To Reimagine Unemployment Insurance Services, Start Small

Ryan Ko:

Hi everybody. My name is Ryan Ko. I use he, him pronouns. I’m the chief of staff at Code for America. Thank you everybody for joining us for the session today, To Reimagine Unemployment Insurance Services, Start Small. This session is one of many sessions that we’re doing and lead up to the Code for America summit on May 12th and 13th, which is a two-day virtual conference around designing and equitable government together. We’re quite excited to be having this, even though it is all virtual. We’d hoped to be in person, but of course, I’m excited that it’s been happening.

Ryan Ko:

You can get tickets at summit.codeforamerica.org. This session is authored by Nava Public Benefit Corporation and we’re very excited to be partnering with Nava to convene this route. Over the course of the pandemic, unemployment benefits that provided a critical lifeline for tens of millions of Americans who found themselves out of a job for months. Unfortunately, failures in service delivery, such as Florida $77 million broken unemployment website and 120-day wait times such as were found in Georgia that prevented people from receiving much needed aid.

Ryan Ko:

Today, we’re going to talk about how we should be investing in technical systems to lay the foundations for a more resilient safety net in the future. I’ll be your moderator today and as we go along, we’re going to be able to do questions through the Q and A. So in Zoom, you’ll notice that there’s both a chat box as well as a Q and A box. Please do put your questions in the Q and A box because it allows us to actually answer both live as well as in the chat. Oh, excuse me, in the text as well. Sometimes because chat, we’ll go and we’ll see people commenting. We may lose a question or two in the chat. So please, if you have an actual question, you’ll make sure to put it in the Q and A.

Ryan Ko:

Before I introduce our speakers, just a couple of quick notes. This is a Code for America event that operates under our code of conduct, which applies to all gatherings. We’re going to put the link to that in the chat and you should feel free please, to chat the hosts or myself directly to report any violations or context to safe space at codeforamerica.org. I mentioned using the Q and A button for questions and as always, we encourage the lively conversation as well on social media, especially on Twitter, by tagging Code for America tagging of a PBC and or using the CSA summit hashtag. And with that, I’m going to invite our panelists to turn on the video as I introduce and then let’s get going.

Ryan Ko:

Allison Hutchings is a Senior Technologist for Senator Ron Wyden, Chair of the Senate Finance Committee. Her portfolio includes AI and automated decision systems and government technology modernization, along with transportation and infrastructure. Before coming to the hill via the TechCongress fellowship, she worked as an aerospace engineer. Thank you, Allison for joining us.

Ryan Ko:

Alyssa Levitz leads the unemployment insurance program with the U.S. Digital Response. She previously, held product manager roles with Rover.com, Facebook, and Microsoft. We’re so happy to have you here Alyssa. Thanks for joining us. And of course, Michelle. Michelle Thong is a Senior Product Manager at Nava Public Benefit Corporation. She previously served as digital services lead for the City San Jose and was co-founder as Code for San Jose at Code for America Brigade. We’re so excited to have you here. Thank you, Michelle.

Ryan Ko:

Okay, let’s start us off by setting the stage here. So a year ago, obviously, an unprecedented wave of layoffs and furloughs left millions of Americans unemployed as the COVID-19 pandemic overtook the U.S., and since then, approximately 40 million people in the United States have received unemployment insurance benefits, yet many more had actually applied. Alyssa, as the unemployment insurance lead at U.S. Digital Response, a nonprofit that pairs volunteer technologists with states that need and has partnered with nine states in counting to help them with UI. Can you maybe just give us the TLDR that too long, didn’t read, on why most states unemployment insurance systems were so overwhelmed so quickly?

Alyssa Levitz:

Absolutely. So unemployment insurance programs were basically built to scale with the hiring and firing of individuals. It was taken as just kind of a given that’s over half of applications would need to be touched by a human at some point. And when the economy is just like going at a standard kind of constant pace, you can plan for that. But when the economy doesn’t go as planned as happened last year with 20X increase in claims over the course of one week, you just can’t scale that quickly.

Alyssa Levitz:

So when you combine that massive increase in applications with a decade or more of underfunding with the need to additionally implement totally new federal programs, including a Pandemic Unemployment Assistance, which was for a population of individuals that had never, before been able to access unemployment benefits, you get this perfect storm of a system, unable to respond quickly to the needs of the population. So they reacted as best as they could and with that, we are still trying to come back from that.

Ryan Ko:

Yeah. I am curious as well. Allison, in your role, what did you see or what was that conversation like amongst governments and what did they really need in order to deliver the benefits to eligible residents, at the speed of that need?

