The White House's Digital Fixit Man Talks Culture Change in Government

Mikey Dickerson, the former Google site reliability manager, speaks out about leading the U.S. Digital Service agency and how he and his team are working to usher in a culture change in the public sector.

by / October 20, 2014
Code for America Founder and Executive Director Jennifer Pahlka interviews Mikey Dickerson at the CfA 2014 Summit in September about efforts leading the nation's first U.S. Digital Service department. Jason Shueh

Government IT efforts can suffer from a host of ailments. There are swaths of delays, gluts of unforeseen costs, accountability deficits, and an absence of foresight in both procurement and implementation. Yet today, White House officials may have finally found a way to start solving some of the problems: They’ve “Googled it,” so to speak.

Meet Mikey Dickerson, Google’s former site reliability manager and now head of the U.S. Digital Service agency, a consulting IT department established in August 2014 and housed within the White House Office of Management and Budget (OMB). Dickerson, officially titled as USDS Deputy Federal CIO, was chosen for his instrumental rescue work on HealthCare.gov, the nation’s health insurance platform that suffered a catastrophic breakdown in 2013.

Since the USDS launch, Dickerson and his team of engineers have spent their time doing IT triage: meeting with agency officials, learning typical procedures and assessing obstacles. All the bureaucratic red tape and minutiae of big government has been placed squarely upon Dickerson’s shoulders. Yet despite his deep immersion, Dickerson remains strategically optimistic about the potential of his team. Recent successes include a growing lineup of private-sector talent to his team, a playbook of tips, and a soon-to-be-announced list of three government agencies the USDS will assist long term.

Known by his penchant for casual dress — suits a concession he refuses to make — and his honest off-the-cuff candor on getting things done, Dickerson said his hope is to embed some of the tech sector’s innovative and less conformist practices in government. In an interview with Government Technology this month, edited and abridged below, he outlined some of his thoughts about plans going forward.

Government Technology: What drove you to work with the White House after such a successful career in the private sector? Why did you want to change things?

Mikey Dickerson: Well ... to change things. I’ve worked the last nine years in a big tech company, obviously, and I was pretty happy doing that. And I could be happy doing that again sometime in the future. What convinced me to change and come to the public sector — at least for a little while here — as just the potential to affect people’s lives in a much more profound way than what I had been doing in the past. So a year ago, I got asked to come join this HealthCare.gov team and I worked on that.

We had a pretty positive impact, I think, on the functioning of that website, and it was a lot more meaningful than anything I went back to in Silicon Valley. So that’s a pretty attractive proposition, at least for a time, to work in government.

GT: What are the biggest differences from working in the White House and working in the private sector?

Dickerson: The government perceives risk differently ... government agencies tend to say and be proud of the stance that they’re very risk averse and can’t afford to let anything go wrong. But like most private companies — even when they’re just kidding with themselves — they see themselves as being risk takers ... government [officials] tend to feel that if they do things the way they’ve always been done, then that’s a low-risk approach. And they still feel this way even though the way they’ve done things in the past has led to bad outcomes again and again. Whereas [in the private sector], there was always a kind of default attitude of trying to do something new. In the end, both types of approaches have about the same amount of risks and they fail a decent amount of time. When you try something hard that’s always going to be the case, they just look at the risk and reward trade-off differently.

GT: What projects are you working on now?

Dickerson: So when we [the USDS team] got here, we spent a lot of time talking to a lot of different agencies and trying to identify what the potential is out there. Because there’s stuff we knew about from the news but thought maybe there were things we didn’t know about, that you have to dig a little deeper to uncover. We did that process and we boiled it down. We’re not a big group, we’re only a handful of people, so the work we’re doing and dedicating the majority of our time on condenses down to just three agency engagements — which we are putting a plan to announce in the near future — but nothing that we’re working on would surprise you. I feel like this announcement is going to be very anti-climactic because if you read the news, and if you had to guess what are the government’s biggest technology problems that you’d want fixed, then I’m pretty sure you’d come up with the same agency fixes as what we’re going to work on.

GT: Looking at culture, what changes could governments benefit from, for example, based on your experience at Google
— if Google could run government, how would it work?

Dickerson: I mentioned earlier that the way people [government and private sector] assess risk is different. That’s a pretty deep-seated cultural thing that we would like to change. The slides talk about it in the [U.S. Digital Services Playbook, a guide for government IT project managers].

Also, risk and reward look differently if you’re looking at it through the eyes of your customer — which in the case of government means the citizen or the public in general. What should make sense from government’s perspective, for example, is if you have a user-centered process, to think about what you’re going to change, what you’re going to do, and what you’re going to spend your technology money on to help customers. It makes sense to try things like saying, "Maybe we need to take this process, which involves filling out a bunch of paper forms, and try putting it online and make it easy to access from the phone."

If you think about things in that way, then it makes sense to spend some of your money and some of your resources to make things easier for the customer as opposed to easier for the agency.

