Mina Hsiang led the U.S. Digital Service across three administrations and was on the original team that saved healthcare.gov. With Becky Hyde, she takes the urgency argument off the battlefield and into civilian services: Medicare payments running at 5% of GDP on federal mainframes, an infant formula shortage with no data on where to ship supply, half a million kids wrongly dropped from CHIP. She makes the case for iterating on policy and not only on software, using the FAFSA Simplification Act and Section 6103 as an example of a law whose letter blocked its own spirit, and closes on the two things she would have leaders do Monday: get clear on the objective, and get clear on how the team works together to reach it.

Becky Hyde:

Hello. Thank you. all right. Let me get settled here. all right. So hi, I'm Becky Hyde. I'm a delivery lead at Rise8, working with Space Force and True Anomaly. When it comes to defense, closing the delivery gap is crucial as Peter just showed us, but that urgency isn't just a defense problem, it's a government problem. And that delivery gap shows up everywhere, whether it's a veteran waiting six months for an appointment, a family struggling to pay their taxes, a senior citizen who can't get the benefits that they need. So my guest has spent her career addressing these gaps, administrator of the US Digital Service across three administrations, original team that saved healthcare.gov. Please welcome Mina Hsiang. Thank you.

Mina Hsiang:

Great to be here.

Becky Hyde:

All right. So we were just talking about urgency. It shows up everywhere in government services. Where have you felt this urgency on the civilian side where it's not necessarily an enemy that's causing the urgency, but it's Americans who need something from their government?

Mina Hsiang:

So we're so deeply reliant on government services across the arc of our lifetimes in ways that I think we often forget until suddenly it's in your face and you need to realize it. And there's lots of pieces of it that we just expect to work all the times, whether it's roads, schools, but there's also more intimate and obvious ways like health insurance. When you're elderly, you show up at the doctor and you expect if you're on Medicare that the doctor will treat you and the claims will be paid. And if they aren't, then that's a huge problem for you. Then you can't go and get the follow-up care that you need. There are thousands of ways that we all interact with the government and that the government addresses crises, big, small, and just day-to-day needs. Some other examples that are just like face-to-face type of interactions are things that we see and feel, right?

Our state public health agencies are constantly monitoring for whether or not there are outbreaks of critical diseases that we need to be aware of and watch out for. We are reacting to critical infant formula shortages. So in 2022, our team jumped into action when there was a critical infant formula shortage across the country. And the federal government said, "We don't know where there's a shortage, how much of a shortage. We're going to go to other countries and start buying infant formula to respond to this, but we don't know in great detail where to ship it." And so we did a ton of work with stores across the country to identify where the hotspots were, where there was not enough so that we wouldn't waste time where infants didn't have the nutrition that they needed. So there are an innumerable number of those that are sort of a face-to-face interaction, but there's also a much bigger macro question about urgency as well.

So I talked about I'm one person, I show up at the doctor and I need my bills to be paid. Medicare payments are 5% of GDP. That is all being run on software systems, mostly on mainframes that are run by the federal government. If those systems go down and 5% of GDP stop circulating, that's going to be massively destabilizing on the economy, right? That is a crucial and structural question for all of us. There's also this bigger question of trust. And I think this plays into what Peter was just saying, which is, do we have confidence and trust that our government is running our locality, state, country in the way that we expect? Do we have confidence in that? What are we fighting for? And if you constantly go to your government and expect them to be delivering something, you vote, there are laws passed and the way we implement those laws isn't actually how your average American or state citizen feels, then we begin to lose trust in that system and over time it makes us extremely susceptible to other types of societal decay and it makes it really.

If we aren't living in the social contract that we all believe that we're living in, how do we maintain the will to fight those critical fights in other contexts as well?

Becky Hyde:

Yeah, absolutely. Great connection from our civilian connection to government and how that connects back to our service members' connection to government and the urgency there. I was watching your talk with Brian at SHIP Summit and you talked about how a crisis or a fire can catalyze action and provide clarity for people knowing exactly what they need to do. And I think those moments are great when we have them, but listening to you talk about all the ways that government impacts our day-to-day life, we want daily services to also function well and for those to only function well when there's a crisis is a scary way to run government services. So what are your thoughts on how we can have the same clarity of focus and forward momentum and improvement in technology and our government services without everything having to be driven by a crisis?

