Interview: Swati Mylavarapu on government service delivery and reinventing citizen experience
In this episode of Rewiring Government, CEO Joshua Goldstein talks to Swati Mylavarapu, a partner at VC firm Kleiner, Perkins, Caufield and Byers, about the challenges and opportunities for software startups, and tech companies more broadly, in helping government improve the way they operate and deliver public services to citizens. The conversation is bookended by some bigger picture questions about the role that technology companies play in solving some of society’s largest problems.
Use the player above to listen, or subscribe on iTunes and Google Play! You can also add our RSS feed to your favorite podcast app. If you like this episode, rate and review us on iTunes, and tell your friends.
A transcript of the interview is below, edited for content and flow.
Joshua: Swati, welcome to the show.
Swati: Thank you. Happy to be speaking with you today.
Joshua: I want to start with a question about you and what motivates your work. You write on your bio that you’re passionate about the role great technology companies play in solving world challenges. I know you’ve looked at these challenges from a variety of angles through your career. We briefly overlapped at Google.org, [where there’s] a nonprofit angle, there’s philanthropy, there’s policy. What appeals to you about the particular tool set that technology companies are bringing to the table, and how would you characterize how well they’re doing at taking on some of these challenges?
Swati: Josh, that’s a huge question, and I think we could spend the whole podcast just answering that one.
At their best, one of the things that really great technology companies do exceptionally well is that they’ve figured out how to reinvent, or in some cases to invent completely new, experiences that delight people or add real measurable value, and then they’ve figured out ways to deliver that experience at tremendous scale.
Think about a company like Google and how ubiquitous that experience of searching for information, how pervasive it’s become in a remarkably short amount of time. As a company, Google’s less than  years old.
Or Facebook, which has really come of age in our lifetime. You and I can argue about the relative value or impact of being able to poke or message people, but I think it’s hard to argue with the way in which it’s enabled people to connect remotely. And now it’s something that one out of six or seven people on the planet are using in some sort of meaningful way. To be able to touch the lives and experiences of so many people in such a remarkably short amount of time, I think, is something that great technology companies can do remarkably well, and it’s that speed of change and experience that makes me most excited about them.
Another great example: Nest, in a short amount of time, has shown us the potential to reinvent an unloved object in your home, like a thermostat. To date, I think they’ve saved over a billion kilowatt hours in energy. Also, as a result of that behavior, they have taught people that there are elements to the experience of controlling climate and temperature in your home that have really great environmental benefits, delight the user in a way that we wouldn’t have expected, and just totally reinvent that experience. So, these are examples of what’s really possible.
When we think about the opportunity to change or improve or update, for example, different ways in which government services are delivered or ways in which government operates, the scale at which government touches lives is obviously compelling. But I wonder if there are ways in which that process of innovation, reinvention of experience, and speed—the best of the technology industry—can be coupled with the best of what government does.
Joshua: I couldn’t agree more.
As someone who’s worked on the policy side of things too, I think that [technology companies bringing the ability to rapidly ramp up, rapidly measure, and then iterate] is incredibly promising for some of these hard, nitty-gritty public sector problems.
Let’s turn to those challenges—government has no shortage of them, right? From your vantage point, what’s interesting to you about the GovTech space, and how, to the extent that you have followed it, have you seen it evolve in the last couple of years?
Swati: I don’t think I’m able to speak to how the GovTech space has evolved, but I see so much opportunity for change and improvement.
Think about any aspect of customer experience (or a “citizen’s” experience) in interfacing with any kind of public service that you tend to interact with a state agency for, and I think you’ll find a long list of opportunities to improve that service level. That’s something that I think the best of software has really reinvented.
I spent a number of years at a company called Square, before coming over to the investing side. I don’t think anybody would have thought that small business payments was a particularly sexy or fun experience. But, lo and behold, here came this company, and [they] made it a tremendously delightful experience. [Now,] people have fun swiping a credit card through that plastic reader, and [so do small business owners after] getting set up and realizing that it’s instantaneous. [They’re] using that point of sale and realizing that it’s so much more than punching keys on an analog cash register.
