Buy, Buy, Buy: A vision for US Federal government procurement

Introduction

This DPI 662 reflection was inspired by the case of California’s Child Welfare Services, and how they disrupted technology procurement in 2015 — deciding to replace their system in modules through an agile development process. This case highlighted the risks, structural impediments, and promises posed by a shift to agile procurement. Admirably, California also smoothed the path for other states to follow their lead, using open source code in their module development and sharing the code freely.

I’m very passionate about US federal government innovation, so this case sparked my curiosity about how we are envisioning the future of procurement at the federal scale — and whether lessons from cases like CWS are being woven into that approach. The federal government spends ~$500B on contracts annually (roughly the size of Sweden’s economy), with DoD leading the charge by an order of magnitude.

Quick review: summary of the benefits of agile procurement

• Risk reduction: A benefit we see across applications of agile is that a modular approach actually can reduce risk — allowing us to test and learn, and change tack based on what we’re learning from feedback and outcomes
• Flexibility: Allows agencies to add additional contractors as needed — similar to what Singapore recently pulled off with their dynamic contracting initiative
• It’s not waterfall: Most waterfall approaches fail, and there is demonstrably a need to test out new modes of purchasing. Some argue that agile procurement is an inevitable and necessary capacity to build; so this is simply a critical piece of capacity-building

Source: CWS case

Recent shifts in US Federal procurement

COVID-19 appears to have, in some instances, accelerated the federal shift to agile sourcing processes. For example, the Department of Homeland Security recently launched The Procurement Innovation Lab, a framework ‘aimed at experimenting with innovative acquisition techniques across the DHS enterprise.’ For each of the procurements it has been involved it, it has shortened the contracting cycle by 20–50%, an impact metric that has encouraged at least 4 federal agencies to adopt this framework in 2021.

With perhaps a greater magnitude of impact, the Department of Defense rewrote Directive 5000.01 in 2020, aiming to comprehensively redesign acquisition policies. It ‘aims to empower program managers, simplify acquisition policy and rely more on data-driven analysis.’ In tandem, the Defense Acquisition University has trained >1400 personnel in software purchasing over the past two years, and DoD is testing a new ‘color of money’ pilot that would allow its contracting offers to purchase software under its own budget activity.

Quick TLDR on this ‘color of money’ business: it means that DoD isn’t beholden to policies designed for purchasing tanks when purchasing software.

All of this sounds quite promising, right? Absolutely. But also, it’s so critical that US federal agencies — particularly DoD — are learning from other case studies and anticipating challenges that may arise.

Some considerations to flag

Placing an end-user at the core of the decision-making process in a traditionally top-down organization “is not without its challenges…Working in an iterative approach with distributed decision-making is really counter to what the state has used for building these systems in the past.”— Peter Kelly (CWS part B)

• Necessity of cultural shift: From my work doing agile transformation consulting within large, bureaucratic organizations, I’ve seen firsthand that it’s not enough to ‘do’ agile, organizations (and stakeholders within them) must ‘be’ agile.
• Role of data: Agencies must enable data-driven analysis to improve acquisition, and must improve data transparency. These are non-negotiable pieces of a successful transformation towards agile procurement. Government leaders must be able to use data in decision-making; to not default to buying closed, off-the-shelf procurement systems where they don’t own the data; to build a truly open and digital system.
• Reminder of the stakes: With DoD and DHS as the case studies we’re looking at here, it goes without saying that the stake of cultural and procedural change are incredibly high. In the case of CWS, we saw similarly that the lives of very vulnerable children would be directly affected by the outcome of the pilot. In addition to the end-user impact, organizations ought to consider 1) the reputation risk for actors involved, and 2) the risks of failure, particularly in discouraging other agencies that have not tried this yet
• Impact on government suppliers: In tandem with these shifts, we should also begin to equip government suppliers with tools and guidance about how to compete within an agile procurement environment (e.g., the value of stepping up as a SME).

Conclusion

As the California Child Welfare Services case illuminates, the shift towards agile procurement in government is incredibly promising, but also bears key lessons and considerations. As we witness a parallel transformation take place at the federal scale, agency stakeholders ought to learn the lessons of such smaller scale pilots — after all, it is the agile way.

This blog post was written in response to an assignment for the course DPI-662 Digital Government: Technology, Policy, and Public Service Innovation at Harvard Kennedy School.