00:00so Patrick welcome so Patrick is the
00:03co-founder and CEO of stripe he launched
00:07the startup we're now pretty big company
00:09in 2010 correct with his brother John
00:13why should we started working on it
00:14full-time in 2010 but it actually your
00:17comment just there about companies
00:18launching it took us quite a while to
00:20launch cuz we had all these kind of
00:21banking partnerships in place and so on
00:22and so we didn't launch until September
00:242011 and we've been working on it for
00:26almost two years at that point and and
00:30every time we saw PG or really anyone
00:32else from YC all they would ask us is
00:34why we did not launch some things don't
00:37change that's interesting to two years
00:42until you have to it was I think one
00:44year and 11 months from sort of first
00:47line of code to public launch which to
00:50be clear I don't recommend that is not a
00:52good idea it's just we had to because we
00:54had to kind of get all these other
00:55things in place and because it sort of
00:58took us to so long to build a publicly
01:01launched we're sort of we tried to be
01:03kind of Airy disciplined about sort of
01:04gradually expanding the number of users
01:06every month and so even though we
01:07weren't publicly available we got our
01:09first user like first production user
01:11just kind of three months in and then
01:13have every month we tried to add at
01:14least kind of a handful of users and so
01:16by the time we publicly launched we did
01:18have about a hundred users which I mean
01:21back then to us you but that seemed like
01:24a big deal that seemed very very large
01:27speaking of actually when I was
01:29preparing for this interview I I was
01:31trying to talk my memories and I
01:33remember specifically because your
01:35office was near here in Palo Alto and I
01:37remember back then people would always
01:40talk about the call us and brothers
01:41running around going to people's offices
01:44and like installing their API into the
01:47web apps and you know in true do things
01:50don't scale fashion and I assume you are
01:52not only trying to make sure they
01:53installed it but also get user feedback
01:55and it happened so much that actually I
01:57don't know if you know program PG now
02:00calls it the Cullison installation
02:03and this is actually something we give
02:04and we tell founders to go do this
02:08do the call us and installation because
02:10obviously you know it in hindsight it
02:12seems so obvious to do well it sort of
02:14served two purposes and one is to your
02:16I mean it's art it's a really good way
02:17to kind of get sort of user research and
02:20to get kind of UX feedback and so on and
02:22I mean if you've done this I'm sure you
02:24had the experience for you design which
02:25are absolutely certain is the most
02:27streamlined user-friendly
02:28straightforward frictionless way to you
02:30know do whatever it is the product does
02:31and then you kind of you you you you
02:34present it and kind of put it in front
02:35of a user and you just kind of ask them
02:37to do whatever it is and you know they
02:39find it completely impenetrable and like
02:41they're clicking all the wrong links and
02:43they can't find the next button even
02:45though the next button is they're
02:46blinking and green and stuff and so it's
02:48invariably so incredibly painful at the
02:51sort of nothing so sobering as watching
02:54somebody use kind of that the first
02:55version of of you know some new product
02:59but the other reason for us was you know
03:01we would suggest to somebody they use
03:03stripe or they switch to stripe whatever
03:04and you know invariably the responses
03:06oh yeah sure that sounds awesome and but
03:08you know it can be postponed and locals
03:10phone it again and just like there's
03:11never a moment or it's like okay this is
03:13the evening or I'm going to switch to
03:13stripe and so us going and sort of
03:16accosting them in person sort of helped
03:19you know great you know if you'll talk
03:21code in sales it's almost like a you
03:22know of why you and I why this and why
03:24now and these kinds of questions and and
03:28going in person could have created a why
03:30now moment it's like what we're here at
03:31your house did you just show up or how
03:34did you I don't think we ever actually
03:36showed up and although perhaps we should
03:39have but um you know we we we try to to
03:43be kind of at least semi invited so so
03:49stripe now today I mean you've come a
03:50long way since back then I mean it's not
03:53even it's really been a decade not even
03:55but I mean today you're what you have
03:571,300 employees across nine offices
04:00across the world you're doing I have a
04:02list here you manage 200 million API
04:05requests a day you process billions a
04:07year for millions of companies across
04:08hundred 30 companies new latest round of
04:10funding stripe is now worth 20 million
04:15billion anyway let's go going on I'm not
04:19gonna stop they're all there otherwise
04:21people think I'm your PR agent
04:22but anyway you've clearly done something
04:26and so I want to spend a lot of I guess
04:28the time today talking about running
04:32your startup from the perspective of the
04:35start-up CEO you yourself and it's kind
04:39of like zoom like what do you think
04:41about from zooming in and the day-to-day
04:43operations to zooming out to the
04:44long-term strategic decisions so maybe
04:47to help us ease into the discussion one
04:50thing is you know when you start off you
04:52know one from the very beginning a lot
04:53of friends get together and they come
04:56for the idea and they're super excited
04:57and they start working on it and then at
04:58some point they need to decide we need
05:00to see you for this company and some
05:04people you know aren't meant to be CEOs
05:07but for you and John you know I've met
05:09both of you you're very smart ambitious
05:11people with great you know qualities
05:15that end attributes that correlate to
05:16becoming great leader how did you and
05:19John decide you would be the CEO well I
05:22think stripe is unusual for you know
05:26John my RFC brothers and we've known
05:30each other for a long time and and and
05:33you know because of the relationship we
05:36real sort of place a lot of trust in
05:37each other and and we really do run the
05:42company together and there's no major
05:45decision that sort of stripe has made
05:47that sort of that we've not both been a
05:50part of and and you know it's not always
05:54the case that you know despite being the
05:56CEO that kind of I'm the person who like
05:58in the case of disagreement it's not all
05:59it's not all the case that I prevail
06:01instead of our kind of dispute
06:03resolution framework it's kind of much
06:05more around which of us carries more
06:07than kind of which most holds this title
06:10earth and I mean John's title as
06:11president and so it was kind of some
06:13both are significant roles and so I'm
06:18yeah in that regard I think we're kind