Mina Hsiang:

So my macro though on that is leadership. The people in this context who can create urgency, who set up the incentives to determine what people focus on, how they spend their time is absolutely those who are in leadership positions. And there are a bunch of different ways that you can help your organization or help drive towards the type of tackling the problems that is necessary. So a crisis is helpful because it does a few things, right? It helps focus the mind. You all know that. So I have spent half my career in the private sector, half my career in the public sector, and I love those. Love is not the right word. There are moments where you walk into an agency and you say, "What is your top priority? What is the thing that you guys want to make sure to get right?" And they'll say, "Oh, we have clarity of priorities.

These are our 24 priorities." Okay, that's not a priority. So I think a crisis focuses the mind, right? It helps everyone say, "Oh shoot, there's a fire here. We need to all work on that fire right now." We have clarity about what good looks like, what fixed looks like. The fire needs to be out or on its way to out. So we all are focused on the same thing and we know what types of actions, even though they have some risk associated with them, are appropriate and what magnitude of risk you're willing to take on in order to tackle that. And so leaders can set up a lot of those parameters in how they lead their team, in how they focus the team with clarity on what is the problem, what does good look like, what steps are we willing to take and how are we going to operate as a team to get there?

Some of it can also just be in how you tell the story, right? It is a crisis for the White House when healthcare.gov stops working. It is also a crisis for millions of Americans who are trying to buy health insurance, although that crisis for them becomes more acute when they try and go see a doctor. But when a kid who's on CHIP, who is eligible for CHIP and should be on CHIP shows up at the doctor, but because there's something wrong in the system that has kicked that kid off CHIP because one of their parents is ineligible for Medicaid, that's not the right policy, but sometimes. So we discovered that that was happening in half a million cases, that kids in states were accidentally, the software was bundling households and it was kicking kids off CHIP even though they shouldn't have been. When that kid and his parents show up at the doctor and he needs care and he can't get it because he's been kicked off CHIP, that is a crisis for that family.

And if the agency that is in charge of repairing that problem can be more viscerally connected to that story of all of these crises for users, it becomes more clear what the solution needs to be and why it's so important that we need to solve that. So a lot of this is also about giving your team clarity and also exposure to the front lines and really understanding the problem more deeply.

Becky Hyde:

Yeah, I love that. Empathy, right? It doesn't have to be a national crisis, but it is a crisis every day for individuals and that's a very inspiring mission. Even if our organization is well funded and operating normally and we feel like we're at peace, there's still really important problems for us to solve. I love that. I'm hearing you talk about technology and policy, so I want to move forward a little bit into this concept that the system is more than the software. We're ultimately a technology group of people. We're here because we build software and we solve problems through technology, but you're talking about the chip example, and I'm sure there are other examples of where we have to iterate, not just on our software, but also on policy. I know this is something you've talked about before, so could you walk us through how you think about iteration in terms of policy?

Mina Hsiang:

Sure. I think this is very topical for all of the discussion around AI as well, because I think one of the things that AI gives us the power to do is iterate more quickly, but if we think. About iteration as only happening within the confines of the spec that we've been given, we'll never be able to achieve what the true mission is. This is a little bit like the spirit of the law as opposed to the letter of the law, right? There's often the case that there's, and literally of a law, there's the spirit of what are we trying to achieve? And then there's the letter of what does it actually say? And very frequently in all of the rooms that we're all in, the team is set up to say, "Okay, we have to perfectly execute the letter." And often that precludes us from being able to execute against the spirit.

And so the question is how can we as a government, broad quotes large group, but this group is a group of leaders. How do we lead our teams in a way that helps make it possible? So to get more concrete and to talk about some examples, maybe a big example. So some of you may recall, some of you may have college age children or have been in college at the time, but at the end of 2023, the FAFSA, the Federal Student Aid Form, there was a pretty substantial crisis where it was not available and this crisis really helped focus the mind. I gave up Christmas, New Year's and every single weekend for about eight months to work on resolving that and fixing the underlying systems. But if you want to take a few steps, excuse me, backwards, in the FAFSA Simplification Act, Congress had said sort of what led us to this crisis moment, right?

Congress had said, "Well, for the FAFSA, we have to know what each family is making, what the parents make in income. We also are the federal government. We run the IRS. So instead of asking the student to submit the IRS data to us, Department of Education, you should connect to the IRS and directly pull those records. It's the FAFSA Simplification Act. We're going to make it so much simpler. So you should directly connect to the IRS, pull in that data and then go forth." I think we all can understand the spirit of that law and what the goal is, but many of you probably are aware of Section 6103, which says that as soon as IRS data is involved, the complexity of your system, the security requirements, the amount of documentation overhead and the set of people who can actually look at the system or the data is significantly changed and it adds complexity by at least one order of magnitude.