I think in the same way that that type of transformation was possible, with what was once thought of as a not very glamorous or delightful experience, [there are] so many opportunities for that type of reinvention when it comes to the way in which we get services from the state. Tax Filing is an obvious example. That’s something that Americans do every year. All of them. And there’s [been] some improvement there with the Intuit integrations and QuickBooks, and the like. But obviously [there’s] more room for improvement.
The way in which you go to the DMV and get a driver’s license—and what you’re going to go through to get your car back today—is another great example. These are things that touch people’s lives on a daily basis.
I think there’s a plethora of examples like that. I’m curious where there are opportunities in the service layer, opportunities also to close the feedback loop. I think that’s particularly relevant in this period. We’re in the throes of a very contentious election cycle and really wondering why we were so surprised by Bernie Sanders and Trump and their support. I wonder if there were ways to communicate with a broader group of people and to understand the ways in which citizen sentiment was changing. Maybe we would have had an political system that was better attuned to citizen needs.
That’s something that I think software companies, the best of them, have figured out how to do. Because your livelihood as a business depends on your ability to cater to end users—to capture their needs and interests and demands and cater to them.
Joshua: I think that’s a really great point. And there seems to be two angles on that. One is using technology to better understand people’s needs and desires, and the other is to be better at delivering on the things that the public sector has promised to people. [This goes along with the question of what drives] the Bernie and the Trump supporters. Is there a role that tech-enabled public delivery mechanisms can play to improve that?
That also gets to something that I’ve been thinking a lot about which is a sense that government is special as a customer, in some sense. There’s another version of the story where it’s just another large enterprise that’s going to go through the same things that insurance tech, that FinTech, that other highly regulated industries have gone through in terms of software transformations.
Do you have a sense of whether either of those narratives are dominant and what that will mean for companies in this space versus larger enterprises overall trying to play with government as a client?
Swati: I think it’s an undeniable truth that as a class of potential customers, from the perspective of a company that would maybe want to offer services, there are a lot of unique attributes to government agencies and government entities. Some of the obvious ones are the pure scale at which a number of them operate to administer systems and tools that touch nearly 400 million people. That’s expansive. Not many technology companies admittedly operate at that kind of scale; a very small, teeny-tiny percentage of them do.
I think you could make the argument that there is room for zero error in a number of government functions. We all saw what happened with Healthcare.gov. There is a degree of reliability that you want in your revenue collection system for the state. You wouldn’t want there to be lots of bugs in your tax collection system, for example. The systems that we use to administer grid access and utilities. There’s a reason why you make sure that there’s ubiquity of access and also reliability in service delivery.
That said, there are a lot of technology companies that deliver services bound by similar constraints, if not on the absolute scale side, certainly on the reliability side. Visa: What would happen if their network went down? What would happen to global payments if their network went down? We know what happens when Amazon Web Services goes down. It’s not a pleasant scenario. These are companies that strive for zero error. And of course, we saw what happened when Healthcare.gov launched. That was far from a zero-error launch.
While there are attributes that make the sector unique, there’s certainly so much room for improvement to take the best from both worlds and see where there are opportunities to deliver zero-error services, at scale, in a reliable manner. And I think that there are a significant number of players in the technology industry that are solving those kinds of problems for other large enterprise players. There are similarities there, and I urge us away from going too far down the path of the unique snowflake conversation.
One of the untold, larger trends of our time is the story of the digitization of a lot of things that were formerly analog. You can argue that there is no core single technology innovation that enables [many of these new, exciting businesses like Uber] to exist. Instead, it’s the combination of a lot of platforms that have been able to proliferate: cloud computing; the existence of smartphones; the proliferation of car ownership; appetite in the insurance industry for insuring new classes. It’s the ability to combine lots of different pieces that make it really unique.
The larger trend here is that we’re moving services that once didn’t enjoy the benefit of a lot of these technology pieces and reinventing them, using a lot of those technology enablers.
Similarly, you have to ask the question what’s possible on the government side of things, on the services that people are accustomed to from the state. A common conversation that I have with my friends, because we’re in an election cycle, is this: How do you access information for making a voting decision, and what’s the process of voting?
Compare that to the way in which we would make a consumption decision today, where so much of it is enabled by your smartphone. With a few touches of a button, you’re able to get the information that you need to make other really important decisions, like where you’re going to open a bank account or which insurer you’re going to sign up with.