06:21of an anomaly he said of the fact that I
06:23became CEO was honestly semi-random
06:28and I would say yeah I think because
06:32we're brothers being able to kind of get
06:33to this unusual situation where we
06:34really do run it together is there a
06:38certain would you sorry may not be a
06:40helpful answer because perhaps you're
06:41trying to encourage all these companies
06:42it's like you know what do you got
06:43guys has got to be the the CEO but do
06:46you think there's like a rubric for this
06:47of you know here are five questions you
06:49should answer maybe then you decide
06:52that's good question um-hmm you know
06:58yeah I'm guessing it's quite okay well I
07:03wasn't one thing which is I think it is
07:06important just have an efficient
07:07mechanism for reaching a decision and it
07:10can't be sort of you know simply orange
07:13around consensus right if there's sort
07:16of three co-founders and sort of none of
07:18you sort of quite want to or you know
07:21nobody's a clear CEO and you don't have
07:24some kind of efficient mechanism first
07:25that of having decisions get made I
07:27think that is a recipe for failure and
07:30and even I don't know doing some kind of
07:32you know sort of quasi democratic voting
07:34is probably not a great idea either and
07:37so so for myself and John we kind of
07:40both have areas we kind of respected we
07:42specialize in and so that doesn't mean
07:44we kind of have absolute ultimate
07:45authority there but so do we be kind of
07:47biased in that direction so he spends
07:49more time for example of working
07:49externally with users I spend more of my
07:51time working to have internally unloved
07:53you know product and engineering things
07:54that nothing to say that he doesn't make
07:56decisions there or I don't here but
07:57again there's a bias in that direction
07:58and then second we have this kind of
08:00additional aspect where you know in case
08:03of a being a major decision and we
08:06should respectively disagree then then
08:08we really do sort of try to make it
08:10based on sort of which of us is kind of
08:11is just more passionate about it and
08:13because that that will correlate with
08:14the outcome one of us really wants to do
08:16something or think that you know X or Y
08:18is the right thing to do simply sort of
08:21wanting is to be so passionately is more
08:24I mean that can become a sort of
08:25self-fulfilling prophecy and so I think
08:27it's kind of not irrational to have that
08:28be a consideration yeah I also see like
08:31the best teams that work well together
08:33are the ones in which everyone they they
08:38all want the best idea to win not their
08:42and so there's a stepping back and an
08:45unselfish kind of way to to get to that
08:48conclusion definitely and I think that
08:51you know something that working on lucky
08:53about that I think is very well exactly
08:57to your point I think he's definitely
08:58present in sort of the most successful
08:59co-founding relationships that I've seen
09:01is some degree of sort of dispassion in
09:04disagreement and you know for us I think
09:07that was you know kind of easier to get
09:08to because you know we've been
09:10disagreeing with each other for 20 years
09:12and so it had lost some of the emotional
09:14charge but I think that sort of finding
09:18other mechanisms whereby you can get
09:19there such that it's not sort of this I
09:21don't know that this kind of but the
09:23notion of disagreeing strongly is not
09:24sort of a scary phenomena and that kind
09:26of both parties or multiple parties if
09:29they're more than two are kind of
09:31suppressing their feelings for fear of
09:33of there being this divergence you've
09:37how many more siblings do you have via
09:39one more sibling yes okay three three of
09:42us in total would you guys would he join
09:44the or is it just you and John are the
09:47special match there well the Tommy my
09:52youngest sibling he's sort of quite a
09:55bit younger than myself and John and so
09:56um you know John is almost approximately
10:02two years younger than I am and you know
10:05when we started stripe I guess I was
10:08about 21 and I think I guess therefore
10:10John was 19 and and Tommy was still kind
10:14of definitively midway through high
10:16school and so it wasn't quite practical
10:18at the time yes finished high school and
10:19now I think you know if you asked him he
10:20probably said he'd never throw his lot
10:22in with you know miscreants like us yeah
10:25cool so in terms of the role is to you
10:28often people say there there's a there's
10:30a threshold in time in which and it's
10:33related to product market fit where that
10:35you have a role as a pre product market
10:37fit to you and then which is completely
10:38different from your role as a post
10:40product market fit to you so I want to
10:43spend most of our time talking about pre
10:45product market fit but just to calibrate
10:48those questions what in terms for stripe
10:51what was product market fit for you like
10:53how did you define it or their metrics
10:55number of employees you are at when you
10:57reached it so forth yeah that's a really
10:59good question and I think you're exactly
11:01right to kind of divide things into sort
11:02of there's kind of the story of a
11:05start-up is two stories and it's a story
11:08of getting to product market fit and
11:09then the story of kind of what happens
11:11subsequently and obviously there isn't
11:13like a totally definitive binary line
11:15between them but I think it's kind of
11:16helpful to frame the narrative in that
11:18regard and I would see for stripe and
11:20actually around the time we launched
11:22publicly I think is is basically when we
11:25had it and that we when we launched
11:27publicly in September 2011 you know we
11:31can rebuild significant components of
11:33stripe kind of multiple times in
11:35response to user feedback like we're
11:36kind of on the third version of our
11:37dashboard and the second or third hand
11:40how you count major version of our API
11:41and so we'd kind of gone through a lot
11:43of iteration response to kind of the you
11:45know evident challenges that these are
11:47hat users had or the deficiencies the
11:49product seemed to possess and and when
11:54we launched we were basically
11:58immediately bottlenecked on sort of
12:00being able to serve user demand rather
12:02than generating user demand and I think
12:05sort of directionally that's kind of the
12:07the sort of the inversion that happens
12:09first it is you know in the early days
12:11you should have really trying to figure
12:11out well okay kind of you know
12:13conditioned on or given some user how do
12:16I make sure that it's if it does what
12:17they need and you know they end up a
12:19happy retained user in a you know a
12:21sufficient fraction of cases or whatever
12:23and then kind of at some point it flows
12:25for okay the sort of well you know I I
12:30think it's kind of very different do you
12:32sort of take you know a hundred users
