00:00hi everyone welcome to the a 6 in Z
00:02podcast I'm sonal and I have two special
00:04guests from the USD s or the u.s.
00:06digital services continuing our DC on
00:08the road theme for those of you that may
00:10remember some of the Fiasco's with the
00:12original healthcare.gov this is a group
00:15that came in and fixed it and more
00:16importantly has been working on many
00:17other services since then the two guests
00:20we have today are Mikey who is the
00:22administrator of the USD s and before
00:24that he was an engineering manager at
00:26Google for a number of years and hailey
00:28who is a deputy administrator at the
00:30USGS she's been working to help
00:32implement some of President Obama's key
00:34technology initiatives during his
00:36administration so Hailey and Mikey
00:38welcome you guys thank you for having us
00:40Thanks yeah well I'm excited to have you
00:42guys because the government is like the
00:44biggest buyer of IT services and then on
00:47the flip side of this we have this idea
00:50that you can have a start-up inside the
00:53government and so there's just
00:54interesting tension between this idea of
00:56the government as a buyer and the
00:58government as a builder and you guys are
01:00sort of sitting at the heart of that
01:01yeah well just to start with how we
01:03operate so we are essentially a start-up
01:06that's embedded inside the White House
01:08and as part of a team of teams working
01:10across government to help improve some
01:13of the core citizen facing services of
01:15the United States government we bring in
01:18some of the best minds from the private
01:20sector and set up tours of duty where
01:22people can come in and help apply those
01:24same best practices and skillsets to
01:26improve the fundamental services our our
01:29country provides I think it's
01:30interesting that you referenced that as
01:32a sort of a tour of duty because one of
01:34the complaints that I've heard you know
01:35we have a lot of government folks that
01:36come in through our doors is that it's
01:38really hard to attract top talent to
01:40work in government because they would
01:42rather be at a start-up or you know a
01:44company like a Facebook or a Google or
01:45an Amazon Apple and so this idea of a
01:48tour of duty is really interesting
01:49because what I was wondering is if you
01:51actually have an adverse selection
01:52problem where the very people you want
01:54to attract don't want to go there and
01:56the people you don't want to attract are
01:58the ones that do and and how you find
02:00yourself selves resolving this to
02:02attract the top talent yeah it's a great
02:05question when we first started you SDS
02:07we had a big question which was were we
02:09going to be able to attract any top
02:11talent to come I believe the Ruiz
02:13why the tour of duty model is working so
02:16well is one very fundamental reason and
02:19that is the mission we offer a huge
02:22opportunity to have impact on things
02:24that really deeply matter in the lives
02:26of Americans and that is our comparative
02:28advantage and you hear often around the
02:30you know the valley and startups the the
02:32interest in wanting to have impact in
02:34humans lives yeah it's funny that you
02:35bring up the adverse selection problem I
02:38talked about I talked about it exactly
02:39that way not quite a bit we don't
02:41encourage people to think they're gonna
02:43advance their own career in any
02:45particular way the recruiting pitch we
02:47give is is very non-intuitive we set
02:51people's expectations that it's going to
02:52be very hard and hopefully in that way
02:55it makes it much more likely that we're
02:58gonna get the right kind of person that
03:00we need and it also makes it much easier
03:02to manage after they're here if they if
03:04they've only been promised that they're
03:06going to have an opportunity to work on
03:07something important that is a promise we
03:09can deliver what exactly are they
03:11working on if these folks are motivated
03:13to have this impact but yet have been
03:15exposed to a very different environment
03:16of getting done for lack of a
03:18better phrase how do they sort of deal
03:20with navigating there it seems like any
03:22large entity forget even the US
03:24government large corporations and
03:25enterprises are like this there's
03:27multiple stakeholders there's lots of
03:29inputs you have decision by committee in
03:31some cases how do you guys operate in a
03:33start-up like way it's tricky as you say
03:35there's definitely a different
03:36environment when you're working on
03:38something that's critical to hundreds of
03:41millions of people you know tax filing
03:44Social Security anything like this that
03:45so I'm going to be Facebook style move
03:48fast and break everything and we try to
03:50minimize that effect by focusing you SDS