Allison Hutchings:

So I think one thing that’s interesting from the federal policy perspective, it is very much like steering a very giant aircraft carrier or something that is not necessarily a normal thing speed of need there. Unfortunately, and we’ll probably talk about a lot this today, but there were things like Alyssa talked about, it’d been a decade of underfunding. We’d already outsourced our mission. So we were already working from a place that was very hard to dig out of.

Allison Hutchings:

The fastest thing we can do is provide money, but like we’ve seen, and I’m sure the people who already worked on these systems saw, providing lots more money like during the great recession doesn’t solve the problem right away. There’s lots of other things we need to do. So I’m going to flip that a little bit and say with a goal of actually building these equitable systems to serve everyone from a policy perspective with Senator Wyden staff, we’re just trying to make sure we can set up guardrails for the future and get out of the way of the people on the ground that are fixing it. I could talk a little bit about why Senate Finance Committee has a role here or we can continue to talk about, would that be helpful, Ryan?

Ryan Ko:

Yeah, let’s go there.

Allison Hutchings:

Okay, great. So Senator Wyden is chair of the relevant committee in the Senate, the Senate Finance Committee. So it has jurisdiction over unemployment insurance. So his long-term goal is to make sure this is a successful system in the future. And he wants to put the guardrails in place to prevent this from happening again. And like I said, this somewhat happened before 10, 15 years ago. We tried to modernize some states, didn’t do the best job before and this is coming back to the mission question. A lot of times the mission seemed to end up being rooting out like waste, fraud, and abuse and things like that rather than actually building equitable systems.

Allison Hutchings:

So over the last year, while all states were responding and while these emergency programs were stood up or these volunteer programs were set up, I worked with the finance committee staff and Senator Wyden pushed us to introduce a bill that got introduced earlier this year with 20 other senators to modernize the unemployment insurance system. So the goal is to prioritize user experience, accessibility, and equity within the bill. Requires using modern software development practices without being prescriptive, but just trying to push towards that and not these huge procurements that we see often.

Allison Hutchings:

Then perhaps what people get most excited to talk about is we would establish a new department of labor digital services team to provide capacity within the federal government to help with this process. We can talk about that more, but one note, just fixing the technology is a huge piece of the puzzle, but fixing technology alone isn’t going to repair a lot of these things. So Senator Wyden introduced a separate bill last week to fix these problems too, that we need to provide adequate financial security across the board for everybody across the country. And not leave workers out like Alyssa mentioned, but I’ll stop there.

Alyssa Levitz:

To build on what Allison said, technology is part of the solution, but it’s not… You don’t want to think about it as like a solution in and of itself. We hear a lot of conversations about how, oh, these systems are 40 years old. They’re still written in COBOL. We can’t even find COBOL engineers and well, that is a problem. The real problem is that these mainframe systems that happened to be written in COBOL are kind of single-threaded. They can do one thing at a time. So a state can’t expand its hours of customer service on the phone because it’s mainframe needs to be switched to focus on processing claims.

Alyssa Levitz:

That is the problem to solve. If all we are doing is switching from one giant COBOL monolith to a giant Java monolith or Python monolith. We’re just changing the landscape of the problem, but we’re not actually solving the problem. So to echo what Allison was saying about that user experience, we need to think about what are the problems that claimants, employers, and agency staff are encountering and administering these programs equitably and in a timely fashion. And focus on solving those as independent problems, rather than just that one giant procurement where six or seven years later, you may or may not have a changed system, but that change system is not going to solve the problems.

Ryan Ko:

Yeah. I really appreciate that point, Alyssa, especially around putting that actual person in that population first and focusing on that. Obviously, we know technology is an incredibly powerful driver and force, but it’s all technology challenges for technology stake. Sometimes is focusing on missing the forest for the trees. I wanted to switch over to Michelle for a second here.

Ryan Ko:

So Michelle, you were the co-founder of Code for San Jose and currently a Product Manager at Nava Public Benefit Corporation. And you and the Nava team actually worked with the State of California on their unemployment insurance program when that service was experiencing unprecedented demand due to the pandemic. What did you learn from that experience and how has that informed your view on, for example, common UI, user interface components that could help make systems easy for people to use and civil servants to maintain and operate as things change?

Michelle Thong:

Thanks for the question, Ryan. So absolutely, our team at Nava, we were privileged to have the opportunity a little over a year ago to work side by side with public servants from the California Employment Development Department and the California Office of Digital Innovation, just as we were seeing huge spikes in weekly unemployment claims from the tens of thousands to over a million a week.

Michelle Thong:

And while this was a unique unprecedented crisis with extraordinary circumstances that we’ve all been living through, it was also a situation where the approach of solving big problems with small flexible tools proved itself to be the right approach for a crisis. And also the right approach now is we learn lessons from that past year and think about how to move forward towards this big vision, a big, bold vision of a resilient and equitable safety net.