12:34and you know some fraction of them on
12:37board and some smaller fraction you know
12:38remain engaged or whatever and so kind
12:40of 100 users the curve sort of a sort of
12:43asymptotes downwards and then you take
12:45100 users and kind of again a meaningful
12:48fraction end up engage and they actually
12:49tell more people about it or they sort
12:51of invite people I don't even have
12:52strictly just in a viral sense but it's
12:54kind of generally speaking just that
12:56kind of lead to and generate more demand
12:58such a thing soon to be sort of a sort
13:00of in some super linear fashion kind of
13:03growing and I think when you sort of
13:06like being kind of less than one is just
13:09very different to being more than one
13:10and it sort of seems again when we
13:12launched I mean that didn't generate
13:14that many users I mean I don't think I
13:16could remember but let's just pretend
13:17that sort of 500 businesses signed up on
13:20the day we launched butts of immediately
13:22those 500 businesses told other
13:23businesses and people heard about it all
13:25the rest so kind of the next day we had
13:27a lot of businesses as well and and so
13:29such that from the day we launched the
13:31challenge became keeping up with demand
13:34rather than tweaking the product to
13:36somehow serve the market when you
13:39watched how many people were you at at
13:41that point employees we're about 10 so I
13:45guess before you launch then but your
13:48day to day sounds like it was just like
13:50we were talking about earlier just
13:51running around talking with users and
13:53fixing issues what was that literally
13:56every single day was like that or what
13:58were you doing everything okay so not
14:00literally every single day but I would
14:02say and we really tried very hard to
14:06understand into the very granular detail
14:10what exactly it was that people were
14:12doing you know where they were tripping
14:14up and so on it's just to give you some
14:16examples we had a chat room for sort of
14:19providing support when people were
14:21integrating stripe and it was it was
14:24actually a public chat room which kind
14:26of had some downsides because you know
14:27if we'd broken something or if somebody
14:29had some kind of embarrassing issues
14:30sort of everyone would see it but we
14:32thought that was kind of good because it
14:33would actually can put more pressure on
14:34us to sort of have the product to be
14:35good and we had literally every time a
14:40sort of for the first I don't know
14:41called 10 years of stripe
14:42every time somebody sent an API requests
14:44to stripe like it sent an email to us so
14:49we were like looking at every single API
14:50request and you know if we saw users do
14:54something weird why did they do that
14:56right or kind of where were our Doc's
14:57confusing or whatever and say you know I
14:58get a dinner in the evening or something
15:00and I've but you know well again it
15:02seems like a lot that I might've you
15:03know maybe 10 email list because stripe
15:05was not handling a lot of API requests
15:06back then but but sort of you're
15:07literally looking at sort of each
15:09individual action and actually stripe
15:11and we just kind of celibate celebrated
15:14or at least pass I'm very good at
15:15celebrating but we passed our seventh
15:17anniversary just you know a few days ago
15:22and so I was looking at our sort of we
15:25had an hourly stats email that we used
15:27to send and and so on the day we
15:29launched we handled sort of 22 payments
15:32in the previous hour which again sort of
15:35seemed huge to us back then but I was
15:37noticing in that email I actually
15:38forgotten this that we had things like m
15:40in the in the email we literally had a
15:42list of businesses that had submitted
15:44three or five or something API requests
15:48but it's sort of not gonna if they did
15:49not launch their businesses wheeler just
15:51had the emails of all these businesses
15:53the idea was that we would then kind of
15:56we literally individually reach out to
15:57them it's like hey you can have seemed
15:59you strive a little bit but you know you
16:00didn't I didn't go live like did we
16:04screw up or you know was the product
16:05somehow inadequate or whatever we did
16:07things like every time anyone ever hit
16:08an error of any sort that generated like
16:12a high priority email to us and so it
16:15sort of try to immediately go Salvage
16:16and that also kind of generated sort of
16:18I think this M this Pleasant kind of
16:20user experience for I mean you know it's
16:22frustrating when you hit an error in
16:23some service but we could then often
16:25sort of 15 minutes later reach out to
16:27them and say hey we saw the tubes have
16:29been counted the problem it's it's all
16:31fixed now and sometimes we did even in
16:32the case were sort of them the users
16:34made a silly mistake and that you know
16:35they'd be kind of mistyped an API
16:37parameter or something and we just reach
16:38out to them they're like hey noticed you
16:40had a typo in your code which you know
16:42perhaps and some of them is a little bit
16:43unsettling but but but at least gonna
16:47helped us you know increase the kind of
16:48the throughput of product feedback but
16:50so I mean these are all kind of examples
16:51I would say of the sort of general
16:52pattern of sort of really trying to be
16:55kind of hyper attentive to all the micro
16:57details of sort of what people were
16:59doing the product and kind of iterating
17:00rapidly in response to it and generally
17:03speaking I think that sort of pre
17:04product market fit metrics are actually
17:06relatively unhelpful and I think you
17:09should have you really want to buy us
17:10very strongly towards kind of as much
17:12sort of inspection and kind of high
17:16throughput qualitative feedback as
17:18possible because probably no not that
17:20many people are using your product right
17:21and so you know if it's 20 users you can
17:24kind of in some sense afford to just
17:25like look at everything they're doing
17:26and try to understand kind of you know
17:29what's working what isn't in some sense
17:31it's how much do they you can tell
17:36actively like how much they love your
17:37product how much are they gonna probably
17:39be really upset if it just disappears
17:42totally and actually on this point we
17:44had a thing at the bottom of every
17:45webpage we had a little sort of text box
17:48and kind of Anchorage to the bottom of
17:50the of the sort of thumb of the browser
17:53frame and and sort of a you know kind of
17:57one line high sort of text input and but
18:00we had placeholder text in it I had to
18:02sort of try to prying people to tell us
18:04things and so it had you know my my
18:07favorite part of stripe is ellipsis and
18:10you know people just fill it out I'm
18:12gonna hit enter now but we also of
18:14course had you know MOOC most of the
18:15problems were negative it was you know
18:17like the worst thing about stripe is or