It’s not easy to do that when deciding who we want to run our local, state, and federal governments. We’re still using paper and pen for voting! And that’s a system that we want to be at scale and 100-percent reliable and where there is clearly lots of room for improvement. I grew up in Gainesville, Florida, in Alachua County. It was a small number of hanging chads that swung the  election one way or the other.
[All that to say that there’s] potential for digital technologies to reinvent, and in some ways, strengthen or improve services that were designed many decades ago, at a time when consumers accessed information, made decisions, and took action in a really different way.
Joshua: That makes me think of some of Ethan Zuckerman’s work on access to information, specifically around this surprising trend that, despite having access to conceivably every bit of information in the world through a web browser, we still have the homophily problem of relying on sources that agree with us already. In terms of accessing the information to make a decision, there’s a lot of design challenges there.
Stepping back to your point about reliability and zero downtime, I enjoyed your podcast with your colleagues, Ryan [Panchadsaram] and Mike [Abbott].
Swati: They lived it. They lived that Healthcare.gov experience.
Joshua: One of them made a really good point that on day one of Healthcare.gov, the amount of traffic was similar to Pinterest. Yet, they rolled it out without testing. It makes me think of some of the “market forces” around the way government chooses to solve technology challenges, and [how that] led to that situation in the first place.
What are the trends that suggest either that we’re hopefully avoiding those situations in the future (or not)?
Swati: Well, to give credit where it’s due with Healthcare.gov, the inclination to unveil a large program over a digital platform absolutely resonates with a lot of what you and I are talking about. I think it’s in line with the way in which citizens are now accustomed to interacting with information and making some of these choices. It made a lot of sense, and it’s hard to critique that part of it. In fact, I think that’s quite laudable. It’s enabled that system to reach a number of people in a remarkably short amount of time.
But to your point, maybe we didn’t get the advantage of how world-class software teams are building leading edge solutions in Silicon Valley (through agile development or by testing and gradually ramping up to scale, for example), perhaps because we’re ascribing to this point of view that there is something unique in the way that we should deliver or build services and systems and tools on the government side.
I think we might have gone too far down the path of not enabling those learnings. And in fact, what we ended up with was a scenario where there was a fire drill, and a few folks from [Silicon Valley] had to work with the great folks from [D.C.] to try and solve it. It makes you wonder what would have happened if that meeting of the minds had happened a little bit sooner.
Joshua: That gets at one of my concerns. I think when you talk to the 18Fs and the USGSs of the world, they’d be the first to say that no fire drill team or core team within government can sufficiently address the challenge of transforming government to digital flexible services, [at the scale necessary]. This then takes you to the thorny problem of things like procurement, things like IT security, and even on a case-by-case basis, a lot of the regulatory and legal hurdles that make some use cases more difficult to transition to digital.
In technology companies, we take for granted that we can replace a system with whatever we think can get the job done the best. But, we’re finding that it’s obviously a bigger uphill battle [within government]. We’re running into situations where the structure of the PDF is mandated in a law, so for it to be machine readable and intake data is a more complex thing.
Do you have a sense or have you seen signals that this “marketplace” is changing for government?
Swati: I think so. And the big question is, How quickly will it change?
We’re speaking from part of the world where I think there are particularly colored expectations around the pace of change. I think our expectations are that things change remarkably quickly and are maybe out of touch with the way they work in a lot of other industries. That’s the huge grain of salt that I want to make sure that we touch on in this conversation.
But, from the few conversations I’ve been privy to, I think that there are convenings of CTOs and the like in D.C. to talk about exactly this. There are really great folks like Ron [Bouganim] and Jen Pahlka and others who are representing the interests of the technology industry and startups and [speaking to] some of the challenges with procurement and the ways in which it might be improved.
I see movement when it comes to policy ideation, for the potential next administration to really take this on and address it. I think we’ll get there. I really do. And I think it’s for the betterment of everybody that we sort this out.
I don’t want to over-simplify just how complicated it is. It is remarkably complicated. But my sense is that it will be better for everybody, because the challenge is not just that government agencies make it really hard for startups to participate in procurement. There’s also a cultural aversion for many startups to start with the public sector as a major client group, in part because there’s this perception that the process of going through procurement and managing that type of client is so much more challenging. It’s something that many of them will constantly punt or put off until later, if at all.
