Mikey Dickerson
‘Obviously, We Made It Up As We Went Along’: Mikey Dickerson on Being the U.S. Digital Service’s First-Ever Administrator
| Mikey Dickerson was the first-ever Administrator of the U.S. Digital Service. He served from 2014 to 2017, and grew the organization from a staff of two to 200. Mikey came to the role after his work on the HealthCare.gov rescue. Prior, Mikey was a site reliability engineer at Google. |
Mikey Dickerson was the first-ever Administrator of the U.S. Digital Service. He played an outsized role in determining the organization’s priorities, its culture, and its relationships across the federal government.
Below, Mikey discusses the difficulties of hiring people to work on crises, tempering ambitions, and what he would have done differently.
April 30, 2024
Emily Tavoulareas:
Mikey, where were you before the U.S. Digital Service, and how did you come to be the organization’s first administrator?
Mikey Dickerson:
I was working at Google. I was hired in 2006 as an entry-level site reliability engineer (SRE), which were words that didn’t mean much at the moment but became more defined over time. I was at Google until 2014, and by the time I left I was middle management.
I went on leave for short times in 2008 and 2012 to work on the Obama campaign as a volunteer. That led to them finding me and other people in 2013 to work on HealthCare.gov. I don’t really know who “them” is in that statement. It’s at least the President and Todd Park.
When that was all over and had been retroactively declared a big success, Todd and Jen Pahlka talked to me about this USDS thing maybe happening. I got introduced to Haley Van Dyck somewhere along the way, and then those three talked to me about working at the White House.
The HealthCare.gov project ended — and had been such a nightmare. It went on and on and on. I think it finally ended April 30, when they declared the last real, final, no-more-extensions-this-time day of open enrollment. I went back to Google, but I didn’t do any more work. I was exhausted and burned out. A couple months later, the mumbling about USDS started. Well, I’m sure it started years before that, but that’s when I first started to hear about it. The summer of 2014 it started to get serious, and I agreed to go. We set a first day of August 11, 2014.
Everybody has their own story of what happened on HealthCare.gov, and mine has me as the main character. That’s how it works. In that story, the four people I spent the first few weeks with were Paul Smith, Greg Gershman, Ryan Panchadsaram and Gabriel Burke. Later we added more people: Jini Kim and Stephanie Van Dyk. Eventually I brought out a lot more people from Google. But the original group of five was put together by Todd, not chosen by me.
Emily:
So your time with HealthCare.gov ends April 30, you hear mumblings about USDS, and then over the summer you decide to join. What happened in those few months that led you to your decision?
Mikey:
I was burned out on Google and wasn’t going anywhere. I had been at Google for a long time and when I left to do HealthCare.gov, a lot of people felt left behind. There was more than a little resentment there. When I went back after HealthCare.gov it was a nightmare for me, and no one else enjoyed it, either. I was gone for two days, then, two weeks, and then months. When I did come back, I wasn’t useful. All I could do was sleep all day long.
My work on HealthCare.gov had also become this big press thing, which people reacted to in a variety of ways. Some people said, “Wow, that’s so great. I understand now why this was so important, and I’m so happy for you.” Other people resented that I got that much attention and that it looked great for my career. When I went back to Google, I promised them that I was going to stay for at least a year, but I left a few months later.
I didn’t feel a ton of pressure to join USDS. Although it’s impossible to talk to Todd and not be somewhat pressured, because he is just so enthusiastic. Jen Pahlka is similar. With how long it took, I’ve always assumed that I was not their first choice. Also: everyone my age watched “The West Wing” and thought it was the greatest thing in the world. So when someone called to say, “You’re going to come work at the White House,” could I really just stay home?
Kathy:
You touched on two interesting points. One, what it was like to leave a role or whole career to come into government, and how disruptive that is. And two, I think it’s irrelevant which choice you were for administrator. Because you showed up and you took the job, and that’s such a theme — just showing up to take on the job.
Mikey:
It gets forgotten that, just like the HealthCare.gov project, we did not know what the story was at the beginning. It might have been a total failure. It might have been forgotten in six weeks. It might have never gotten any attention at all. I might have gotten to Washington, D.C. and found out that everything I’d been told was lies. I might have had an intern desk at the end of the hallway and no one even remembered that I was there. All of these were definitely possible, and they were possibilities for all the early people that went out there. They didn’t have much to go on except my word. Obviously, I didn’t know what I was talking about. I’d been there for two weeks. All of those people took massive risks and paid high costs to leave their old job, their old home, their old partner.