18:19the worst thing that this page is or I
18:21really hate the way stripe does or
18:24whatever and and again we would sort of
18:25sort of at that stage you have to be
18:27kind of masochistic and so again we'd be
18:29sort of always waking up to you know all
18:31these emails you know telling us all the
18:33terrible things about stripe but that it
18:35was you know if I helpful to-do list for
18:36the day ahead how did you stay happy if
18:39look in the early days I mean for every
18:41story complaints what makes you think we
18:42did so it was it was what happy is such
18:49a squishy concept right and because like
18:54there are lots of things that we are I
18:56guess when I look back and look maybe
19:00this is the rationalization I taught
19:01myself but when I look back and sort of
19:03through life other things that I'm sort
19:05of most glad I did I wasn't exactly
19:08happy well doing them like often I was
19:09very stressed out or you know I had to
19:12work really hard or whatever but to the
19:13things that kind of post hoc brought the
19:15most fulfillment and and you know I
19:21think that well you know part of it
19:23there's a rich literature here and so I
19:24won't kind of died too deeply into it
19:26but I do think kind of happiness is a is
19:29a tricky concept to kind of pin down
19:30because you know is that kind of
19:33happiness in the moment is it sort of
19:34the sense that you have a year later
19:36looking back and so on and so I think
19:38and you know language is squishy and
19:39it's not you know completely
19:40specifically defined but I think kind of
19:42generally speaking kind of a better
19:43utility function and kind of gradient
19:45reddit line is that of fulfillment and
19:49and so I would take an experience of
19:51doing stripe was it was not especially
19:54happy because we were sort of you know
19:56constantly sort of incredibly aware of
20:00all the ways in which the product was
20:01severely deficient and you know all the
20:03challenges we faced and I mean it seemed
20:05like there was no FinTech category back
20:07then it was just like two teenagers or
20:10you know just just over teenagers trying
20:11to compete with PayPal which you know
20:13many people told us was not an
20:14especially kind of promising Avenue to
20:17pursue and so not especially happy per
20:22se and but it did feel kind of
20:24fulfilling and like I enjoyed working
20:26with John and the kind of people were
20:27separately hired and it was really fun
20:30working with the kinds of kind of
20:31customers we were serving and they're
20:33just sort of businesses doing all these
20:35kind of wonderful things and they were
20:36kind of really smart people and it felt
20:38that like if it worked it could be kind
20:41of consequential in the world and so I
20:43would say kind of it's it's sort of felt
20:47like I mean well you know I don't know
20:50what it feels like to be sort of a
20:51scientist or something but I'm guessing
20:52when you're sort of pursuing all you
20:54have this kind of big question and
20:56you're appreciating all these kind of
20:57haven't used to sort of try to better
20:58understand - and I'm guessing day to day
20:59it's sort of not especially happy
21:01because most your experiments don't work
21:03or whatever but sort of in for perhaps
21:05there's some analogue there in terms of
21:07them it still feels like it was in some
21:09sense meaningful yeah they're in some
21:12sense I mean it did add a like you said
21:14you just you're running around with your
21:15head cut off but maybe on a weekly or at
21:18least monthly basis is just taking a
21:19step back and seeing like what progress
21:21have I actually made because week to
21:23week it's like one percent here one
21:25percent here it doesn't seem much but
21:26from a monthly basis or you know maybe
21:29longer than that it it seems like
21:32ya know I think that's true and I think
21:33III don't know like I think it's almost
21:37took me a bad idea to sort of work on a
21:39company or to work on anything if if
21:40you're like truly unhappy working right
21:42and so there's kind of charities and
21:48and I think I mean a lot of well they're
21:51sort of so many differences of them
21:53difficult judgement calls to try to make
21:55in a start-up but you know of course
21:56part of it is like if something is
22:01making you unhappy or if it just doesn't
22:03seem like an especially promising Avenue
22:04or it's not really working whatever like
22:06your time has relatively high
22:08opportunity costs like you don't get to
22:09start that many startups in your life
22:12right and so kind of knowing when to
22:13sort of call it quits I think it is
22:15actually I mean sort of you know in
22:17start-up them we really extol and serve
22:19uphold the virtues of sort of you know
22:21determination at all costs you know
22:23never quit and so on and you know that's
22:25clearly not the right answer and that's
22:28like sometimes you should quit and so um
22:30so I think how much of getting it is
22:32true and right for I think that does
22:34need to be some sort of some happiness
22:36satisfaction fulfillment this good book
22:40of that of that name satisfaction which
22:42I recommend kind of week to week month
22:44to month if you're on the right
22:46trajectory oh man I wish I'd known that
22:49I could have you know what I remember
22:52when I was trying to when I was about to
22:54put stripe myself I had these bad
22:57memories of trying to implement PayPal
22:59and like stacks the documentation and
23:01taking like five days to figure it all
23:03out and so I sat down I was like it's
23:05because you guys are saying oh it took
23:06you ten minutes it's like no way and it
23:09probably took me actually five minutes
23:10at that point and I was I was super
23:14happy I mean that's like but maybe I
23:17should have sent you a review yeah
23:19especially if the review noted sort of
23:22criticism for us but no it's definitely
23:26helpful to have competitors with not
23:29very good products yes okay so you were
23:34in the early days you were doing a lot
23:36of stuff yourself at some point I guess
23:39too two-part question one is one of the
23:41things that you weren't good at and then
23:43too at what point did you hire someone
23:45to what what I assume is to do those
23:48things well I mean I'm I'm not that good
23:56at most things right and I say there's a
23:58non sort of false modesty or
24:01I don't mean artificially
24:02self-deprecating but said for almost
24:05anything that stripe has to do those are
24:06the people who were better at that than
24:09I am and so I think to some degree sort
24:11of the story of say you know product
24:13market fit and to you know say our
24:15current stage is kind of implementing
24:19the recognition of that in kind of more
24:21and more areas that said I think maybe
24:25sort of embedded in that question is you
24:28know okay well acknowledging that you're
24:35playing out the world's expert in in any
24:37single thing it's kind of in what order