GT: What’s the advice you give most often, is it on high-level innovation or more grounded in project management fundamentals?


Dickerson: It’s really basic stuff. A lot of our advice to implement projects it seems should be kinda common sense. Such as, we kind of hammer on the agile development methodologies — and that’s agile with a small ‘a.’ I don’t care a ton about anybody’s specific 10-step process to software development. I just care that we follow some kind of modern-ish process, by which I mean, we don’t commit a humungous amount of money at the very outset of the project, that we don’t lay out every single detailed specification at the outset of the project, and [we] set up a project plan that doesn’t require measurable outcomes until years into the future. "Don’t do things that way," is what our advice boils down to a lot.

There are absolutely big problems and difficult problems out there, including in the government space, that are going to take years to solve and might take, in the end, billions of dollars to solve. When you’re going to take on a project like that, you’ve got to somehow carve it into pieces that are measured in weeks or months, and then have a way to measure, at the end of a week or month span, whether we’re making any progress before we’ve committed all that money. Because there are cost overruns all the time, and before we get ourselves into that hole where it feels like the only way out is to spend more money, you need to measure your trajectory all along and make sure that if you have a team that isn’t working, or a contract arrangement that isn’t working, or something isn’t leading you to a successful outcome, that you need to have a way of detecting that and course-correcting before years and billions of dollars have gone by.

Like I said, it seems silly to even say out loud, because it seems like common sense, but that’s one of the biggest things we repeat a lot for how you should run a specific project.

GT: How much time is spent getting authorization versus acting on a plan?

Dickerson: Well, in our experience, because of what we do at the U.S. Digital Service, we spend a decent amount of time on this. For the engineers who are working for us directly, I would say a good amount of their time — and probably I would still say less than half — is spent trying to deal with procedural issues like that. It’s a larger fraction of our time at the U.S. Digital Service because that’s specifically our job to take on those structural and procedural issues that make it hard for others to do their best work. So I would say if we’re being successful, then my engineers are taking a decent amount of time taking on bureaucratic issues like that, and the agency’s own engineers or their contractors are spending less time on those issues. That’s a winning trade because there are a lot more of them than there are of us.

GT: In what areas do you think you’re making the most headway?


Dickerson: Well, being that we’re only in our second month, I would say that I’m really pleased with how much success we’re having with bringing quality top-tier talent from outside government to join government. We’re getting a ton of interest. We have not had any difficulty finding top-quality people that are interested in our jobs, so we are, consequently, spending a lot of our time — to get to your earlier point — taking on those bureaucratic obstacles that make it difficult to bring those top-tier people from the outside and put them in places where they’re going to make the most difference. We have put a lot of energy into that and we’ve had good success there so far. My own team is already about 10 people or so, which is good progress. Because even when I was at Google, and my job was to build new teams from scratch, it would be rare that I would be able to acquire 10 people in two months, and it’d be extremely rare that’d I’d be able to get somebody that’d be my first choice for the first tier of candidates.

GT: What type of working relationship do you have with 18F, the government’s technical team?

Dickerson: We have a different revenue model than they do. They are operating as if they were a supplier to the rest of government, so they’re like a shop you can hire. You can do an interagency agreement with them and transfer money to them in exchange for them implementing a project for you in much the same way that you would hire a contractor. But it’s a lot less work and it’s easier for the government to do. It’s not a whole procurement process, and 18F is setting a higher standard for how well those projects can be executed and at how cheap of a price.

And we’re not like that. You don’t have to come asking for the U.S. Digital Service and then execute a whole agreement where you effectively are buying our services as an agency. We are a smaller group and go where we feel there is the most need. So the way we’re working with 18F is that we hold them up as an example of what’s possible, what can be done with a high-performing technical team. We direct a lot of work their way when we’re talking with agencies. We may happen to be the first stop when they’re asking about building something — a new service, a new website, or whatever — but a lot of times our answer is to go to talk to 18F and see if they have capacity to take on that work for [agencies] because they’ll do it and they’ll do a good job. So that’s kind of the nature of how we work with 18F. We meet with them regularly and we go to same happy hours and interact around the same kinds of people.

GT: Long term, or even three to five years down the road, what’s your hope for USDS as it develops?

Dickerson: My greatest desire — if we were a complete and total success and I was thrilled with how everything went — then two or three years from now we’ll be much lower-profile than we are now, by which I mean most people will have forgotten that we exist because the government’s technology problems are not nearly top of mind for people as they are right now today. If we help solve a lot of those extremely painful interactions that people have with government technologies, we will kind of fade into the background. But, with one exception to that, I hope that we remain a strong brand name of our own, as in we become an organization that the country’s top technical talent wants to work for and they continue to seek out because we’ll still have a role for them. The government's always going to need really capable technical people for a variety of jobs. That need is not going to go away. So I hope that we remain exciting as a place to come work and do public service. But other than that, I hope nobody even remembers what we do.


Jason Shueh former staff writer

Jason Shueh is a former staff writer for Government Technology magazine.