Emily:
One of the things we’re fighting both in the classroom and outside is the mythology that develops over time. You’re reminding us that no one knew what was going to happen. No one knew how this was going to go. It could have evolved in any number of directions. I’m curious: Deep down in your heart when you took the step, what did you think was going to happen?
Mikey:
I had no idea what I was getting into. Well, that’s not totally true. I had been introduced to Jennifer Anastasoff. And I had heard Todd’s big dreams about hiring 200 people and transforming the whole federal government. Whether I really thought that was possible or not, I don’t know.
I pictured more projects like HealthCare.gov: A smaller group of people working on something where my job was a lot more operational. I imagined I would do two or three projects like that and then call it after a year, if I’m lucky. I knew that it had to be a more regular job than HealthCare.gov, because I couldn’t work seven days a week, 16 hours a day forever. But that is not quite how it turned out. It was never like a regular job.
I had very confusing signals from all kinds of people. Around the time I accepted the job, I talked to Beth Cobert, who was Deputy Director for Management at the Office of Management and Budget (OMB) and Steve VanRoekel who was Chief Information Officer (CIO) at OMB. In the original conception I was reporting to Steve. And Steve reports to Beth, so these were my managers. They said, “Oh, you could do this job half-time from California. You could telecommute.” That did not make sense compared to my HealthCare.gov experience. But it was a clue that people had wildly different ideas of what this job was going to be. Obviously, we made it up as we went along.
Kathy:
I want to build on that idea of “We made it up as we went along.” Within the first six months or year, how did the structure of USDS evolve? Was it out of necessity? And what was the thought process behind that evolution?
Mikey:
On day one, USDS didn’t exist yet. When I first walked through the gates, it was me and Erie Meyer, the first employee. She agreed to sign up before she had even met me. A couple days later, Jennifer Anastasoff showed up. Then Haley showed up. Viv Graubard was there, but she didn’t work for USDS yet. She worked in the Office of Science and Technology Policy (OSTP).
All the other offices are grudgingly realizing that they’re expected to include me: “Great, another person at the White House. Now I have to add another person to our meeting on everything under the sun that touches technology.” From the minute I got off the plane, it was 14 meetings in a row per day, and I didn’t even know where any of them were. Erie is taking me from place to place. There was a camera crew because they were doing this West Wing video. So I’m going to an HR orientation and it’s totally for show. I’m sitting there for two minutes while they film me and then I’m rushing off to the next thing.
What I needed immediately was what I would now understand as a principal’s staff. By the time I finished at USDS, there were at least three or four people whose main job was making sure my schedule was managed, that I knew where to go, and that I was something resembling prepared. The first few people on the scene did that role, whether it was what they signed up for or not.
We were going to build project teams after that. Because Erie was there on the first day, I assumed she wanted to be front office staff. But she said, “No, that’s not what I want to do at all. I want to go to the Department of Veterans Affairs (VA) and work on projects.”
It was also my job to figure out our relationship with the White House. I was thinking, “Which meetings do I get invited to? Do I sit in the front row or the back row at the deputy’s meeting? If both me and the CTO are in a meeting, who is going to talk?” All this is being figured out on the fly, but they really care about this stuff.
At the same time, there are these outposts scattered all over, like 18F at the General Services Administration (GSA). Emily, you were already at the VA. There are Presidential Innovation Fellows (PIFs) here and there. There’s the Office of the Federal CIO.
Right out of the gate, they are our “allocational rivals,” as Anthony Downs would say, because we’re paid out of the same budget. And that means we’re going to have a difficult relationship, at best. The relationships started all over the spectrum, from “This is so great, we’re going to have more connection and visibility,” to “This is so terrible, now there’s another person breathing my oxygen.” Every meeting I walked into, if I had any clue, it was whatever Erie had frantically told me in the hallway before we opened the door.
Kathy:
Mikey, can you talk more about figuring out USDS agency teams? And also the Communities of Practice (COPs)? Because those happened pretty quickly out of necessity.
Mikey:
Agency teams got cooked up as an idea very early on. We had this notion that if USDS was going to transform the whole entire government, it had to colonize. Yes, I know how fraught that word is — that’s why I’m using it. And as I sit here now, I think, “That had zero chance of success. Why did I ever let that get tried?”