Michelle Thong:

There’s three key reasons that I want to highlight about why starting small and working towards that bold vision is the right approach. So as Allison and Alyssa alluded to, one of the problems that made things worse in the crisis was that many state unemployment systems are currently these monolithic systems that where each component all of the different functions, whether it’s applying, certifying, payment processing, fraud detection, they’re all very tightly coupled. That means that as Alyssa was saying, you can only work on one problem at a time and every change that you make has a risk of bringing down the rest of the system.

Michelle Thong:

So instead when you take a modular approach where you are picking specific problems to solve and specific functions or user needs or staff needs in the system, you can break those off into a modular component that can be largely independent from the rest of the system. And then loosely coupled through what we call an API-first architecture, where you can have different components and you can improve one part of the system while continuing to use the existing legacy system and then slowly add on new components.

Michelle Thong:

So that’s really valuable both in crisis situations and moving forward for a few reasons. First, it enables you to learn as you go along. You can get something working, put it in front of real users, put it in front of staff, and see if you’re moving in the right direction towards that vision. And so that reduces your risk. So in California, we actually started by just building a plain language guide to navigating the unemployment application process. That was just the very entry point of the process. It didn’t affect the processes farther down, but that enabled us to learn a lot about how UI, how people were navigating the process and where they were confused and where they needed the most help.

Michelle Thong:

The second big reason why starting small and working on modular components is really valuable is you can work in parallel paths as Alyssa was saying. Certainly, during the crisis and even going forward, the reality of unemployment programs is that there’s a balance between the day-to-day operational pressures of processing claims and getting payments out and investing in a new system. And when we were working with California, especially a year ago, their absolute focus was implementing policy changes and getting payments to people out. It wasn’t the time to be completely redesigning the service experience.

Michelle Thong:

But we were able to contribute as were other teams by building standalone apps that were more loosely coupled with the rest of the system. So standing up a certification process in a matter of weeks, on cloud infrastructure that still had some integrations with their data and client authentication. We’re building on that approach. We see that other state agencies are taking that approach and that gives them more flexibility to work on multiple things in parallel.

Michelle Thong:

Then the third big reason is really about the people in the process. Shifting from the approach that states have typically taken, to managing their technical systems for service delivery from big procurements to managing multiple modular projects and taking on a human-centered design, perhaps shifting from mainframes to the cloud, that’s a huge change in process. So if you start small and then gradually build on it, that gives the people in the process, the time to adapt as well.

Allison Hutchings:

And really quickly, I’ll just jump in. One thing about the bill that Senator Wyden put out is it really… He wanted to listen to the people who had done this and what we learned from responses to COVID, but also lessons learned from some of the great digital service teams within the federal government, like USDS and 18F and how they build products to solve problems for people. So the bill specifically asks these modular set of capabilities that would allow us to build different pieces. It asks for piloting. This is not something that you roll out everywhere. It start very small and then slowly build more couple other pieces.

Allison Hutchings:

So just explicitly requiring this user experience type work, where you’re in consultation with a whole host of different people, including as everybody here knows, but the end impacted claimants in this case, but also the workers making sure that their experience is good. Then just a lot of other pieces, I’m just looking at the summary of the bill to make sure I capture it. But accessibility requirements that came from talking to some of the experts in UI reform space to make sure this is just accessible for all sorts of different people, simply it has to work on cell phone. So a lot of people only connect to the internet on a cell phone.

Alyssa Levitz:

Yeah. And to build on both Allison and Michelle here, as we are seeing these modular pieces and we can talk about them within one state, one of the things is that… So there are 53 unemployment insurance systems, the 50 states, do you see Puerto Rico in the U.S. Virgin Islands? And Guam also had a poor program and they’re looking into building their own unemployment insurance program now. But right now the U.S. Department of Labor is paying for the same things, 53 times. And while the specifics of each state are different, there are a lot of state laws that indicate how to do certain pieces or what the variable is here and there.

Alyssa Levitz:

The general functionality is the same. And with these pilot projects that the bill is talking about and supporting, we’re able… We could figure out the right ways to try to build this once with 53 variants rather than from scratch, 53 times. And in so doing, we can use that money more efficiently and have the time to focus on things like a better user experience, like equity and various ways of accessing this portal, so that people can get these benefits more effectively, efficiently, and equitably.

Ryan Ko:

I’m seeing a couple of questions in the Q and A and I think these are more for Allison. A couple of folks have pointed out there are already teams that exist that do digital services. And at the federal level, such as AKF within the GSA or the U.S. Digital Service. The bill that you’re describing provides for a digital service for the Department of Labor, what would that be similar and different? And they’re also just a couple of folks asking for a pointer perhaps to the bill number itself, so that they can actually look it up and read it.