03:54on projects that are very high urgency
03:57and very important to the leadership
04:00that we're working with because
04:02government can move fast when it wants
04:06to or when it needs to so when it's
04:08urgent basically is what you're saying
04:09right you mentioned the healthcare.gov
04:11story and I was part of the group that
04:13worked on that and we got a tremendous
04:15number of things done in the short in
04:17the space of a very short amount of time
04:18six eight weeks or so because it was
04:22mission-critical and it was incredibly
04:24important and everybody in the whole
04:25organization knew that and when we have
04:27those conditions we can get things done
04:28there's a world of difference between
04:30moving a little slower than you're used
04:32to and not moving at all what are some
04:34of the similarities between say you know
04:37a regular enterprise and government is
04:39it like every large organization or are
04:41there like other differences one of the
04:44thing that's funny to me about
04:45government is is people feel like
04:47bureaucracy is you know patented by
04:49government large companies don't have
04:52huge bureaucratic structures as a result
04:55there are a lot of similarities between
04:56how you would operate in the private
04:57sector and and here and a lot of the
04:59things that we have gone back to as the
05:02proven method for for creating change is
05:06keeping very very small teams and having
05:09them hyper focused on a specific mission
05:11and then doing what they do best which
05:14is bringing in a lot of the best
05:17practices from the private sector the
05:18things that everyone knows a second
05:20agile development and bringing them into
05:23government where they are not second
05:26nature most of the challenges of trying
05:28to organize a large large group of
05:31people and get them to get something
05:33done and move those are exactly the same
05:34it wasn't different at Google if you
05:36have a thousand engineers and not a good
05:38decision making mechanism and they're
05:40not going to agree on how to do things
05:41big companies are struggling to move
05:45into shared services cloud hosting off
05:50of mainframes like there's a lot of
05:51industries that are in exactly that
05:52place today and that's a lot of
05:54government agencies as well and the
05:57mechanisms that work best that Haley
05:59just pointed out one that we get the
06:01best results with is a small group of
06:03very capable and empowered people and
06:07that works that works the same in the
06:08private sector or the public sector in
06:10my experience yeah it's actually almost
06:11like the tension between small companies
06:13and big companies but all in one place
06:15I can see now why you guys would call
06:16yourselves a start-up inside government
06:17just like a startup going into a big
06:19company and some of the similarities you
06:21described you know small teams shared
06:23mission agile you know the urgency all
06:25these factors are very startup like the
06:28difference however and I think this is
06:29an important one to bring up is that
06:31when you are you know a big company you
06:34have repeat customers and you guys are
06:37essentially working across very
06:38different you know stay
06:40holder's every time you have like sort
06:42of a repeated knowledge that then builds
06:43up that you can carry across because
06:45it's so different with every group I
06:47would imagine you know it's funny
06:48looking at government from the outside
06:50and it's easy to think of it as kind of
06:51one single entity or something that
06:53operates similar to a company but in
06:55reality it's a gigantic massive industry
06:58each of the agencies of the federal
07:00government if you peel them out into the
07:02private sector would be larger than most
07:04fortune 500 companies so it's a huge
07:06space that we're operating in so yes we
07:08do have very many different customers or
07:10two partners that we work with but at
07:11the end of the day it's very easy to
07:13stay focused on the same consistent
07:15stakeholder which is the end user or the
07:17citizen or the beneficiary of the of the
07:19service that we're trying to improve and
07:21that relentless focus whether it's a
07:23veteran that we're trying to get better
07:24access to benefits or someone who are
07:26trying to improve getting access to
07:28student loans it's it's very easy to say
07:30focus on that on that common user and a
07:32lot of the tactics end up becoming the
07:35same across various parts of government
07:37various agency the extent we have
07:39principles that we feel are near
07:41universally useful a bunch of those we
07:44put into the playbook they're pretty
07:45high level things they're like I think
07:47number one is like have one person in
07:49charge who is going to have the ability