But we were going to colonize all of the federal agencies by implanting a digital service team lead. We were going to use the power we had at OMB to mail out guidance to the agencies that said, “You should request money for your USDS digital service team.” It was a 40-step plan.
We wrote a position description and we got it through the Office of Personnel Management (OPM). I spent a lot of time on that. They’re still using it: A thing called “Digital Service Expert” which is graded at a GS-14/15. It was written really broadly because I didn’t want to do it three more times; that’s why the same position description was used for engineer, designer, and product manager. We had this very vague notion, which never really came into focus, of a bureaucracy hacker or operations specialist. It was something that we needed, but we couldn’t figure out how to give it an identity that made sense.
We also did this roadshow with PowerPoint slides and all of this ceremony. Because scheduling a meeting with the deputy secretary of an agency is like meeting with the prime minister of a mid-sized country. Based on HealthCare.gov, I believed there was no hope of USDS working without the patronage of a high-up executive; someone everyone in the agency is afraid of. There’s politer ways to say it, but that’s what matters. So we aimed for deputy secretaries. We met with nearly all of them. This dragged out for a year, and it took up enormous amounts of attention.
It wasn’t a great idea. Those meetings are theater. I am going there with these PowerPoint slides that someone has spent hours and hours on, and they’re in complete and total fantasy land. The agency is also showing up with slides that took hours and hours, and it was also total fantasy. Like, “What if we had a 20-person team here that was extremely capable and extremely empowered. What could we do?” The thing was, we did not have 20-person teams for everybody; we weren’t able to put a team at all 24 CFO Act agencies. That was never within the realm of possibility. We were only ever going to have a handful, at best.
I was also meeting with Kristie Canegallo, the Deputy Chief of Staff for Implementation. She worked for Denis McDonough, who was the Chief of Staff to President Obama. Kristie was very involved from the beginning. I would bet that with the agencies where things actually moved, it was because Kristie had a relationship with them. My roadshow had nothing to do with it.
At the time, there were four or five places where things were actually going to go somewhere, and 19 other agencies where nothing happened. Some agencies don’t get a lot of attention from the White House and this was going to be a big opportunity for them. We didn’t deliver anything, and they resented that years later. It was even brought up four years after I left the White House, in meetings with people who used to be in charge of those agencies.
We met with Labor later in the roadshow, and by that time the scope of what we could do was better known. Labor said, “We could do this, this, this, and this.” And I said, “I don’t have the people for any of this. I don’t know what we’re actually going to be able to do here.” The answer, of course, was actually “Nothing.”
Emily:
At what point in the course of that year did you think, “This is not feasible. Why am I even going to these meetings?”
Mikey:
I don’t remember a clear turning point. It developed over time. Halfway through, it was clear this hadn’t been a good idea. We had bitten off way more than we could chew. It was also becoming clear to me who I could actually hire. When we started, the only model I had was HealthCare.gov. Todd and I had this idea that, based on the HealthCare.gov experience, there are people rattling around in the tech industry who have already made enough money and they’re losing interest in their middle management Google job. They would come out of a desire to do public service. It was like tours of duty; Schedule A hiring, which is term limited.
I wrote two years as the term that got approved through OPM. We were looking for people for whom this was not a career, but an interlude. They would go back to the tech industry. It was deliberately taking an outside culture and putting it inside the government to see a disproportionate impact in a short period of time. They were not going to become the new agency CIO. But it got muddled immediately because we had this press attention. And the agencies have like 15 billion open CIO jobs, deputy CIO jobs, and CISO jobs, and they’re trying to fill these things all the time. It’s very, very hard to do.
So Jennifer and Todd are bringing people into the organization who are not there for the HealthCare.gov project-style work. They want to do something else. Some of them want to wear suits and go to fancy OSTP meetings. That became a category of people who started appearing in my world. Some of them were people who came from the tech industry, were very capable, and wanted an executive-type job. A lot of them became agency team leads or filled those CIO-type jobs. But I hired them first, so they’re technically my responsibility — even though Todd had recruited them, and who knows what they got told in that process. So they don’t really think of me as their boss. They think of me as a person who signed a paper that got them their important job.
Also, I only wrote one position description for USDS hiring and it was graded GS-14/15, and the very lowest salary is still pretty high on the government scale. So some people were attracted to USDS because it was a career path — a big jump in rank, a big jump in salary. This is now a population that needs to be managed differently, and that I created by accident. They’re in it for career benefits; it’s not a tour of duty to them. Suddenly I’m having arguments about work-life balance inside USDS, which was never meant to happen.