Allison Hutchings:

We can drop the bill in the chat in a minute, but actually, I should know the bill number and took too many things. So the discussion of why the Department of Labor have a specific digital service team versus just let’s say USDS role in was actually… That’s a great question. We talked about that a lot. We do have such expertise at 18F or USDS to do something like this. We want every federal government agency to have this type of expertise. We don’t want it to be something that they have to ask for help from somebody.

Allison Hutchings:

If you have an executive within your agency that maybe not interested sometimes there’s trouble bringing in that expertise. But also having experts on the digital service team that truly are integrated in department of labor’s mission and don’t have to get up to speed. All right. They’re not replacements at all, but having these expertise at all agencies would be a future we would be excited to have. And so that was why we made that decision. I’ll go grab the bill and put it in chat.

Michelle Thong:

Yeah, I would love to chime in and double down on the value of having designers and user researchers and other digital capabilities in house, who can really work hand in hand with the policy and program implementation teams on UI. One of the teams that helped make our work in EDD successful was they already had an internal web content and usability group, that was both really understood the policy and then also really understood the needs of users in California and were able to advocate for content improvements that would help users in their experience.

Michelle Thong:

I think what’s really exciting about the bill that Allison was talking about is how it talks about prioritizing the user experience. Just to shed a little bit more light on that, I think a lot of the pain that we saw people experience in the last year trying to get unemployment benefits, in most cases, it was not because the state was trying to make that experience hard, but because the policies and operational processes were complicated and were implemented in a policy-driven way rather than necessarily using the claim and experience as an organizing principle.

Michelle Thong:

What we’re seeing is that when you use the claim and experience as an organizing principle for how you provide information that is customized to what that claim. It needs to know at each step and you help make it a clear reassuring experience all the way through, that is going to help drive that simplification so that people aren’t feeling as overwhelmed, navigating the process. And also has operational benefits because it can reduce the customer support load. By having the people who have relatively straightforward cases, if we can help more of those people through the system quickly, that frees up more bandwidth to deal with the cases that are more complex.

Ryan Ko:

One thing I’ll actually time in and add here as well. I think in many ways, USDS and 18F are actually obviously early and innovating, but the case could and could be made by all of our panelists here that all levels of government, the ways of working espoused by digital services teams should be funded and exist. It’s really quite interesting to see how that grows.

Ryan Ko:

Obviously, some of the federal agencies, well, not federal agencies, like 18F and USDS has slightly limited scope. They can’t always reach everywhere they can. So the more cases we see where like a state, or a county, or even a city, or a department like Department of Labor can carve out funding and resourcing for digital services team. But more than I think we’ll see the practices that we want to see grow.

Ryan Ko:

I’m seeing this a lot as I curate the question and answer here. There’s a lot of people pointing out things like modernization should not be just for modernization’s sake and it could be focused on actually operationalizing policy. These are all just absolutely great points. How to actually paint parts and minds about the way of working so that it’s not a traditional outsourcing approach as Allison mentioned, but actually more of working with and building in a modular way as Michelle and Alyssa were mentioning.

Ryan Ko:

I’m seeing a lot of these themes together and I think it brought together the Code for America view is, the more we can help cultivate and grow this digital services mindset and being a really human-centered data-driven iterative in our approaches, really mindful of equity and empowering public servants to drive their mission in support of vendors or in support from vendors, not purely outsourcing the mission as Allison mentioned. All of that should add it together.

Ryan Ko:

So we in general see the rooming of digital services as really exciting, exciting trend. But that’s probably too much talking from the moderator. So I’m going to actually direct the next question at Allison. Allison you’re Congress alumni and you currently serve as one of two senior technologists for Senator Wyden’s office and you obviously had a major role in crafting the UI Modernization bill. For those who haven’t been following the chat, this is a bill that is to upgrade the technical underpinnings of the country’s unemployment system. And it shows in the language of the bill modular, digital service, islet state, starting small.

Ryan Ko:

Based on this, maybe can you pull back the onion on this a little bit and talk about how you and other policy makers were able to bring and bridge that gap between the traditional policy state and more of the technology space? And how were you able to actually successfully advocate for these practices?

Allison Hutchings:

Yeah. And I think I’m going to not give you the most satisfying answer and I’m going to say we got extremely lucky in certain ways to have this terrible event happening, but we had the chair of the Senate Finance Committee being Senator Ron Wyden, who deeply cares about technology. He long has hired staff with expertise. He wants it to be used for good and not be used in harmful ways. And he will poke his nose in all sorts of things. So that was that. He’s the only member I know of that has two technologists on staff. So TechCongress is pushing to bring more people to the Hill for that purpose, but we’re growing.