Starting to address this via conversation and ideas, contributed from both sides of the relationship, will go a long way—not just in making it easier for startups to go through procurement, but also, frankly, making it a more palatable revenue option and source for companies out here.
Joshua: There are really promising signs.
Are you familiar with the California Child Welfare Services project?
Swati: A little bit.
Joshua: [To me, that project] is one of the signals.
For folks who aren’t familiar, the folks from 18F, in collaboration with Code for America and others, helped the State of California break up a many billion dollar thing—which was poised to be another Healthcare.gov—into much smaller projects.
Wen you break those projects up, it actually makes the playing field a little bit more level. You no longer have to be one of few companies that can navigate government. You can just be a great software company and take on a modular piece of this. That’s really encouraging to me.
You mentioned the government being a potentially slow first client and that sort of “grain of salt” with our expectations. That’s definitely something that I’ve observed, coming from a company that started in the very earliest days in D.C. and then came out here to the Bay Area. I think there is an inherent tension between the two, with the growth that you need to accelerate and with some of the expectations and funding models.
I wonder if there’s room for innovation around that or what needs to happen to encourage more folks to enter this space.
Swati: Part of what makes the procurement conversation so important and interesting is also related to demographics. The long and short of it is that we are on pace to have too few computer-science capable employees and graduates in this country in the next 10 years to fill the number of roles and the demand for those kinds of positions, both in the private sector and, candidly, in the public sector. [This is] in part because of the other demographic shift that I just talked about, that we are in the age of the digitization of many things. It’s not just about a small class of technology companies that are generating a few online tools.
One of the greatest value-creation stories today is about the creation of software to replace things that once ran over paper or other kinds of analog systems. That’s why you see so many phenomenal enterprise software companies these days; they are taking these once-analog business processes and figuring out how to basically build a remote IT system. It’s part of what made Square so exciting. They were building large-firm software and making it available for free, or at very low cost, to small businesses.
Think about what that does, not just in terms of a low-cost product that a company pays for. As a small firm, you don’t have to take on the responsibility of hiring your own IT firm. I think that’s true also for medium-sized and larger-scale firms. The best enterprise software solutions enable you to augment and complement, and maybe actually have a smaller in-house staff of dedicated IT professionals.
But still, when you look at major government agencies, you see large, full-stack IT teams. We want to reengineer that whole stack from the bottom up. But, how do you compete with the best graduates that are also now looking at opportunities from places like Google or Uber or Facebook or Amazon? Acknowledging that there is a near term future where the talent market itself is going to become that much harder to solve for, it seems like it just makes better sense to ease up procurement and give ourselves the option to take at least pieces of our IT solutions from firms with a comparative advantage and provisioning exactly those services.
Joshua: We see that as well, where our best customers are those frontline users who have gone through some traumatic event, either with the internal IT team or outside consultants, and they just want to control their own systems and do so with drag-and-drop and be able to do everything in a web application.
I definitely see that trend, but you’re right that with some of these systems, the organizational systems are not quite ready to handle that, particularly because a lot of folks will end up choosing a sort of “grey” infrastructure. (They’re not going through the official approval process because it’s too painful or too complicated.)
To me, one of the surprising things on that front is that procurement is a whole world of things that you have to navigate, but there’s also a parallel world of IT security approval for products, for cloud products in particular. At least at the federal level, I think that this has been even more of a barrier to adoption of good SaaS tools.
There’s this whole idea in the federal government of authority to operate (ATO) that, again, was very well-intentioned when it was implemented—we should make sure that we are choosing secure tools. But these are multi-year processes and cost millions to get through. For example, AWS is the first cloud service provider to go through FedRAMP, at a very high cost.
However, there’s a whole generation of earlier-stage companies that want to serve government, but going through that formal approval process is prohibitive. There’s definitely room for a movement and a conversation to ask, “How can we adapt this to a new generation of companies?” so that they’re not blocked out of the process.
Swati: It makes good sense to me, but at the same time, I hear the other side.