07:52to make decisions across the program and
07:53another one is you know pay attention to
07:56what the end users actually want they're
07:58so high level that they're that they're
08:00nearly platitudes but like you cannot
08:02take them for granted in any big project
08:04here or in the private sector people
08:07it's amazing how often big groups people
08:10lose sight of what the thing is they
08:11were trying to actually do in the first
08:12place well share some examples at me I
08:14mean I think some of the most I find
08:16most personally interesting are examples
08:18of the work you guys did for veterans
08:19services and student loans those are two
08:21important constituencies that are often
08:23overlooked those are two great examples
08:24we helped create a product called the
08:27college scorecard which lets people
08:29search for information about earnings
08:33after graduation and other stuff that's
08:35relevant to what college you might want
08:37to go to there are a lot of ideas
08:39education policy it gets complicated
08:41fast as you might imagine and we helped
08:45bring some like direct feedback from the
08:48target audience so we had user
08:50researchers talk to high school students
08:53just for just a positive that for a
08:55moment that feels so obvious to a
08:56Silicon Valley startup and how they
08:58operate our test products it's a it's
09:00not something you can assume is gonna
09:03maybe I was just too far away from it
09:04and companies I worked in the private
09:06sector but it was not a thing I could
09:07take for granted there either I mean
09:09there's also kind of a counter-narrative
09:10of like the hero product visionary right
09:13like I mean I don't think I've ever read
09:15stories of Steve Jobs like going and
09:16asking high school kids about the iPhone
09:18like he's just we just assume that he
09:19knows everything there is to know you're
09:21right and yet there's this truth as well
09:23despite whatever the biographies say
09:25that there was also a tremendous amount
09:27of work that happened beneath the
09:29surface that wasn't all just about only
09:30his vision and how they execute on it so
09:32it's a really great point yeah I think
09:34you can't take it for granted anywhere
09:35and it's in it's worth writing down and
09:38reminding people reminding whole teams
09:40from time to time that like oK we've
09:42iterated this design eight times now did
09:43we ever at any point go back and ask the
09:45high school students are they ever going
09:46to use the search feature
09:47so it's interesting because you know
09:48there's this movement in in usability
09:50and user testing where there's sometimes
09:52a tension between asking people don't
09:54actually do what they say watch what
09:55they do how do you resolve that one my
09:58quarter field is engineering so I won't
10:00pretend to be expert at this but I've
10:01only ever seen that work by doing a
10:04direct observation it's exactly as you
10:06say like it does not work I mean there's
10:08that you know everybody has their
10:09favorite aphorisms there's like Henry
10:10Ford saying if I'd ask people that what
10:12they wanted they would have said faster
10:13horses right right and by the way there
10:15was a funny HBR article a few years ago
10:16that actually debunked that he's even
10:18the one who said that but I think the
10:19point remains valid and that is a very
10:21widely cited quote whoever was the
10:23Senate I agree it absolutely doesn't
10:26work to ask people what kinds of like in
10:29the case of the veterans employment
10:31center veterans healthcare application
10:33facilities locator all these types of
10:36things that are services that we need to
10:38work well for a large number of very
10:42diverse people with different needs it
10:44doesn't work to ask a group of people in
10:47a conference room what features they
10:49would use you have to go and watch them
10:52and it's much better to go where they
10:53are and watch them in their own
10:55environment we learn crazy stuff there
10:57was a there was a we've done some user
10:59research with the veterans where our
11:02researcher came back and said like her
11:05mind was blown that this guy
11:07trying to operate the VA website with
11:09his laptop like balanced on an ottoman
11:11like with one hand and like kept
11:13dropping a laptop on the floor and like
11:16he was only operating it with one hand
11:17and is having a hard time and that's
11:20probably not a use case that we would
11:22have thought of if we just sat around
11:23and brainstorm the focus on the user is
11:26just so critical to everything we do and
11:27at the at our work at the VA it's very
11:30very apparent and so if you look at the
11:32user experience from a vet's perspective
11:34today there's hundreds of different
11:37callin numbers that you can call to get
11:39help in thousands of different websites