Allison Hutchings:

But that was paired with him being chair of the relevant committee. And so he had the expertise on the Hill in terms of the UI policy side right there and we were living through the code of pandemic. So it was obviously systems were not serving the public. We didn’t have to make the case that we needed to do something. The lift was already there and people were asking for help. So it wasn’t convincing people, there was an issue and we should do something. It’s what do we do? We need to do something now. And so that’s just a very different environment.

Allison Hutchings:

On the modern human-centered practices side in particular, that’s somewhat of the same thing. The UI reform advocates who have been pushing for a lot of the policy reforms, they know the system is not built to serve people in many states. And you can look at these really bad examples of Florida seemingly designing a system to purposefully not serve people.

Allison Hutchings:

So starting with the conversation about user experience and human-centered was just not a big lift from the federal policy space at the time that I came in to the conversation. But I think it’s a good example. Now is the example for other spaces. We need to use this as yes, it took this emergency to have this conversation, but why are we not having the same conversation? Not only for every other benefit system, but beyond that in government services. It was not the sort of answer your question.

Alyssa Levitz:

To build on that and also to answer the couple of the questions in the Q and A, I think what Allison was saying about how UI is a starting point. It was one of the first systems to really go down, in a sense with the pandemic. And there’s been a lot of focus on the ways in which it is broken. It has not gotten the resources it needs and the ways that it’s not helping people as effectively as it could, but there are a lot of benefit programs that are not as effective or as equitable as they could be. With these pilot programs through the federal government, at the UI level, we can also think about how do we connect this to other benefit programs that many of the same people might be participating in.

Alyssa Levitz:

So almost every benefit program, you’re going to need to verify identity. So what does it look like to have a integrated identity verification solution, so that every single time someone is applying or renewing benefits, they don’t have to reupload all of their documents, find all of those documents in the first place, go see someone in person? How can we do that? Have that once. And again, that’s thinking about the user experience and the user problems that we’re trying to solve and UI is one way that we’re able to enter and begin that conversation, but there’s definitely a desire to think about that holistically as part of the social safety net.

Michelle Thong:

Absolutely. I think the good news is there’s a lot of prior work that we can build on, even as we look at UI specifically. So Code for America, and Nava, and other organizations have partnered together looking at integrated benefits initiative. Nava specifically, worked in Vermont where document uploading, the moment in the journey where someone who’s applying for benefits needs to provide evidence of the information that they provided. In a lot of states, that’s still a manual mailing process that can take days.

Michelle Thong:

And so what Nava did there was, we built a document uploader, which is one of these modular components that could be used across states and systems, where we were able to increase the percentage of claimants who submitted their benefits within 24 hours from 11% to 55%, by enabling them to easily submit their documents online over their phone. That’s a solution that is now working at scale across multiple benefit programs in Vermont and we’re seeing the need for it in lots of other benefit programs.

Michelle Thong:

So I think when we look at the other types of needs that Alyssa mentioned like eligibility, ID verification, claim status, certification forms, there are so many opportunities for us to learn and reuse. Even if you’re not exactly reusing the exact code base, there’s a lot of lessons and patterns that we can be sharing across states and programs.

Allison Hutchings:

From the federal perspective, from policy side, whole of things I think about, we need to help get out of the way. We need to help make sure that there’s not regulations that are in place that are making it hard for people to share things that would be helpful across the states and things like that. We want to encourage collaboration. But one thing Alyssa said and then one other, the identity side, there’s such a big federal policy question on who should be doing this, what guardrails need to be in place. A lot of these identity solutions are moving towards using biometrics like face recognition and things. One, are we making sure that it actually works on everybody, but two are we normalizing something that there’s very real policy discussions about in other contexts. And we can’t put blinders on and just ignore that.

Allison Hutchings:

So there’s such big picture questions that the UI space alone can’t answer. That’s not a UI answer, it’s a bigger answer. And then the other side is just, there’s so much data that would be useful to share, to help with a lot of this. How do we do that in a way that protects everybody? I think I saw some question in the Q and A about that and it has to be part of this. There have to be guardrails once again. If the states are going to share data for UI, are they guaranteed it’s not going to be used for enforcement of state immigration?

Alyssa Levitz:

I think identity is one of the ones where there’s a lot of urgency around it now and so there’s a question of, should states wait for the federal government to figure it out or how do they work with this on their own? Everyone has their own unique situation right now, but it’s certainly something that the federal government seems to be looking at. But as states are looking to solve these problems in the short term, if they need to do things like these systems can’t just wait for the next big bang procurement or big bang solution coming from federal government or a vendor, whoever else, some other state.

