What if Zohran’s PIT crews were on a federal racetrack?
The first administrator of the U.S. Digital Service thinks it's time to think bigger
Mikey Dickerson was the first administrator of the US Digital Service under President Barack Obama and now leads Searchlight’s Tech Viaduct — a team of technologists working to develop a gameplan for a future federal administration that is committed to effective government service delivery.
Earlier this month, Mayor Zohran Mamdani launched New York’s new Public Interest Technology (PIT) crews — five teams of local government technologists working to improve the city’s digital public services. It’s a much-needed bit of good news that has triggered a huge outpouring of support and enthusiasm in the “public interest technology” community, including from the Tech Viaduct team. It also raises the urgency of our work to reach a clear-eyed understanding of what we have all learned in the last decade of “digital service” reform.
As veterans of the last two Democratic administrations, we are acutely familiar with the many obstacles government puts in the way of doing the right thing. And we know how to circumvent them with the usual array of hacks and incremental workarounds. But we had never really tried to think about what big-picture digital service reform would actually look like.
In the last few months, our loose federation of fellows, contributors, advisors, and other well-wishers has interviewed about a hundred people, read thousands of pages of reports, and participated in adjacent projects and convenings with the Federation of American Scientists, Future of Government Initiative, Center for Civic Futures, and others. As a result, we have new insights about why this goal is difficult, and how we can chart a path towards real government service reform.
The “digital service team” era was organic, not strategic
Before I share some of these insights, it’s important to understand how we got here. I was the first administrator of the US Digital Service, and as I look back on my ten years of experience working with “digital service teams” in various American jurisdictions, it’s hard to make sense of the results.
I’m not the only one who has had this experience. As a recent report from the Federation of American Scientists found, there isn’t much agreement among the people on these teams as to what their strategy or purpose was. One reason is that when USDS was founded 12 years ago, it never implemented a grand plan. The projects it worked on were important, but never grounded in any particular unifying vision.
We inherited the ruins of a number of big government-reform projects that had their roots in the George W. Bush, Clinton, Reagan, and Eisenhower administrations. Experts debate how effective they were, but certainly none of them were finished.
There wasn’t anything we could do to clear these previously stalled strategies out of the way either — each one was embodied in a whole bureaucracy of its own. We instead tried to do as much as could be done, with the time and people we had.
This is the reality of USDS, the erstwhile GSA-based team called 18F, and all the similarly styled teams of the last decade. The budget, funding source, and reporting structure are decided once, at the beginning, when we know the least about the problem. These decisions constrain and channel the work that the bright-eyed team takes on. They have a large influence on the projects and shape whether they succeed. Eventually, various people retrofit a “strategy” or “theory of change” to the projects and mythmaking of the past. In my experience, the bright-eyed team members themselves are not aware this has happened. So, downstream, you find a multiplicity of strongly held beliefs about what the strategy was.
For instance, the Obama-era USDS ran significant projects at HHS, DHS, VA, and the Department of Education. The actual reason those agencies were chosen boils down to the fact that they were what was possible given the internal politics at the time. But the rank and file had little visibility into that. One could look at the set of choices and reason that the “strategy” of USDS was to build citizen-facing services. Or, one could conclude that the “strategy” of USDS was to support presidential priorities.
The two theories predict extremely different outcomes when the presidency transitioned to Trump. If the purpose of USDS was to support presidential priorities, well, those priorities are tariffs and deportations. So is that what USDS works on? Longtime staffers, changing leaders, and new political overseers have different answers. The resulting conflicts weakened and distracted the organization throughout the Trump and Biden years.
While researching the book Crisis Engineering, we found that the pattern is the same in any large organization. Most of the time, a “strategy” is a story that was retrofitted to events that were forced by a crisis. Stories are important because they hold the organization together, but you can rarely reason forward from one and predict what an organization is going to do.
The next era of digital service reform needs to be grounded in a broader mission.
We need to think much bigger
A second, related observation from tech-reform research is that almost no one is thinking beyond problems that require immediate solutions. Digital service teams have evolved in isolated, hostile environments. Like the plants that grow wild in the desert around my house, they are extremely parsimonious with resources, and put most spare calories into defense mechanisms.
Today’s population of civic and government-tech practitioners has been working for years in isolation from mainstream government processes, with shoestring budgets, and little influence over policy design. Thus, we are mainly good at generating ideas that act as band-aids on failing systems.
The irony is that “propping up failing systems” is another way of saying “preserving the status quo.” I find Jen Pahlka’s Three Horizons framework to be a helpful way of sorting through this problem. Horizon 3 is the future of government reform we’re working towards, Horizon 2 is the place where that change occurs, and Horizon 1 is the unacceptable present. Band-aid work fits into what Pahlka calls “Horizon 2 minus” — it is needed, and it’s happening now, but it ultimately serves to entrench the unacceptable status quo.
It follows that, ten years in, we find digital service teams that had success making incremental improvements to a dilapidated program, and have internalized their working constraints so well that they are resistant to any further change.
We have met the enemy, and he is us.
Of course, this mindset was partly created by the impoverished environment. Even digital-service teams structured to do interventions have been like ambulance crews with no hospital to drive to. The digital services team that was tasked with assisting healthcare.gov during the crisis period stayed open for more than ten years after the launch. That was excessive. The need for frequent changes persisted (and always will), but the war-room model dragged on because the agency was not capable of managing those changes outside of a multi-year procurement cycle. The situation became akin to keeping a patient on emergency-room dialysis for ten years because there was no capability to do the needed kidney transplant.
Searchlight fellow Marc Dunkelman examined the history of progressive politics writ large in Why Nothing Works, and found that the pattern of “poor implementation leads to timid policy leads to poor implementation” is by no means limited to the gov-tech arena. It is hard to get anyone anywhere to think about “Horizon 2 plus” ideas, meaning the near-future actions that accelerate progress towards a “third horizon.” But that is exactly the sort of thinking we need to usher in this new era of digital service reform.
Where we go now
As people who still believe that the government could function better than it does, we are in a window of opportunity that won’t last long. The two challenges above happen to mirror what’s been called the “Stockdale Paradox” — the need to both maintain the faith that success is possible, while still staring down the cold hard truth of how hard that will be. With future-facing efforts like Tech Viaduct, we need to expand our imagination of what’s possible and realize that we can get there. But we also need to confront the frustrating facts of how hard that will be, especially given that even more capacity may be lost in the Trump administration.
Recent history tells us to avoid policies that depend on empty announcements, “pilot programs,” or hope that masses of people will see the light. Tech Viaduct aims instead to find points of higher systemic leverage, as defined by Donella Meadows.
Past efforts have gone no higher than “modifying information flows,” which Meadows ranks as the fifth-most effective type of change. We now need to overhaul the rules, incentives, constraints, and distribution of power. This is the only way to break the cycle of perpetual emergency-room dialysis.
We need generational changes in the federal systems for workforce management, procurement, and data governance. These problems are enmeshed in, and connected by, technology. Tech Viaduct will figure out those parts, but will not have the complete solution. We need our partner organizations, aspiring candidates, and fellow schemers to push their writing and thinking to be bigger. Err on the side of being too ambitious, even though you may be laughed at in the DC scene. We can always water down the policies later. It’s really hard to water them up.
 |
|