24:39do you sort of replace yourself and for
24:43stripe the sort of the the important
24:47thing they were sort of most obviously
24:49not that great at was kind of
24:51partnerships and you know working with
24:53the various banks that we had to sort of
24:55get on board and and so on and so kind
24:58of we made them in fact Geoff Ralston
25:01who is here in the room today I think
25:06sort of was you know some combination of
25:10sort of very supportive and perhaps in
25:12the back of his minds of slightly
25:13pitying you know where he saw us kind of
25:15trying to get these partnerships with
25:16these kind of big banks in place and we
25:19really weren't sort of getting and they
25:20were very fast and so he told us that we
25:23had to hire this guy Billy Alvarado and
25:26I told us just don't ask me questions
25:28just hire this guy and at the time
25:32everyone everyone at stripe was an
25:35engineer and so we just we couldn't
25:37quite understand what a non engineer
25:39would do like you're not earning code
25:42what I was you know is there and and so
25:45we were kind of suspicious of this idea
25:47and and we sort of we went back to Jeff
25:49we met Billy he seems like a wonderful
25:50guy we really liked him but we couldn't
25:52quite get past this okay but like what
25:54does it actually how does this work out
25:55in practice and so Jeff told us that we
25:57should hire Billy and it didn't work it
25:58over the first few months or whatever
26:00that if it didn't work out Jeff would go
26:03back and pay his salary so it's kind of
26:04a zero risk move for us and and so and
26:09we didn't have a whole lot of money at
26:11the time so that was not insignificant
26:15we heard Billy and that was sort of an
26:18immediately trajectory changing event
26:19for you no previous leaving me gone and
26:21talked to a bank I mean I know they're
26:22literally doing this but it certainly
26:23felt they were kind of pushing the
26:24security button under their desk it's
26:26like what are these guys in my office
26:27and and then suddenly things start to go
26:30kind of much more smoothly and and Billy
26:34is still with stripe today and day has
26:37been an enormous ly kind of the pivotal
26:39presence and so that was um
26:41turned out to be very helpful advice
26:43that is interesting just as an
26:45inflection point for stripe never new
26:48cool so it sounds like you had hired
26:50actually people before Billy I guess
26:54first your very first hire the third
26:57person between with you and John it was
26:59an engineer I'm assuming like well he
27:01actually wasn't an engineer but he
27:03joined stripe and he started to become
27:05an engineer so like he'd written a
27:09little bit of code previously but he
27:12joined a lot of codes I had to written
27:14it was a guy I'd known from college he's
27:16actually also Irish and yeah we love
27:21code to be written and so he was the
27:23kind of person who you know would sort
27:24of survey that landscape and just do
27:25whatever it was was you know most
27:28important and required and at that time
27:30I was writing code and so he went so
27:35kind of the transition from pre pardon
27:37mark fit to post product market fit I
27:39want to see use when they think back
27:41this is like the one point in which they
27:43think I could have done that better so
27:46how do you how do you create your
27:47success of that transition because a lot
27:50of it is you know just taking stuff
27:52delegating stuff better and doing other
27:55functions of the business so what
27:58exactly was your transition like and you
28:00know how would you how could I if I was
28:03going in your shoes tell that I was
28:06doing it well yes I I don't think I did
28:08it especially well and I think it kind
28:11of you know fortunately sort of you know
28:13it worked out but I think if if I was
28:15doing it all over again I think we could
28:17have sort of accelerated ourselves by a
28:18year or two if we'd sort of gone about
28:20it in a somewhat more disciplined and
28:21kind of self-aware fashion I think the I
28:26think one of those kind of pernicious
28:28a sort of mental models you can have
28:32here is the kind of you are on some
28:35growth curve and it is sort of your job
28:37to sustain or you know marginally
28:41inflect upwards or kind of somehow and
28:43you know that much of them is curling
28:46the sport where you're kind of you know
28:47wiping the ice while the overt is the
28:50weight a precedes down along it sorry
28:53I'm not Canadian but um and but kind of
28:56somehow you're kind of making kind of
28:58these very small interventions and
28:59perturbations on some underlying growth
29:01curve and I think that's an easy mental
29:03model to have and I think it's kind of
29:04you know actively kind of fallacious and
29:07mistaken I think I have a much better
29:09mental model to have is you're serving
29:11some market and that market is I mean
29:14for its relatively finite in size I mean
29:16you can always you know change the
29:18project and increase the market size but
29:19sort of you know you can you can think
29:20of it as being finite and then there's
29:22sort of the percentage of the market
29:24that you're serving and then you know
29:27whatever percentage you are not serving
29:28is kind of well you just haven't built
29:30the sort of go to market functions and
29:33organization that's kind of capable or
29:35at least has yet sort of brought the
29:37product to those market segments and the
29:39kind of the growth curve or adoption
29:42curve is just kind of a function of of
29:44that go to market apparatus but
29:46basically it's it's it's not some kind
29:48of cosmic trajectory it's something kind
29:50of very much of your creation and under
29:52your control and so what I what we did
29:54not do but what I wish we did is kind of
29:57maybe whatever you know just after we
30:00launch me there's a whole bunch of kind
30:01of immediate scaling work we have to do
30:02but say six months after launch that we
30:04should have sort of mapped out the kind
30:06of concentric circles of our market like
30:08maybe there's kind of you know very
30:10early-stage startups there were kind of
30:12a you know our sort of initial initial
30:14constituency then there's kind of all
30:16technical startups but not necessarily
30:19very early stage and it's kind of I
30:21don't know you you you keep going these
30:23kind of successive increments until
30:24eventually got to say all companies
30:25handling online payments or something
30:27right and I'm gonna sort of figure it
30:29out okay kind of what's the size of this
30:31market sort of you know kind of what
30:34fraction of are we currently serving
30:36what would it take to serve more and so
30:38on and I think it's quite striking you
30:39see that kind of when and repeat
30:42start companies they almost invariably
30:44are willing to kind of build the
30:45organization you know post prime market
30:48it so it's invariably willing to kind of
30:49build the organization ahead of where