There were a lot of feelings about term limits and a lot of shenanigans started to happen. Like, “We can put you on a different payroll over here and start the four-year clock over again…” If you wanted to build an organization that way, it should have been done differently from the beginning. We didn’t create any internal promotion paths. It wasn’t meant to be a long-term home for anybody. To make it a long-term home for everybody, you would have to create a lot more structure and process than I did.
So in short: Over the course of those 18 months, I got more clarity on the type of people I’d be working with. We disproved the theory that there would be lots of people who wanted to temporarily leave their big fancy tech job and come to the government for a tour of duty. We only got a good couple dozen. By now, if that theory had been successful, we would have had a thousand people like that, instead.
We started this whole thing two-and-a-half years before the end of the Obama Administration. So after year one — which is not very much time in the life of a startup — the end was in sight. We barely established the first generation of USDS, and we were already starting to have meetings about institutionalization and transfer to the next administration. By the end of the first year, I had a lot more information on what the agencies were capable of. And just to be blunt, the answer most of the time is “not much.” The digital transformation everybody fantasizes about almost never happens. Most agencies are not even interested.
Every single one of them will say it’s what they want to do. But they also say, “We cannot bend any rule anywhere. So before we can get started, we need to do this nine-month CISO architecture review.” Most of the agencies are actually not going to move, period. It doesn’t matter how hard we try. It took some time to figure out what we, USDS, were really going to be capable of. I had a much better picture of that, but it wasn’t like a big epiphany one day. It was a blurry JPEG that got clearer over the course of a year or two.
Emily:
So you were figuring out, “Based on the humans that I have here and what I understand about the agency, where should we put our energy and our time?”
Mikey:
Yeah, and we had some good success. If I had four or five people who worked on the guts of big tech transactional systems — like the Google web search cluster or Twitter — I could send them to a place like the Department of State and expect to change core operations very close to the mainframes. They could re-engineer processes that were COBOL- and mainframe-based and untangle things at a very fundamental level.
But if I have people that know Rails and can make websites, yet haven’t ever worked on anything like that, we can still do stuff. At the Department of State, we can do the visa status tracker, but it’s going to be attached to these agencies’ existing systems, and it’s going to have to be something that works with the existing data and API. Those are still valuable projects, but they have a lot more user-facing surface area, calling for a lot more UX research and design. They have a lot less business process reorganization to them, so we had to change what we were aiming at.
Emily:
So there was a tension between the firefighting and the dev-type work. Your assumptions coming in were that USDS would do firefighting SRE-type stuff. But over time, you end up getting fewer of those people and more of the people who built user-facing products.
Mikey:
Yes. It was really, really hard to walk that line. I hurt lots of people’s feelings by saying that the original idea was this other thing. But the dev-type stuff needs to be done, too. At no point did I think that the government doesn’t need all of those teams. It absolutely does. My only claim is that we didn’t keep our eye on the ball on the firefighting, because the government really needs that, too. If USDS doesn’t do it, then we’re back to where we were — there’s no one who can do it.
We’re watching that right now with the Federal Application For Student Aid (FAFSA) issue. This was exactly what we were supposed to be there to fix or prevent, and it’s not happening. For posterity: This is spring of 2024 and the Department of Education is a year behind schedule on the simplified FAFSA. They’ve missed all the deadlines. I work at a college right now, and they don’t have their financial aid decisions for admitted students for the fall. It is super fucked.
And I teach this in my class now: It is nearly impossible to preserve the conditions where people are not prioritizing their own self-interest in their career path. Some of them have this transference where the thing they now care about is the reputation of the organization. But anybody with that motivation isn’t going to sign up to work on HealthCare.gov or FAFSA, because there’s a lot of risk and it might not go well. I know this. I understand.
You have to construct a different isolated culture that doesn’t assimilate. USDS needed to be an exceptional culture that did not assimilate to the rest of the government around it, or it would just act like the rest of the government. The people there need to be interested in fighting fires and responding to different incentives. To drag out this analogy more: If a fire gets big enough, the firefighters aren’t going to be able to put it out. They’re going to do harm reduction the best they can. And when it’s all over, you don’t say, “This skyscraper burned down because the fire department was incompetent.” That doesn’t make any sense. They’re there to fix what they can after the disaster has happened.