11:40that you can go to which is unfortunate
11:42because that's not an ideal scenario in
11:44terms of providing services for veterans
11:46our team is building an integrated
11:48experience that's easier for vets to
11:50find access to the benefits and services
11:53that they are looking for instead of
11:55just just thinking about what that might
11:57be we actually went out into the field
11:59and did a lot of work with veterans to
12:01figure out which were the services they
12:02were looking for the most to start with
12:04those in a very agile and iterative
12:07fashion we wanted to start releasing you
12:09know streamline services quickly instead
12:11of waiting to redo all of them so we did
12:13a bunch of user research with vets and
12:14realized that the services they were
12:16looking for the most were education
12:17benefits and disability benefit but
12:20another user experience piece of this is
12:22not presupposing where the fix is going
12:25to be so when our team started working
12:27really closely with our agency partners
12:28there we realized unexpectedly that one
12:31part of the problem that could actually
12:33help improve that's getting their
12:36benefits was actually working on
12:37improving this the the tools that the
12:40the VA employees themselves were using
12:42which was obviously not a problem we
12:43went into the VA focused on we released
12:46a new service called case flow which is
12:48a new tool to actually help the
12:50government adjudicators themselves more
12:53efficiently process applications and do
12:55their job more easily yeah
12:57case flow is a good example because it
12:59is not a flashy website when we started
13:03digging into those claims does this is
13:06disability claims process inside the VA
13:09the biggest and the greatest impact we
13:11could have there was really to just
13:13speed up the internal workings of the VA
13:14from the time the claim is submitted to
13:17or when the appeal is adjudicated those
13:20are products that we're pretty proud of
13:21that nobody's ever going to see the
13:24magazine or anything because it's not
13:25glamorous like that but it has a
13:28potential to take months or years off of
13:31the time that the people are waiting for
13:33their disability benefits that's hugely
13:35important that something that impacts
13:36human lives in a very real way you know
13:39I'm gonna ask you guys a tough question
13:40though which is going back to this
13:42tension between government as buyer and
13:44government as builder and I think it's
13:46interesting because you guys are
13:47navigating that by being the startup
13:49inside government but the question I
13:51can't help but ask is that so what was
13:53the incentive for creating this model in
13:56the first place why wouldn't they just
13:57then just outsource the work to a bunch
13:59of different agencies who maybe our
14:00startups outside the government
14:02so outsource to contractors right
14:03exactly we are always trying to optimize
14:06for what's going to serve the citizens
14:10the customers whoever the service is
14:12meant for the best in the fastest in the
14:14shortest amount of time for the most
14:16reasonable amount of money that we can
14:17put into it from the government most of
14:20the time for any project that's any
14:23non-trivial size there's already a team
14:24in place and the best use of our people
14:28that we bring into the government is to
14:30work with that team and help them
14:33sometimes it's a matter of applying some
14:36best practices that they may be a little
14:38bit less familiar with a lot of times
14:40it's just a matter of being able to
14:42traverse boundaries and improve
14:45communication between different parts
14:46the organization honestly that is the
14:49main service we're providing an awful
14:51lot of the time I've been on a few
14:52intense web projects and one of the
14:54things that I learned very early on is
14:56that these projects are actually proxies
14:59for the organizational dynamics at play
15:01and how it's when you get down to
15:04actually building something that you
15:05really begin to see like the the whims
15:07of people the psychology of
15:08organizations and it's funny because you
15:10can't use technology to solve that
15:12necessarily although it can be a driver
15:14to facilitate those very conversations
15:15that you're describing there's this
15:17concept of Conway's law which tells us
15:19that organizations always build systems
15:21which match the shape of the
15:22organization and vice versa
15:24so then the question I have is we had a
15:26podcast earlier with Teresa Carlson from
15:28AWS for public sector and on
15:30and international and one of the things
15:32that you know she talked about is that
15:34in having these conversations with
15:35certain stakeholders you're having to