Alyssa Levitz:

As states continue to make modular changes, the key is that they should be modular. They should be open source and the state should own the code. The state should be able to change the code. They need to have these contracts that enable them to do this. And they should be thinking to the degree that they have the capacity to do so and I know States are still super overwhelmed right now. They’ve been understaffed constantly and definitely having difficulty now, but thinking about how might someone else use that system? And how could we build this system, especially if it’s open source and focused on a user problem? That’s something that another state could come to and be like, “Oh, hey, we have that problem. Yeah, we have a different backend. We’re going to have to need to connect it differently.”

Alyssa Levitz:

But to what Michelle was talking about, about an API-based infrastructure, that API can connect to a different backend and you have that module there. And so I think that that is an important part of this modernization system. It’s not modernization for modernization’s sake, we have problems that we’re trying to solve. There are short term solutions that can plug into this longer term vision over time.

Ryan Ko:

Alyssa, on that note, I’d love to actually heat something up, which is, we actually get asked at Code for America a lot. The audience today does have both technical and non-technical folks from all across all levels of government. And so we get asked a lot about this thing that you talked about during a panel or an event or a white paper that you published, but how do we bring this life? What are the actions that we can take and to bring them back to our departments and agencies based on what they learned? Let’s open it up to all three of you. What are things that folks, both technical and non-technical consider bringing back?

Michelle Thong:

This is a great question. I think as I reflect on where we started, we’re saying start small and then we’re talking about really big ideas about how we want to see this bold vision come to life in a whole system. But yeah, let’s start small with, what are things that folks on the call today can go back to their teams and start doing? I think one that’s my favorite is to just get extremely concrete in your knowledge and understanding about what your user’s experience and your staff experience is in navigating the unemployment process and where the pain points are. That can be very easy. It can be just creating a test account and going through all the screens that you need to go through in order to certify or putting the questions in front of someone and seeing what is confusing to them.

Michelle Thong:

We had great experiences in California being able to bring along policy and communications folks to listen to real users. And for some of those folks who were very well steeped in the policy, it was still their first time listening to one of the beneficiaries of the programs that they run, actually talking about what their experience was navigating it. That is something that can be made into a habit of doing usability testing or ride alongs to help your team and your leadership understand what that customer experience is like. And then secondly, following on that, once you find a problem or a pain point, start thinking about it iteratively and try something to improve it.

Michelle Thong: We’ve noticed sometimes, there’s a tendency when people think about their systems as being monolithic, as just saying, this is something that’s a problem. We’ll fix it next time in five years when we have the new system. What we say is, figure out if there’s something that you could do. Is there a sentence that you could add to your website or to the application form or the certification form right now that might make it easier? And decide what your goal is for improving that problem and then measure it and see if what you did made a difference. Also, build that organizational muscle and habit of picking a goal and measuring it, and then seeing how you’re doing and iterating.

Michelle Thong:

We found that worked really well. With our new certification form, we stood up that we all had agreed on goal where we could tie together the program objectives with the user experience. And then seeing progress on that really motivated people and made it feel more real to try out new things that maybe we hadn’t tried before in our communication strategy, so that we could move the needle on those metrics. So those are the two, get experience with the user’s experience and then pick a problem and start iterating on it.

Allison Hutchings:

I’ll jump in and advocate for capacity. I’m speaking as I used to build things, but not in the way that my colleagues on the call do. So I’ll count myself as non-technical in this type of product and advocate for at least having some expertise in your organization. So many things are going to go so much more smoothly with things like procurement and things like that, even if you’re not building your own systems. I don’t know when I was supposed to do this, but I’ll jump in and advocate for everybody on the call. Senator Wyden and Senator Murray are going to be reintroducing a bill in that matter of weeks and I’ve seen a couple of comments about that in the Q and A.

Allison Hutchings:

But why did the UI bill focus on the federal government? Well, we had to start somewhere and we are starting with Department of Labor. But we realized that state and local governments absolutely need support in standing up these types of teams and building this capacity. So we’re introducing a bill to provide startup funding for digital service teams for state and local governments, along with expertise from these teams we talk about, like USDS and 18F to help them stand it up, to help with all of what we’re talking about here. So stay tuned and I’m sure you’ll hear about this more later, but help us advocate to get something like that passed because it will help on a lot of these broader questions, just to have capacity within these teams.