30:51things are today and which I think is
30:53exactly the right thing to do because
30:54they're thinking okay I have the right
30:56product now that's gonna work backwards
30:58from okay what would the organization
31:00looked like that was serving the entire
31:01market and let's just start building
31:02that organization because again the
31:04growth curve is under my control
31:06and you know of course it's not like
31:07100% under your control but I think it's
31:09much more under your control than then
31:12things that people don't think there's
31:13also of course a difference here a major
31:15difference when you can use or I think
31:17about sort of consumer versus kind of
31:18b2b use cases where consumer it's a bit
31:21more I mean people don't necessarily
31:22know exactly what they want you know
31:24what's the the market size for like a
31:27news reading app or a dating app or
31:28something I mean it's it's a kind of it
31:30depends a lot on the product like you
31:31know maybe when more people should be
31:32reading news or something but for fer
31:35sort of a for some service or a product
31:38that sort of serves a discrete logical
31:41concrete function that sort of set of
31:43businesses or entities to finish early
31:45have or don't have I think it's kind of
31:47much more sort of much more rational and
31:50much more mapable and I partly had this
31:52epiphany Vienna Aaron Levie who's the
31:56CEO of Box John I eventually we started
32:00an Palo Alto we moved up to San
32:01Francisco I don't well Aaron Levie um
32:04had Facebook message on and and we've
32:08never heard of him and heard of most
32:10people in Silicon Valley and but he'd
32:11Facebook messaged us like very early on
32:13asking to invest in stripe and I think
32:16John didn't know this is a random guy
32:17and so we never replied and but then you
32:22know we heard of Fox and we heard of
32:24Aaron and you know we've read his funny
32:25tweets and all the rest and so we moved
32:27to San Francisco and and we invited him
32:30to our housewarming I think was the
32:31first time we'd ever met him and sort of
32:34you know like we're not very good at
32:37partying and so you know by a sort of
32:3911:00 p.m. or you know midnight or
32:41something kind of everyone was going
32:42home and but Aaron was still there and
32:45an Aaron state owns like 2:00 a.m. I
32:47still remember kind of sitting in our
32:48front room telling us how much better at
32:51woz be building b2b software than
32:53consumer software for this reason our
32:56for it's so hard to predict sort of what
32:59people want they don't even know
33:00themselves what they want you're kind of
33:01you're it's such an amorphous space
33:04whereas when you're selling software to
33:05businesses you know businesses are kind
33:07of are mostly rational are mostly the
33:11opposite of inscrutable is scrutiny I
33:12guess and and you can sort of work
33:15backwards in a way that sort of it's
33:20just much more comprehensible and so um
33:21I still have this kind of you know it's
33:24like the angel and demon on your
33:25shoulder I still have sort of you know -
33:26am Aaron Levie sitting on my shoulders
33:28of extolling the virtues of building
33:30software for entities that know what
33:33they want you've made the right choice
33:35yeah in some senses there's a lot less
33:37variables or in consumer there's a lot
33:39more variables to to consider and
33:41they're quite unknown I want to take a
33:44step back and you talked a little bit
33:46about you know thinking about thinking
33:48ahead about what your team is or what
33:49your company structure is going to look
33:51like how do you maybe this is too big of
33:54a question so I mean we can whittle it
33:56down a little bit but how do you think
33:57about that like like I'm sitting if I'm
33:59starting a start-up today I am close to
34:02maybe a product market fit but before
34:04that stage like am I thinking about this
34:06sort of thing i watch my team look like
34:07what should my culture be and stuff like
34:09that I think that then well okay so pre
34:14prime market fit um I mean the goal is
34:19really is just to gas to product market
34:21fish and so I actually wouldn't bother
34:23thinking too much but all these kind of
34:24distribution mechanics questions now you
34:26can get probably a good fit sort of
34:27relatively quickly and so kind of that
34:28pre-primary at this stage might only
34:30last six months you know you're not
34:31necessarily like stripe or you're gonna
34:33in it for various reasons for first of
34:34multiple years and so it may not last
34:37very long however until that point I
34:38really think you should just be think
34:39about okay how do I get there the main
34:41thing that I think companies screw up at
34:42that prepartum are could fit the sage
34:44and is sort of speed of iteration I
34:47think okay well what determines or you
34:49kind of speed of fruitful iteration and
34:51and that sort of I mean if you're kind
34:55of completely or if you're repeatedly
34:57rebuilding our product but sort of
34:58national response user feedback I mean
35:00that's that's kind of far less likely to
35:02be kind of bringing you in sort of a in
35:06a productive direction but if you have
35:09some kind of meaning
35:10albeit perhaps smalls of initial set of
35:12users and your rapidly iterating in
35:14response to sort of their feedback and
35:16sort of observed behavior and so on then
35:18I think that's like a really good spot
35:20to be in and I think again at that
35:22juncture a pre prod market fish it is
35:23kind of this you should be sort of doing
35:26everything you can to tighten that sort
35:28of feedback cycle and there's a fighter
35:32pilot in who kind of revolutionized
35:34airborne combat in the US and said of
35:36the second half the 20th century so
35:37Korean War onwards and called boid and
35:40he had this concept of the OODA loop
35:42which was sort of a sort of a similar
35:45notion I was the most improve you see
35:47people thought you want to just like the
35:48fastest aircraft or said that you know
35:50most sophisticated weaponry and so on
35:52where he was all about sort of no you
35:54actually want an aircraft that instead
35:56of pilots and training that are really
35:58oriented around sort of the fastest sort
36:00of responsiveness in iteration to the
36:02cut particulars of the situation in a
36:04way that kind of subsequently went on to
36:06really inform sort of modern aircraft
36:08design and so I think you want to be
36:10like and one of these sort of these
36:12modern fighters every year sort of
36:14really optimized to respond as quickly
36:16as possible to sort of rapidly evolving
36:18situations again and it's kind of
36:20pre-primed market at stage and so then
36:21from a hiring standpoint I think I think
36:23it should really be about okay well
36:24what's gonna sort of what's going to get
36:26you there faster and I mean I think it
36:28an early stage it's most likely well
36:30sort of just people who help you kind of
36:32build the product but of course not too