If you can maintain that mindset, then you can have people who are happy to go to a place like Ed and work on FAFSA. But it can’t coexist with people who are worrying about maintaining their relationships because they want to get a job at Ed five years from now. I let a whole bunch of things coexist that really can’t coexist. That led to the perpetual fight over, “Are we here to make friends or are we here to force people to change?” Because you can’t do both.
I don’t even know how many meetings were spent agonizing over, “How are we going to justify 18F and USDS to Congress?” I wanted the answer to be what you said, but I was not policing USDS hard enough. Which meant that the stuff USDS was doing was bleeding out over all kinds of projects. 18F was doing the same, trying to survive, and so they were not staying in a lane, either. There was a ton of duplication. We could have cleaned that up, but did not get very far with it before I left in 2017.
Emily:
Ironically, over time they seem to have become more similar, not less.
Mikey:
They’re evolving in the same niche now. They have the same population of people to hire, and they have the same population of projects available. It makes perfect sense to me that they would be indistinguishable at this point. They would be Coke and Pepsi.
Emily:
I’ve long wondered about how focused we should be on longevity… is the goal to build an organization that is going to last and grow, or is the goal for talented people to go in, do the things, set up internal capacity and walk away… or both?
Mikey:
Yeah. That’s where the rubber hits the road: Are we going to take on something that has forever maintenance obligations? Another thing to know about agency teams is when we do successfully colonize an agency, it means they’re paying their own people on their own budget under their own hiring process. What control does OMB or the USDS Administrator have over them now? The answer is none.
Emily:
You’ve been talking about how the intent and original vision evolved over time, especially in terms of hiring, organizational structure, priorities, and approach. From your perspective, have these changes over time at USDS changed its ability to make an impact?
Mikey:
Well, the latest spending bill [in 2024] proposes there be no appropriated money for USDS at all. Instead, it is allowed to exist only if agencies pay it to exist. That fundamentally changes its nature. If USDS has to now convince existing project leadership to hire them, it loses all of its valuable functions like challenging existing assumptions and being able to bring an independent perspective. All of that advantage is gone. They’re just another vendor now, and they should be expected to be treated as just another vendor.
For example, if the Free Application for Federal Student Aid (FAFSA) process burns itself to the ground and the Department of Education passes the buck, USDS can do nothing about it. Because Ed has to decide they want to hire us before we can go there. That’s completely different from being a fire department that puts out the fire first and then worries about who’s going to pay for it. The federal government needs that same function for its broken projects. Otherwise it’s like having a contract with a private fire agency, and when my house is on fire, I call them up to start negotiating the price.
Emily:
Considering all of this, what are your hopes for the future of USDS?
Mikey:
The version of it that was worth doing is the version that was originally put in place: a smaller group, more responsive. There is a book by Arthur Squires about his experience working on the uranium gas diffusion plants in Oak Ridge in World War II. It is shocking; all of the systemic dysfunction with the vendors is exactly the same. All of the problems of government are exactly the same.
It’s a book length study of operating as a breakout organization, because the Manhattan Project was exactly that. There was a big challenge which was unlike anything the Department of Defense (DOD) was currently doing. So they hired a whole bunch of weird people they were unlikely to hire under ordinary circumstances, they sent them off to the desert, and they let it run hands off for a while in hopes it would produce something. And it did.
You need to be flexible in your hiring authorities. You need to let the recruiting proceed through existing networks of expertise. You absolutely cannot staff this thing in the Office of Personnel Management (OPM) style of, “We’re going to read resumes…” So to your question: The government will need a breakout organization. This isn’t a matter of opinion. It will rediscover it again when it needs it again. It’s not a priority right now, and that’s why everything has atrophied. It has assimilated to behave the same way the rest of the government does. There’s no reason to be using up all of these EOP passes, office space, and all the rest of the resources for something that’s no longer any different from the rest of the government.
Emily:
Is it possible to carve out a pipeline if and when the need for a new breakout organization occurs?
Mikey:
In my last year on the job, I was compelled to go to this endless series of meetings that they called “institutionalization,” which was about how we’re going to make sure that all of this continues forever and ever until the end of time. It was a goal statement that I did not buy into then and I don’t buy into now.
It is a very bad idea to try to institutionalize the exact things we’re doing at a given moment, which are this big grab bag of tribal beliefs. There’s going to be a new fad in five years from now. I can point to 100 other examples in the federal government of things that seemed like a good idea in 1972, and now we still have somebody whose job it is today because that’s just how this works.