Alyssa Levitz:

I guess I’ll use this as an opportunity to do a little bit of a plug. If you are a government agency or are in a government agency and are interested in trying to understand, what would it mean to do modular procurement or how would I think about solving this user’s problem in isolation, as opposed to going down the rabbit hole, but it’s connected to this and this and this and this, but being able to really focus? U.S. Digital Response is absolutely happy to help. We are pro bono, so we’re free to the government and we are here to help in whatever way makes the most sense. We have user researchers, we have engineers, we have designers, we have systems architect experts who can help figure this out. You’re not alone in this, there are also great places like Nava that can also help with this.

Alyssa Levitz:

In doing this, governments don’t need to be alone in this. They don’t need to reinvent the wheel. There are a lot of resources out there to understand product design thinking and user-centered development. And to the degree that you are able to develop those muscles in-house, that is amazing and really important. So tying back to what Allison was saying at the very beginning, you don’t want to outsource your mission. You want to think about what it is that you as the agency, what is your goal and how can you regain control of that to best serve your constituents? And part of that is understanding how to make changes that will improve the experience for people a little bit at a time.

Allison Hutchings:

Alyssa, actually, can you perhaps say more about the procurement piece? There are many folks who, for example, have multi-year contracts with the large technology providers and in many ways, are working with them to react and change and move quicker. So when you say, don’t outsource your mission, can you tell us a little bit more about the ways that you might operate with them in a more modular product management way?

Alyssa Levitz:

Absolutely. And first I will do a shout out to 18F’s amazing procurement guides. There are, I think two or three blog posts and then like a 45-page primer on this that have been the source of a lot of my expertise on this. But the idea is that you have someone in-house who’s playing that product owner, product manager, customer experience, expert role that is going to help understand all of the pieces of the system and prioritize what is most important. So what is the problem, that if we figure out how to solve, is going to make a really big difference in the speed with which we are able to get claimants their benefits or the reduction in the amount of staff hours it requires to process a certain percent of applications?

Alyssa Levitz:

Someone who is thinking really hard about all of that and has the individual expertise within that environment. With each of those problems, a small procurement process is established. So it’s not a 200-page RFP contract. It’s a 20-page contract and 18F also has examples of these contracts to focus on one thing at a time. And when you have that product owner in-house, it’s their job to manage multiple of these things at once and understand how they fit together and to continue to make progress.

Alyssa Levitz:

The thing with these contracts, you are paying for the developer’s time. You’re not paying for a certain requirements checklist outcome. So if things aren’t going well with a contractor and in this different modular contracting mode, it’s just like, okay, well, we’re not going to contract any more hours with you. And so that person in-house is collaborating with the procurement officer, as well as all of the agency staff to help prioritize and fix problems independently.

Michelle Thong:

Nava’s on the other side of the procurement process and we absolutely agree with the value of bringing that capability for ownership, not just of the technology, but also the vision and the strategy in-house. Certainly, companies like Nava and others like us, we’ve worked best when we have partners on the government side who have that strong vision about what they’re trying to achieve and really care about taking human-centered approaches. So it’s really exciting to be having this conversation from different parts of the ecosystem.

Ryan Ko:

Thank you so much for elaborating and sharing those examples. I’m keeping an eye on time which is less than 10 minutes now and just really appreciate this conversation so far. I’d love to shift a little bit with more Q and A, so maybe let’s go through most of the rest of the hour on the Q and A side. And the first question here from Sarah is, “Do you have any tips on moving toward government service design? I’m currently a UX designer with four years of experience with the job seeing so few and competitive, such as 18F or USDS.” And looking at this panel, and we all had a bit of a different experience on how we perhaps broke into or started working at the intersection of government and technology. Perhaps, I could tap on each of you to share your experience of how you started to get involved. Maybe, go ahead Michelle.

Michelle Thong:

First of all, I’ll say Nava is hiring. And I would say with respect to my own path, I started getting into civic tech in 2013 and it was by going to Code for America Brigade meeting OpenOakland. So there are volunteer opportunities you could volunteer with USDS, but I will let Alyssa speak to that. There’s a growing number of opportunities aside from 18F and USDS. As we’ve been talking about there are more and more state digital teams, as well as local digital teams in local governments and firms like Nava.

Michelle Thong:

But also, I’m a big advocate, especially if you already have really solid experience in your discipline, whether that’s design, or product management, or it’s taking a job in government that may not be labeled as a civic tech digital service. But working for the Parks and Recs Department, or working for the city manager’s office, or working in a UI Department, there’s a huge amount of influence and positive impact that you can have by bringing your skills and approach into their program or policy discipline. So there’s a lot more opportunities where you can widen that net and just look at, you will learn so much just by being in a government agency and seeing how they operate.