15:37explain a lot of basic concepts because
15:39they're not necessarily tech savvy from
15:41the get-go and that's true both inside
15:43and outside government but what do you
15:45find especially with repeat folks or
15:47other people you talk to on a regular
15:48basis - people change in how they react
15:51as a result of these projects I think
15:53that the the place that government is in
15:56right now very much mirrors a lot of
15:57what we see big companies going through
15:59where essentially we're removing
16:01technology from around what it used to
16:03be the periphery into the core mission
16:05of business and what we do so one
16:08example one of our agencies we were
16:10explained to them the concept of user
16:11testing and how and why we make product
16:13decisions based off of the feedback we
16:15got from users and you saw all these
16:17light bulbs go off that we could also be
16:19user testing policy development as well
16:21we didn't just have to apply these
16:22practices to building applications this
16:25could be applied to anything so so yeah
16:27it's been really fun to kind of watch
16:29how this culture is kind of taken
16:31written a lot of places that we weren't
16:32expecting one one nuance so that is I
16:35have seen some really positive
16:38second-order effects from executives
16:41learning what is possible just what
16:44their expectations should be or could be
16:47like I don't we don't need cabinet
16:50secretaries to know how to like run a
16:51you know scrum process or something they
16:53have they have other stuff to do
16:55but it is incredibly valuable for them
16:58to just have like a ballpark idea that
17:00you started the question asking about
17:02AWS just the fact that it exists and you
17:05can do small things for small amounts of
17:07money instead of there being like a
17:08baseline floor cost of setting up a data
17:11center for you know tens of millions of
17:13dollars every time which is kind of what
17:14they're used to from these enterprise
17:15style projects just knowing that it can
17:18be done and goes a long way yeah I think
17:20that's the front line of the kind of
17:22culture change that wrapped right now so
17:23you know one of the things I am curious
17:25about again going back to this notion of
17:26government aspire in government as
17:27builder and one of the things that we're
17:29obviously very interested in is
17:30government aspire because they are as I
17:32mentioned one of the largest buyers of
17:34technology and there are a lot of
17:36startups that sell to government and one
17:38of the stories I heard from one startup
17:40not one of ours was that it can take up
17:43to get something bought and it just shed
17:45light on this idea that procurement
17:46itself as a process can be such a
17:49long-winded thing you'd love to get your
17:51take on how that works and from where
17:52you sit the government finds it solves
17:54in these situations where by the time
17:56the process is finished the thing that
17:58you're buying is already out of date or
18:00doesn't serve the needs anymore
18:01it has to be faster and competition in
18:05that marketplace needs to be more
18:07effective which is a combination of
18:09lowering the barriers to entry and doing
18:12a better job with the management of it
18:14all to be clear we like most people
18:18doing government work like we're always
18:19in favor of buying if that's an option
18:21that's good to hear for things that you
18:23can just easily buy off the shelf like
18:25100 percent just there's no need to
18:28reinvent email this year exactly you
18:30know I need the government are really
18:32not different from everybody else in
18:34that respect there are occasional places
18:36we end up in the building role or the
18:37government is doing something that
18:38nobody else does there's nobody else you
18:41know running Medicare there's nobody
18:43else running Social Security so there's
18:44gonna be some stuff where there's no
18:46choice but for a government to strike
18:50out and do something that hasn't been
18:51done before so that procurement work is
18:54really the long-term goal and two years
18:57that we've been here we have put a
18:59decent amount of effort into it we see
19:01we see strides happening in the right
19:03direction yeah you don't buy software
19:06the same way you would buy a battleship
19:08so one of the programs that USGS has
19:10worked on setting up is this contractor
19:12training program which is actually
19:14focused on helping improve the literacy
19:17of contracting officers and how to buy
19:19software and things that it wasn't
19:21trained to buy you know as a government
19:24several decades ago so it's I guess you
19:26know didn't to wrap up one last question
19:28I have is you know been wrote a really
19:30popular blog post a couple years ago
19:32about the notion of a wartime versus the