36:34many because I mean at some point like
36:37not be able to do more but you might
36:39actually be less responsive because
36:40you're bigger team to manage or
36:41something right and so as an empirical
36:43matter it seems that somewhere between
36:45kind of you know two and ten depending
36:48on exactly what you're building is kind
36:49of the optimally responsive size but but
36:54I think it really is about that sort of
36:56him time from observed sort of necessity
37:01or deficiency or just you know
37:03characteristic of your users behavior to
37:06executed fix or improvement and whatever
37:10it is that kind eyes is that is there I
37:12mean you say two to ten which is helpful
37:14but is there are there observations or
37:19evidence in which you have hit the peak
37:22like you should not add an X like this
37:24extra person is going to be negative
37:26value add to the whole operation yeah I
37:31mean every person every person well
37:35startups in general involve all these
37:36kind of impossibly difficult of judgment
37:39calls in this kind of you know high
37:41dimensional possibility space and so it
37:44would be great if you could sort of
37:45collapse it down to kind of a fairly
37:47simple set of trade-offs that is I don't
37:48think you can however I think in
37:49principle the calculations that are
37:51trying to run is ok and each successive
37:53person takes time to hire and so slows
37:55us down in that regard takes time to
37:57unmoor slows that in that regard kind it
37:59takes time to understand meld with the
38:00sort of culture and learn the stack on
38:02the stuff that that also takes constant
38:04time then involves subsequent ongoing
38:06cost of just like coordination and
38:08alignment and sort of you know you you
38:10should now have you know the
38:12organizations that distributed across
38:14more neurons and and so there's that
38:16kind of persistent tax and that's not
38:18just necessarily a linear cost but it
38:19you know it can be sort of you know
38:20quadratic or something kind of given you
38:23know the multi way communication problem
38:24so you of all these costs to an
38:25additional person so the question is
38:27okay but this person can also help us
38:29with like get more done right they
38:30can write more code or talk to Moe users
38:32or whatever and so it's kind of is that
38:34fixed benefit to sort of that additional
38:37person worth all these other costs and
38:39again with I think the ultimate arbiter
38:41being will we be more responsive to what
38:44we're learning about our users given the
38:45presence of this additional person and I
38:47think you know whether or not you will
38:48be dependent the complexity the product
38:50and the complexity of the market and you
38:51know all that stuff but but I think
38:54that's in principle the calculation you
38:56might be running I've heard you said and
38:58speaking of hiring people I've heard you
39:00said you know the key qualities that you
39:02look for in a future stripe employee is
39:04intellectually honest cares a great deal
39:07and just loves getting things done which
39:09are great attributes because some people
39:11just don't fit in those categories so
39:13it's good to have that separation there
39:15how do you when you meet someone how do
39:17you figure out this is the right person
39:23it's very hard and and I wish I had a
39:27more sort of a sort of definitive rubric
39:30and kind of particular set of questions
39:33although if I had a particular set of
39:34questions maybe you know they'd you know
39:36they'd stop working because kind of
39:38would learn how to how to game them or
39:40something and so I mean it well it's
39:45hard to fake just being smart right
39:47and so that one's kind of not as hard to
39:51discern and and I it's oddly hard to
39:59fake being intellectually honest and and
40:02so the characters who kind of being able
40:04to see multiple sides of a debate or an
40:07argument like there are so many kind of
40:09complicated questions for so the only
40:11thing that I'm really skeptical of is
40:13kind of certainty in either direction
40:14just because you know the questions are
40:16intrinsically sort of involve very
40:20contingent trade-offs it's also it's not
40:23impossible but it's hard ish to fake
40:27just being nice and like something
40:29something we care a lot about it's
40:30stripe is is just sort of people who are
40:32pleasant and warm and sort of just you
40:35know make it others happier as a result
40:37of their presence and you know if well
40:43you know perhaps if you can fake that
40:45perfectly forever you know that's fine
40:47right but but yeah they're all and sort
40:54of actually there's no clear rubric for
40:59them and I'm not sure that the peer
41:00rubric could even work yeah so you talk
41:05about like get people who help move the
41:08organization faster if at all possible
41:10and I think there are two issues usually
41:13that start slowing down the organization
41:16as you start scaling just cuz each
41:19person adds just more complexity than
41:21organization but also I think one is
41:23asymmetric information just people don't
41:25want that guard never on the same page
41:27and I think you've talked a lot about
41:29this I think and there so everyone can
41:31google around for lectures you've given
41:32on this on you've done email
41:34transparency you know obviously you know
41:36having metrics transparent to the home
41:38ization but the second problem to that
41:41is if you fix that there's also this
41:43asymmetric interpretation problem which
41:45is everyone's there's just like black
41:48box function that how people interpret
41:49eight even if all the inputs are
41:51the same the outputs are all different
41:53and it's official at your skill now it's
41:55nearly impossible to figure out
41:57everybody's function so when you're
42:00thinking about creating your
42:02organization and building it out like
42:04how do you reduce that reduce that noise
42:07and and try to get everyone on the same
42:09page yeah and well the first thing is I
42:15wish we had actually between say five
42:20and fifty people and I think we were
42:23consensus oriented we of course weren't
42:25completely consensus oriented I mean we
42:28couldn't have gotten anything done if we
42:29were but I think we kind of biased too
42:31much in that direction and again I've
42:33seen a it's just not that efficient
42:35right and it's sort of it's it's
42:37necessarily not the case and so kind of
42:39you like you can sort of perhaps
42:42maintain some kind of fiction for you
42:44know more or less time but but sort of
42:46ultimately speaking the third there is
42:48no way of sort of sustaining that and I
42:52think that's a relatively common mistake
42:53right and because you know you almost
42:55invariably come from so some kind of you
42:57know n Musketeers for so small and sort
43:02of kind of orientation whereas it is
43:05there's no particular need for sort of
43:07formal decision-making mechanisms or
43:09sort of you know subsequent
