But there are other steps that governments can take to dramatically enhance the transparency of the public procurement process without adding steps that slow the process down and discourage many prospective vendors from participating.
Some things we can do to make procurement better
With all of this in mind, here are, in no particular order, five suggested changes that can be adopted to improve the government procurement process.
Raise the threshold on simplified / streamlined procurement
Many governments use a separate, more streamlined process for smaller projects that do not require a full RFP. In the City of Philadelphia, professional services projects that do not exceed $32,000 annually go through this more streamlined bidding process. When I worked in Philly, we had some success in using these smaller projects to test new ideas and strategies for partnering with small local vendors. There is much we can learn from these experiments, and a modest increase to enable more experimentation would allow governments to gain valuable new insights.
Identify clear standards for projects
Having a clear set of vendor-agnostic technology and data standards to use when developing RFPs and in performing work can make a huge difference in how a project turns out. Clearly articulating standards for:
- The various components that a system will use
- The environment in which it will be housed
- The testing it must undergo prior to final acceptance
…can go a long way to reduce the risk of uncertainty inherent in technology projects.
It’s worth noting that most governments probably already have a set of standards that are usually made part of any technology solicitation. But these standards documents can quickly become out of date and they must undergo constant review and refinement.
In addition, many of the people writing these standards may confuse a specific vendor product or platform with a true standard.
Require open source
Requiring technology projects to be open source during development or after completion can be an effective way to reduce risk on an IT project and enhance transparency. This is particularly true of web-based projects.
I’m not suggesting that every vendor needs to open source their entire platform, but if governments are contracting with outside companies to build new, custom software solutions—and there are no security or other considerations that would make it impractical—then why not make them open source?
Government RFPs should encourage the use of existing open source tools—leveraging existing software components that are in use in similar projects and maintained by an active community—to foster external participation by vendors and volunteers alike. When governments make the code behind their project open source, they enable anyone that understands software development to help make them better.
Develop internal capacity for IT project management and implementation
Governments must find ways to develop the internal capacity for developing, implementing, and managing technology projects.
Part of the reason that governments make use of a variety of different risk mitigation provisions in public bidding is that there is a lack of people in government (particularly at the state and local level) with hands-on experience building or maintaining technology. There is a dearth of makers in government, and there is a direct relationship between the perceived risk that governments take on with new technology projects and the lack of experienced technologists working in government.
Governments need to find ways to develop a maker culture within their workforces and should prioritize recruitment from the local technology and civic hacking communities.
Make contracting, lobbying, and campaign contribution data public as open data
One of the more disheartening revelations to come out of the many postmortem reviews of the healthcare.gov implementation is that some of the firms that were awarded work as part of the project also spent non-trivial amounts of money on lobbying. It’s a good bet that this kind of thing also happens at the state and local level as well.
This can seriously undermine confidence in the bidding process, and may cause many smaller firms—who lack funds or interest in lobbying elected officials—to simply throw up their hands and walk away.
In the absence of statutory or regulatory changes to prevent this from happening, governments can enhance the transparency around the bidding process by working to ensure that all contracting data as well as data listing publicly registered lobbyists and contributions to political campaigns is open.
Ensuring that all participants in the public bidding process have confidence that the process will be fair and transparent is essential to getting as many firms to participate as possible—including small firms more adept at agile software development methodologies. More bids typically equates to higher quality proposals and lower prices.
None of the changes list above will be easy, and governments are positioned differently in how well they may achieve any one of them. Nor do they represent the entire universe of things we can do to improve the system in the near term—these are items that I personally think are important and very achievable.
But it’s time for us to stop pretending that the problems with government technology procurement are an accident.
If we’re going to fix things, we need to design a better system.