So in addition, it means that every kid who wants to apply for FAFSA now has to go get their parents enrolled. Their parents have to set up their own account, link it with their kids because the kid can't pull the IRS data. The only person who's allowed to pull the IRS data is the parent in their account. You can see where I'm going, right? I don't need to get into every piece of this, but suddenly this Department of Education is stuck in a position where there are only three employees at the Department of Education who can look at stuff in the testing environment and it is illegal for them to create the types of test data that any of us would expect in order to launch a system like this because it would be falsifying tax data. The IRS is unwilling to send them the formats.

So all of a sudden you've created a ton of complexity. Now imagine if at the moment that you figure out, okay, this is going to be more complicated, we go back to Congress and we say, okay, we understood what you're trying to do here, but here is why that is going to cause us lots of downstream problems.

You can imagine a more collaborative and iterative process, just like in any company, if a CEO hands down any high performing company that operates well, if somebody senior hands down something that is not actually feasible, you go back and you figure out how to achieve what the overall objective is. There are truly opportunities for this. So when we were implementing the Medicare and CHIP Reauthorization Act, there were a lot of features that's like a real mouthful and it's a obscure and esoteric piece of legislation that basically was designed to rejigger how Medicare pays physicians so that they would be paid based on the quality of the care, not just on the quantity of their care. And once again, it's an admirable goal and it's an important goal, but there are a lot of complexities in how you should do that. And at the time, one of the essential features that we had was a really strong working relationship and leadership that understood that iteration was going to be key.

So when there were things that were unclear, there was a whole section that basically said, if you're the kind of doctor who does value-based care, lots and lots of words, so we built the tools according to the law and we tested it with doctors and they all said, "I have no idea if I'm that kind of doctor. I don't know what that means.That's a lot of legal jargon." And we were like, "Ah." And we were required to put that in there, but that allowed us to go back and say, "Okay, we need to define all these terms in a very different way. We need to give people ways of testing it. We need to look at their data and help them figure out whether or not they're that type of physician." And that's easy to do because this is in software and we already have all their data.

So it doesn't have to just be asking them these yes or no questions that they don't understand. So I think it's important to understand that everywhere, user experience isn't just iterating within a tiny loop where you say, "Oh, let's change the order of the words in this sentence to see if it becomes more understandable." It's really about saying, "Okay, what are we trying to achieve for people? What problems are we trying to solve? Are we solving them? And if not, how far up do we need to go to make iterative changes to get us to the goal that we're trying to achieve?"

Becky Hyde:

Yeah. It's like this combination of technology iteration and policy or process iteration and it has to happen in concert and technology just keeps moving faster and policy frankly has never kept up, but it's getting better. And I'm hearing you talk about collaborating cross-functionally between people who are experts in the technology, people who are experts in the policy and the end users of these systems. That sounds like a dream scenario where we could all get together and collaborate on this stuff and iterate when we find out we need to make a change. How do you get the right people together with the empowerment to have that dream scenario?

Mina Hsiang:

That's on all of the people in this room. I mean, I think somebody was talking earlier about incentives and every person who is here is here for the mission. Every person who you work with in a collaborating contractor and in an agency. I mean, it can feel more complicated obviously. You get down in the weeds and you're arguing about story points and everyone loses track of the fact, but no one here, if you all wanted to make the most money that you could possibly make with your technical expertise, you'd be managing systems for a bank or you would be somewhere else, but you are here because you want to solve true problems in people's lives, make this country stronger, preserve our way of life, support your fellow Americans. So taking a step back and continuing to create that feeling, that environment to remind people of the big picture constantly is the opportunity that each of you has as a leader to create the one team, one dream environment.

Does it matter if someone is a fed, if someone is contractor A, system integrator or contractor B, the person who's being bean counted or agency B, the IRS that's now trying to collaborate. It is our job as leaders to help the team understand the goals and then to understand how they're going to work together as a team, that it's not about a blame game. There's no incentives for. You have to create the world where there is no incentives for pointing fingers across at each other for what anyone is, who made it so that that deployment didn't work. Doesn't matter. Let's figure out how we're going to fix it.