I think it’s a legitimate question and [presents] an opportunity to dig deeper. Maybe it’s not a binary—”Should we allow this or not?”—but more about what solutions or what size of startup is appropriate to address each size or class or type of challenge on the government side.
If you were to ask me whether a five person company should be solving cybersecurity for the Pentagon, I could point to the obvious risks. I can understand why part of the procurement process biases towards larger, more established, more proven firms. This business of starting new technology companies is a remarkably high-risk business with a lot of existential threat to newcomers.
[On the other hand,] are there opportunities for an entrant that’s providing a relatively straightforward SaaS product to cater to city and state governments in slightly less risky scenarios? For the digitization of forms, as you all are taking on, or feedback from citizens or the social media based engagement of citizens. Those things feel like they might be more appropriate for a loosening or a changing of the way we procure.
I think that’s the conversation we’re ready for and that examples like the one that you just described, in the State of California, help guide us in the direction of. I’m excited for the richness of that conversation to play out.
Joshua: I definitely agree that the city and state level is very rich for a quicker transformation and opportunities.
In the last few years in GovTech, there really does feel like there’s this sense of excitement and a movement with a social change phenomenon to it. [We’re] bringing technologists into government with programs like the Presidential Innovation Fellowship and even Code for America. But you also see a lot of early stage companies that have had trouble getting to product-market fit, which of course is going to be the case in any sector.
Do you have any advice for entrepreneurs who are interested in the civic tech or GovTech space in particular about how to position themselves?
Swati: This may seem slightly counterintuitive, but one of my early pieces of advice would be not to be too constrained by this sector label of being a “GovTech” company and focus much more on the real problem that you’re trying to solve and for whom, along with the most important aspects of that product and service delivery experience that you need to nail to have a delighted customer.
Show that there’s value in that experience, that there is an appetite to pay and a willingness to stick with the solution over time, and then start looking at the different use cases to which that applies.
If you step back from all of that and realize that you are really positioned to serve public agency players, at this price point, in this way, that’s the aggregated lesson that comes out of that zero to one process of iterating and finding product market fit. But I’d focus very acutely there, rather than just saying out of the gate, “I’m a GovTech company.” I think that kind of category focus can sometimes be too restrictive, and you might find along the way that you’re able to sell to certain private sector clients at the same time. I mean, there’s a lot in the solution that you and your team are building that is not necessarily public sector specific. Look at where those opportunities arise.
Joshua: I think that’s spot on. There’s often research on the civic tech space and GovTech space that asks, If government or the public sector had access to any tool in the world, how are you solving the problem better, by 10x or whatever? [That answer matters more] than the sector itself.
To finish up, I wanted to ask an inverse version of what I asked at the beginning: Where is there still a role for the public sector to take on these public goods and solve problems that only the public sector can take on? What does that give-and-take look like?
Swati: At this intersection between government and technology, the really exciting potential is not just in technology companies building better products and services for government and in government making it easier to implement or incorporate. It’s also [the potential for] public sector actors, engaging in a richer conversation and exchange with the technology sector writ large, creating incentives and signals and pressures to encourage more technology players to cater to the long tail, or the group of people that we oftentimes overlook.
I find it staggering when you look at the numbers of American households that don’t have dedicated reliable Internet access in their house. How many of them actually have access to unlimited data and smartphones for all family members? Lest we assume that there is a future where everything ought to be distributed digitally, let’s also acknowledge that there is a real public good component to the infrastructure via which these services are delivered. There’s so much more room for improvement there, for us to be able to make sure that these capabilities and opportunities are in the hands of all Americans.
That was [the goal of] the postal service, of the public education system. These are some of the best of what government services have delivered and created and enabled for the vast majority of Americans, and there is an element of that that I think this industry would be so well served for incorporating as well.
Joshua: So, I think we have to wrap up have to get back to navigating the government process of getting my car out of the City of San Jose tow lot….
Swati: Don’t you wish there was an app for that?
Joshua: Yeah, I’m working on that as the next project. Thank you so much for taking the time and for your excitement and enthusiasm about this space. It’s going to be really fun to see where it ends.
Swati: I agree. I can’t wait. Thank you.
Joshua Goldstein is the VP of Product at CityBase.