Alyssa Levitz:

I think of the panel, I am the newest to civic tech and I will say that I actually came in as a USDR volunteer. That was my first foray into civic tech. I had been in private tech for the prior eight or nine years and was actually on unemployment for a lot of 2020. And as I was trying to get things together, a friend introduced me to USDR and I was working as a volunteer on an elections-related project. I was just like, “Oh, this exists. This is what I was looking for in my life.” I didn’t even realize like civic tech was a thing, honestly, but as soon as I found it, I was like, “This feels like my people. This feels like something that I could do to help impact real people’s lives in a meaningful way.”

Alyssa Levitz:

And then when USDR posted job hiring for the unemployment insurance team lead, it felt like just a super perfect fit because I had been on unemployment. I was very lucky, I was not reliant on the UI benefits to make sure that I was still going to eat every week and that I wasn’t going to lose my house, and wanting to contribute in a way that helped people and made the system better for others. So with that, it was really serendipitous, but there are always tons of opportunities with USDR.

Alyssa Levitz:

I will also echo what Michelle was saying with joining government wherever feels like a good fit. I personally have not been in government, but I was talking to a friend the other day and he was like, “Just take the first position within government when you can. Good work will be recognized and you will continue to have an influence there.” And I think that’s really powerful.

Allison Hutchings:

I think I’m even the weird one still. I do technology policy for the U.S. Congress, thus, I still am not even working at the intersection of civic tech. I’m just trying to help make policies that clear the way for everybody on this call to do good work. I came to The Hill through the TechCongress program and so for folks that are interested in that policy purely, I mean, it’s not purely, but so policy-focused role within Congress, it is a phenomenal program. It is constantly growing. I will flag a new program that just got set up this year that there’s actually a congressional digital service that started last year in the house side that just got lots of great shout outs at the house modernization committee.

Allison Hutchings:

So for folks interested in that, pay attention to see what the future of that is because Congress also needs help with its own technology. I do believe that these opportunities are going to grow. I think the stories are being told a lot better than they were before, so I have a lot of optimism that there will be more available.

Ryan Ko:

And of course, a couple of additional ones that I’ll mention are the Aspen Tech Policy Hub Fellowship, as well as the Presidential Innovation Fellows programs. Of course, Code for America has a 25,000 members strong-volunteer network across 85 plus brigades across the country and of course we’re hiring as well. But I’d love to just close with perhaps one last question. What are words of encouragement that you’d offer the audience, many of whom are still working through different crises?

Allison Hutchings:

We have competence in things being delivered when we’ve been told they’re delivered and that was very vague. But the number of vaccines we’ve seen rollout, I’ve been seeing encouraging science from my viewpoint across government in terms of things actually happening when we say they are. When an agency says they support something, they support something. So I’m extremely hopeful as we continue to move forward.

Michelle Thong:

As we look forward, I also want to look back and recognize the enormous amount of good that was done by unemployment insurance benefits and policy and implementation happening very quickly. Just over a year ago, with the CARES Act, and PUA, and so many public servants working at incredible speeds under incredible pressure, one in four workers in the country relied on unemployment insurance for at least some point in the past year. That’s 46 million individuals.

Michelle Thong:

So we did see the enormous benefits of really groundbreaking policy that was then implemented to get that out. We also learned a lot of painful lessons about what wasn’t working and what was brittle, and now is really this key moment, critical moment for us when we’re still fresh from those really painful lessons to take what we’ve learned and translate that into momentum to build towards this big, new bold vision of a resilient safety net.

Alyssa Levitz:

Wow. Michelle, I feel like you said what I was going to say, but so much more eloquently. I think, reflecting back on this last year, the staff and the people who have been doing their very best under really hard circumstances, including death threats and weekends away from family to help people, and sometimes knowing that there was not that much they could do, but still coming to work and doing it for their neighbors and other people in their community. I think that is incredibly meaningful and incredibly powerful and I thank them so much for their work. And I hope that we really are able to take this moment and use it as an opportunity to make things better and more resilient for next time. Because unfortunately, this is not the last procession we’re going to have, but I think that we can use this to make sure that we are better prepared.

Ryan Ko:

All right. Thank so much Allison, Alyssa, and Michelle and thank you of course, to Nava for sponsoring this session. If you enjoyed the session, there will be many more on this and similar topics at the upcoming Code for America Summit, May 12th and the 13th. You can get tickets at summit.codeforamerica.org, you can check out the speakers and the track sessions. Tickets are $50 for another week and then they go up to $75. Sosummit.codeforamerica.org. With that, we are past time. If you’d like, I did also invite panelists to share LinkedIn or email addresses in the chat. Probably, folks want to follow up, I think we’re all on Twitter and very active there as well. So we’ll direct both there. Thank you everybody for joining. Have a good day.