Every program that I worked on as USDS administrator, setting up this dynamic appropriately was a humongous percentage of setting ourselves up for success. So all up and down the stack in every agency, there are extremely lossy communications. People don't have the same understanding of what the current state of play is. Then there's across contractors, what are we trying to achieve? Who's responsible for achieving that? The biggest thing that we always did was get everyone together and say, "We're all one team now. It doesn't matter. We are all trying to achieve this objective. Here's how we're going to get there." And having all of the leaders of all of those groups buy into that and set those set of shared expectations for their team is crucial for allowing everybody to work really well together. A, for example, is when we were implementing, I think somebody mentioned direct file earlier, but direct file was a...

You may have noticed that doing taxes in America is not the greatest experience. In our peer countries, you don't even have to do taxes often. The government sends you what they think you owe and you check it off or you say, "Oh, actually, here's a couple modifications," and you send it back. So direct file was the first step in that long process of maybe getting us as a country to that place. We had actually a pilot where it was pre-filled, but it was just free tax filing with the IRS. We launched it three years ago to. It was incredibly well received. The users loved it. It was incredibly hard to build because all of these agencies, as you know, are full of lawyers who are like.

The normal way to do this would have been to take all the questions on a tax form and put them on a webpage and then you would still be confused because you'd be like, "I don't even know what that question is asking me." And so what we did is we created these shared teams. It was user researchers who were with users and they were working closely with the lawyers because every time we wanted to change the language of what we were saying to somebody, we had to go back to the lawyers and say, "Is this clear enough?" And we would work together. So they were on the team. The leadership is on the team trying to get us to, "Okay, this introduces a tiny bit of risk that it could be misunderstood. What are we going to do about that? How is the agency going to make sure to manage that?

What are we going to say about how we're thinking about this?" So this iteration process of working with users, but also with lawyers, but also with leadership, but also with customer service, all of that has to be very tightly coupled because only in that way can we actually do something that accomplishes the overall objective. And all of this enables us to enact all of the policies that the US chooses to enact through the tax code. We want people to buy houses, have children, serve in the military, all of these things have tax credits, go to college, but all of that has to be understandable to people. They have to have an easy way recouping the benefits from those if we want that policy to work.

Becky Hyde:

Am I hearing you say we have to work more closely with lawyers?

Mina Hsiang:

We do.

Becky Hyde:

I joke, but I think -

Mina Hsiang:

But you can make them on your side, right? Exactly. And the leaders in the agency, which is some of you, have to make that true, right? Lawyers' jobs cannot just be risk avoidance. It has to be being part of the team and how do we move this forward?

Becky Hyde:

Yeah. And my background as a designer, that resonates really strongly with me. It was often the people who I felt like I was most opposed to in what we cared about were often my greatest partners in designing and delivering a great experience because I needed to understand their world as much as they needed to understand mine. So yeah, instead of pointing fingers at each other, if we work together, we create better stuff. Okay. We've just got a few minutes left. So looking forward, the theme of Prodacity is uncharted and we're in this moment where AI is sort of upending our assumptions about everything, including how government delivers. So what is your advice? If you were writing the playbook for the next four years for government services delivery, what would you tell this room to do differently starting Monday?

Mina Hsiang:

It's on all of you to write the playbook. I think from a how do you deliver perspective, the things to start doing Monday. Number one, be super clear about what your objective is. How do I know if it's solved? What outcome am I prioritizing for my team and what is my objective? Because AI allows us to iterate so fast that if not everyone is super clear on where you're trying to get to, you will iterate yourself into oblivion, into a different corner, and it will be lots of time wasted, harder to get back, investment. So number one, be extraordinarily clear with yourself, with your team about what your priorities are and what those objectives, how you evaluate those objectives, what problem are you trying to solve? And second, be extraordinarily clear with your team on how you want them to achieve those, how you want them to work together.

This is coming back to this one team, one dream. It's also a question of who do you recruit, right? You need to recruit in people who have the expertise in execution that you are trying to achieve. So be clear about what you're trying to achieve. Be clear with your team about how you want them to work together to achieve that.

And you guys are going to fix a lot of things.

Becky Hyde:

Thank you so much, Mina.

Mina Hsiang:

Are we taking questions?

Becky Hyde:

Any final thoughts? We weren't planning to take questions, but I'm sure we can. We've got a couple minutes left if anybody's got a question out there. Okay. We sprung it on them. That's okay. All right. More time for dispatch. Thank you so much, Mina.

Mina Hsiang:

Thank you.

‍