19:34peacetime CEO and I love it because it
19:36actually goes off one of my favorite
19:37movies The Godfather where there's a
19:38wartime and peacetime consigliere yeah
19:41right and anyway how does that play out
19:43for you guys because it feels like
19:45you're actually a boat at the same time
19:47some organizations can take turns as
19:49being in a wartime peacetime phase even
19:50though there may be an overlap yeah
19:52that's right people sometimes want to
19:54think of the government as a
19:56business when it's actually like an
19:58entire sector of the economy of its own
20:01so as you say we do not get the luxury
20:04is this a wartime period or is this a
20:06peacetime period obviously the
20:07government has Wars but like in this
20:09context of this metaphor it's always
20:11both in different parts of the
20:12government at different times I think
20:14that's one of the things that's been
20:16hardest and I think also one of our
20:18strong suits is that we are able to kind
20:20of do both the peacetime and wartime CEO
20:23together on one hand where we function
20:26at our best is often responding to very
20:29urgent very you know important needs
20:33that have short deadlines and really
20:35need to kind of lean on a lot of the
20:37wartime CEO tactics to be effective on
20:39the other hand we are setting up a team
20:42of teams this network of innovators
20:44across government that needs the
20:46nourishing and attention that you often
20:48find in more peacetime CEOs to make sure
20:50that that culture can cultivate and be
20:51strong enough to continue to create
20:53change and provide a platform for doing
20:55this across government and for the next
20:57several years to come so we have to sort
20:59of switch between the two very fluidly
21:00depending on what context were in the
21:03article really spoke to us in a few ways
21:06it's maybe not an accident that since I
21:10the only organizational management type
21:13strategy theory reading I have done that
21:16has really felt like it had anything to
21:17say that was any use to me is very old
21:20stuff from the 1940s 50s and saying I
21:23don't read hardly anything you know in
21:26the like management section that feels
21:28like is anything but empty platitudes
21:29there's a Rand report from like 1965 by
21:33Anthony downs and it became like
21:35foundation for a lot of social science
21:38in sociology that followed the u.s. had
21:41just come through World War two had just
21:44completed the Manhattan Project had
21:46launched the Apollo project and was
21:48building like the space program the
21:50nuclear industry and the you know the
21:52military as we know it today and they
21:54had just gotten these things done under
21:56an existential threat like you want
21:58urgency there's no more urgent situation
22:00than like the United States 1944 and
22:03they were not screwing around and they
22:06captured a decent amount
22:09of the knowledge that we still need
22:11today in order to actually get things
22:15done with any urgency that's great all
22:17right I would be remiss if we didn't
22:20close on the shameless plug that the
22:22United States digital service is
22:24obviously in dire need of additional
22:26people to come join our team and help us
22:28on these important missions Ted and in
22:30penny Pritzker they did a podcast a
22:32couple weeks ago they are also
22:34advocating for this concept of tours of
22:36duty like look you know no one expects
22:37anyone to stay in any one place let
22:39alone government or a start-up or a
22:41company for you know 20 years and so
22:43it's perfectly fine to have this model
22:44because it's a way that you can actually
22:46attract people but it's also sort of
22:47like a sabbatical who like to work
22:53really really really hard on
22:55I think sabbatical of course has an
22:57illogical origin of something like a day
22:59of rest concept but it what I mean it is
23:01an academic sense where it's actually a
23:03time when you can work on a project
23:05that's really important to you a passion
23:07project some work a book something new
23:09and to your point like if this is a
23:11mission that people want to support yeah
23:13we you use the word like tradition of
23:16public service which i think is
23:17fantastic phrasing you look at the even
23:19the legal field or medicine and it's
23:22very common for people to take little
23:24sabbaticals or do tours of duty from the
23:26private sector and help with the last
23:27mission of public service and that's
23:28what we're really hoping to create in
23:30the tech industry recognizing that there
23:32hasn't been a strong tradition of that
23:33in recent history and we're and that's
23:35what we're trying to cultivate here well
23:37I just want to say thank you Haley and
23:39Mikey for joining the eight 6nz podcast
23:40thanks so much for having us thank you