The way you could have done that is with a skeleton crew. It might’ve been a great outcome to let the organization shrink down to something that is significantly smaller. Let’s imagine the parallel universe where that happened. When something really rose to the level of crisis and the West Wing is actually paying attention and cares — the Afghanistan withdrawal comes to mind — you could have had the skeleton crew who remembered what was written on the old stone tablets and had an idea of how you spin this up, and there was a big budget sitting there ready to be used. You could run the whole show again, you could put out the call. You could turn on that bat signal again. You could recruit from that pool, you could do it with all of the priority, the urgency, and the whiz-bang, razzle-dazzle, and could run that whole thing again.
Emily:
Looking back, are there any major milestones or turning points that stand out to you?
Mikey:
The first time the whole dynamic changes is when you go from zero to 15 or so people. Everybody fits in a room and knows everything that’s going on all the time. Everybody feels like they get to participate in all the decisions. This is Jeff Bezos’ “two pizza team” idea.
When we got to 20 or 25 people, everybody didn’t fit in one room anymore. People started to feel like they were not being consulted on all the decisions. There’s a lot of conflict, and it’s still an issue to this day. Not everybody landed in the place that they wanted to. Specialization started to happen, and who gets to be in the room became a thing people fought about. This was something we were up against within the first six months.
At 50 or 60 staff, another thing happens: You start to not know everything that’s going on. We had to do intentional things to manage that, like the staff meeting where people present their work. But it was another really high-anxiety period. This period was also characterized by low points of trust in leadership. We had a team offsite, and Haley and I were 30 minutes late because we made a wrong turn. The staff present told the facilitator, “They probably just abandoned us.” That’s how little they understood or believed that we were there for them.
It cooled down quite a bit after that offsite, which was the point. That’s why you have to do offsites. I’ve run organizations of this size inside Google, and they fall apart if you don’t invest all of this time. Especially if it’s growing and it’s changing. It cannot remain a coherent thing that everybody feels like they’re part of the same team. But government doesn’t understand that.
Another inflection point is when you get to Dunbar’s number: 150 people or so. At that point, not only are you no longer part of every decision, but you also don’t know all the people anymore. There were people working at USDS who you’d never met and would not recognize. It was very obvious when this happened to USDS: Around that time, when we had 153 employees, someone’s iPod got stolen off of a table in Jackson Place. That kind of behavior doesn’t happen when it’s 15 people.
Those were the internal growth dynamics at USDS. Externally, for a while, they did everything but throw us a ticker tape parade. We were the heroes who had just bailed them out of this HealthCare.gov clusterfuck, which everybody whose career depended on the Obama White House knew was their legacy. I was a government employee for all of three days before they had me come and speak at Camp David to the Office of Management and Budget (OMB) leadership. That honeymoon lasted three to four months. Then everybody moves on and starts to forget why we’re there, and the teacher’s pet phenomenon begins: USDS is competing with other wings of the White House for office space, budget, and most of all attention. That’s the most coveted resource at the White House. We started to be “accidentally” left out of meetings and having to deal with a lot more pushing and scratching.
That had fully set in by six to eight months or so, right around the time of the offsite and the internal trauma. We were actively disliked by people that had power in the West Wing, and that led to outcomes like never having another offsite. It was specifically rejected with prejudice. At the same time, there was an attempt to reorganize. There was a question of whether USDS was going to continue to report to the CIO or not. During my tenure, I found reporting to the federal CIO was a recipe for endless, non-productive conflict. We needed to report to the deputy director of management, but that was also rejected. By this point, the sheen of specialness had worn off. We were just another entry on Deputy Chief of Staff Kristie Canegallo’s schedule. The default answer to everything became “no.”
By the time I left Washington, D.C., I was reporting to Andrew Mayock, the Deputy Director for Management at the time. That happened because in August 2016, I said I was quitting. The wormhole to the autonomous zone — the only space where we could get things done — had closed. We were now operating under regular order.
I was happy to be done. This job was unbelievably stressful. It was not very fun. So what did I do? I fucked off from Washington, D.C. and moved to the middle of the desert.
Emily:
Mikey, is there anything in particular that you are proud of when it comes to USDS?
Mikey:
The things that got done where I could see with my eyes that they were actually done. Our number one project was always ensuring HealthCare.gov did not have any more major meltdowns, because that wasn’t a foregone conclusion. It probably needs work right now, 10 years later. We launched Vets.gov. And we did the College Scorecard. There was also the visa status tracker.
The things I feel the best about are the times where we found really broken things that were pretty easy to fix. Like the service records interface between the Department of Defense and the Department of Veterans Affairs. We were slogging through that process and found all these cul-de-sacs: If some set of conditions happened wrong, then you ended up with contradictory values in two different rows of the spreadsheet and you would never be picked up by the next COBOL script. It was as if your application had fallen behind the file cabinet, never to be seen again.
There were a couple instances where intrepid engineers like Alex Gaynor and Marina Martin followed those file folders around the building. We found process screw-ups that had tens or hundreds of thousands of people in them, and no one in the VA was aware of this. These were called “cracks” or “seams.” When we found that kind of thing, we oftentimes were able to fix it but never allowed to tell anybody what had happened, because that would mean admitting how bad it was screwed up.
There were similar examples in visa processing and refugee processing; overlooked leaks in the bureaucratic machine where, because the machine is so goddamn big, a small looking process error can easily have 75,000 veterans denied service. It was worth the price of funding USDS even if they only did things like that a half dozen times a year. It pays for itself 10 times over. Those are the slam dunk cases to me.
Kathy:
Is there anything org- or culture-wise you think of fondly?
Mikey:
I can still recite the values: Hire and empower great people. Find the truth, tell the truth. Go where the work is. Design with people, not for them. Optimize results, not optics. And create momentum.
I must still care about these, because I get mad when I think they’ve been forgotten. For example, I heard that for reasons of convenience, we have reinterpreted “go where the work is” to mean “meet the public servants where they are.” But I meant literally. If all the work is happening in some horrific beige office building in Iowa, then you need to go to Iowa. That’s where they’re doing work. You need to see it with your own eyes.
Emily:
Ok, a related question. So much of what happens in a new organization can’t be attributed to a single person. But is there anything you felt you very specifically had your hands on?
Mikey:
There was a lot of stuff I was individually responsible for, because the government is intensely hierarchical in a way that other places are not. For example, no one but me could sign the form that hired a new person and put them on the EOP. That never changed while I was there.
When I left, we had more than 200 people. And on balance, they were a stronger group than what would’ve happened if we followed the OPM hiring process at face value. And there were lots and lots of things that ultimately no one but me was going to be accountable for, like term limits. I would still defend that one. I wrote or at least edited the final draft of what became the position description for Digital Service Expert. The idea is that it should be a Schedule A appointment that is only good for a couple of years.
Another thing we were successful at: People hung out outside of work. There was a social core. There was a lot of going to karaoke and bars. That is evidence that we successfully created a place where people felt fairly safe. It wasn’t perfect. But compare that to the culture of your average DMV — those people don’t go out drinking together because they would never, ever trust each other enough to do that. And everything that I’ve observed since then tells me that this is leadership’s responsibility. I was reasonably successful at creating a space where people felt like they had protection from noise from above, that their immediate coworkers weren’t going to knife them in the back. That is not the natural culture of Washington, D.C.
Emily:
One of the recurring themes throughout our interviews is how critical that atmosphere was to people’s sanity. Because people left everything on a dime to come do this work, and they needed each other. Having that space and having each other created a very special bond.
Mikey, I’d like to mention milestones other people brought up and get your reaction, like the TED Talk and the agency roadshow.
Mikey:
The TED Talk is what I was referencing when I said that by the end of the administration, we were up against active dislike instead of just massive jealousy. There are some things I did that contributed to us being disliked. For example, we had a visit from the White House counsel’s office to discuss a social media policy. Our people were sharp, bordering on disrespectful in their questioning. I tolerated it because I put up with it myself every week, but I probably should not have. That meeting did not go well, and I had to do an apology tour.
I was not responsible for the TED Talk, but it made people very angry for reasons that I don’t think are good. I still do not see what was so problematic about the content of the talk, and I was never able to dig us out of that hole. It became clear that TED Talk-gate was going to determine the fate of the rest of USDS for 2016. It was why we were not allowed to do an offsite or even use the General Services Administration (GSA) building. Once that picture became clear, I went to war and compiled a folder that was an inch-and-a-half thick with emails. Because everybody kept telling me, “It’s not that we don’t like the TED Talk. It’s not that we don’t like the USDS deputy administrator. It’s that you didn’t follow the proper process” — which was a pile of horseshit. I had the “CSI Miami” file of who said what to whom on what date. I brought it up with leadership and they said, “We’re not going to play Who Shot John.” Which is White House code for, “We have no interest in understanding what went wrong here because it’s going to be embarrassing to somebody higher up the totem pole from you. So we’re just going to sweep this whole thing under the rug and pretend it didn’t happen,” except for, “We’re absolutely not going to pretend it didn’t happen. We’re going to continue to hold this grudge and punish you for it, but we will not allow it to be discussed in any kind of open forum.”
It’s hard to imagine anything else that happened that pisses me off more than TED Talk-gate. Again, the Manhattan Project and the “Oppenheimer” movie are an excellent illustration of this: “We didn’t have to convict you of any crime. This isn’t a trial. We just don’t feel like going out of our way to help you anymore.” And that is the difference between a working Manhattan Project and nothing at all. The entire thing turns on that kind of attitude change. And we experienced that attitude change in 2016.
Emily:
And what about the agency roadshow?
Mikey:
The roadshow had this line of reasoning: “USDS is here and will last for 1,000 years. Its job is to take over the whole entire government and transform all of government technology everywhere, and boil the whole entire ocean. And in order to do that, we need to put ourselves in front of every single one of the 24 CFO Act agencies and get them excited, tell them they’re going to have a USDS agency team, and it’s going to solve all their problems.”
That was the attitude behind the agency roadshow, and it was not a good idea. There was never any chance that we were going to establish 24 copies of the USDS team while I was struggling mightily to establish just one at the White House. Everybody’s eyes were way too big for their stomach. And that’s something the government does over and over: find something that works once, and then immediately have the impulse to copy it 1,000 times. But USDS and breakout organizations just don’t work that way.
We did get a couple teams off the ground this way. There was a Department of Homeland Security agency team, a DOD agency team, a Health and Human Services agency team. But they often went haywire immediately. I was supposed to be following up with deputy secretaries from 24 departments, which is not possible for anybody to do. And deputy secretaries are not interested in having their time wasted. We set a lot of expectations that we absolutely had no business setting. That resentment had accumulated by the time I left in 2017.
Kathy:
I personally learned a lot from that time. At first I thought, “This is so great. We’ll learn so much.” But quickly discovered that regardless of what we learned, we’re being extractive of organizations and taking up their time if we don’t deliver anything. And they’ve gotten used to people coming in and saying, “Tell us all your problems, because we might fix it,” and running away.
Mikey:
If I had it to do over again, there would be no roadshow.
Emily:
Mikey, last question: How did USDS change you?
Mikey:
I got an education — one that cannot be purchased — working on complex systems problems. “Complex systems problems” is how I usually phrase it, because I don’t have good language for everything that we do. Everything that’s affecting people’s lives coming out of the public sector is some massive, sprawling, entangled, hybrid human-computer monster, that literally no one alive understands because it has been grown, accreted, and evolved over two generations or more. The original computerization of business functions happened in the ’60s, and those people are not with us anymore. So the reasons they made decisions they did are not known by any living person. And we’re three generations past that, and people are just doing their best to work on these foundations that, for the most part, they don’t understand.
How did that change me? Well, I understand that this problem is much deeper and harder than anyone wants it to be. Almost every time I read a LinkedIn post or listen to a talk where someone’s got the prescription for all these problems, I think, “You have understood 5% of it.”
People said the same stuff about USDS when we were new — they absolutely said the same thing about me. And that was accurate. We got started because we had some success on HealthCare.gov. But the computer problem is only a small piece of it. That’s the easier part to fix, in fact.
So that is a perspective that’s different post-government. I was only interested in fixing computer problems before. I would still prefer to fix computer problems, because it’s so much easier. But I know that it’s not usually what’s needed.
To be perfectly blunt, the level of dysfunction in a lot of these old bureaucracies is so much worse than even the most cynical person on my local Facebook group thinks. People have such low opinions of government agencies, frequently the ones that they depend on. And anyone can see that most of these policy ideas end up not working. Oftentimes, they end up making things worse. People believe that it’s because everyone is corrupt, or the politicians are lying. People think that Joe Biden personally makes decisions about their veterans benefits. They have no concept that this machine has 2 million people in it. There is no person, including the president, who has any clue what’s going on most of the time, any more than one ant can see the deforestation of the Amazon.
The levels of dysfunction that can take root in an organization with 250,000 employees that is not accountable, that isn’t measured against any market — that isn’t measured against any external standard at all — can be so much worse than people realize. That is sobering — and I think it’s connected to why I now live in the middle of the desert.
Kathy:
Thank you, Mikey.