43:11communication of said decisions or
43:13whatever right but then you can hit
43:14fifteen people and now there is and come
43:16I think companies don't adjust quickly
43:18enough to sort of that new necessity
43:20again very much as included and so I
43:22actually think we did relatively well on
43:24the kind of ambient availability of sort
43:28of context and information so on but
43:30actually kind of in some sense poorly at
43:31sort of deliberate explicit
43:33communication of of decisions decisions
43:37being I mean actual kind of tactical
43:39decisions or even you know bigger
43:40decisions like you know what are our
43:41priorities the next six months or things
43:43of that nature and so the higher order
43:48piece of advice and perhaps I'm over
43:50extrapolating from our personal
43:51experience but Hydra advice just kind of
43:54reflecting back would be to kind of it's
43:56kind of like the pre and post pop
43:57probably market fit thing it's sort it's
43:58actually about the size of the
43:59organization we're kind of when you hit
44:01a certain size again maybe I'm just
44:02gonna say ten people for the sake of
44:03simplicity a bit before it's a bit after
44:05I think you to kind of adjust more more
44:07deliberately to this kind of explicitly
44:09eunuch ation model of sort of of being
44:12sort of quite quite firmly non consensus
44:16based and and sort of I mean nobody
44:19nobody likes the idea of sort of being
44:21hierarchical that sounds pejorative but
44:23like in some sense hierarchical sort of
44:26the top-down like it does move things
44:28faster maybe some people don't like that
44:30yes it does that's right that's right
44:33and there is this can delicate balancing
44:35act you you want to sort of orchestrate
44:39we're kind of on the one hand you want
44:42to sort of really prioritize speed and
44:44agility which involves being kind of
44:46somewhat a hierarchical because those
44:47are sort of efficient sort of symmetry
44:50breaking mechanisms and ways to sort of
44:53have shots get called but on the other
44:56side you really do want people to have
44:57this kind of strong ownership mentality
44:58and and real sense that sort of they
45:02they can cause think things to change or
45:06identify problems or sort of inject new
45:09ideas even in unrelated areas and so
45:12it's this a delicate act where I mean
45:14you definitely can be excessively
45:15hierarchical but sort of how do you sort
45:19of facilitate enough autonomy and agency
45:21them but also not have things devolve
45:23into this kind of I don't know sort of a
45:27uniform morass of sort of all these kind
45:31of of a Brownian motion and and and I
45:36think that's hard and requires all these
45:38kind of you know ongoing sort of tweaks
45:40and nudges and I mean is dependent on
45:42the personalities involved and all the
45:43rest and again I you know I really wish
45:45there was kind of the sort of wrote
45:47simple I'll just do XY and Z and you'll
45:49be good but sort of if such XY z-- and
45:51Z's exist you know no one's told me yet
45:54if only there was a formula for
45:56everything so riding on the same theme
46:00you know as stripe grows you strike me
46:02as a company that does the opposite of
46:04most companies which was they get bigger
46:06they just slow down and they get less
46:08innovative but for you guys and it's
46:10hard to even keep up like you're just
46:11pumping out new products and which seem
46:13to be successful to me like stripe Atlas
46:15try press stripe terminal so and so
46:19early days you know when there's just 9
46:21or 10 of you if someone had a new idea
46:22it's probably really easy to get to you
46:25and you know then you guys would decide
46:27how has that changed over time to make
46:29sure that someone who's like six level
46:32six degrees away from you that a good
46:34idea actually gets to you and then how
46:37does that decision made to actually
46:38execute on upon it um well I'm it's very
46:44nice to you say that I feel that we're
46:45getting faster as we get bigger because
46:47you know we're constantly sort of
46:49self-flagellating over how like freaking
46:51slow we are and like paranoid about sort
46:55of I don't know degenerating into some
46:59kind of you know sort of immobile stupor
47:03and so to whatever extent we do get
47:08things done or or appear fast I think
47:10it's largely because we we're very
47:15paranoid in this dimension and and I
47:17think that kind of the the default
47:19outcome of companies of course as they
47:20scale is to become sort of more ossified
47:23and rigid and sort of closed to new
47:29ideas and directions and things that
47:30contradict their prevailing orthodoxies
47:32and all of that as how we kind of avoid
47:37that and I mean well there's lots of
47:42kind of obvious things like partly it's
47:44it's sort of I mean having leaders who
47:46care about the rate of progress and and
47:48just love seeing things happen quickly
47:50and lots of things of that nature
47:52and I think maybe sort of deeper rooted
47:55as we try to be a kind of yes and
47:58culture in that I mean I personally love
48:03just ideas and potential things we could
48:06do and I mean most of them are or I mean
48:10no matter how good the idea is sort of
48:13we should not pursue most ideas right
48:14even if independently it could be a
48:16super successful comfort or something
48:17but like there is a fairly finite number
48:19of things we can do and so sort of you
48:21kind of on the one hand need to
48:22recognize that intellectually and just
48:23like not pursue most things while on the
48:25other hand enjoying the exercise of
48:27contemplating well what it looked like
48:28if we did it right and so you know one
48:30thing we do every year for example is we
48:33crazy ideas process.we college where we
48:35literally sent a document to the whole
48:36company a hackpad document that that's
48:39open to everyone and people can sort of
48:42add ideas to it that they think are
48:44probably a bad idea but if they worked
48:47might be a great idea
48:49and it's very important that they have
48:51to be probably a bad idea because if
48:53they're like probably a good idea then I
48:54mean it's not that risky and I mean kind
48:57of by definition we should ply do it and
48:58so we know whatever it may be just do it
48:59and so people like really self-censor a
49:03lot in most organizations because they
49:05don't look stupid and they don't be sort
49:08of associated with just having all these
49:10like wacky bad ideas and so on and so we
49:13try to be fairly firm that if you add
49:15ideas it has to be probably a bad idea
49:17and and actually a lot of cool things
49:20we've subsequently got on and done the
49:23first came from the process it's try
49:25Pappas you know being one of them
49:28and and also we we sort of helped
49:33stellar it could off the ground this
49:34sort of cryptocurrency back a few years
49:36ago instead of that also sort of came
49:38from this process and so that's one of
49:41the mechanisms by which we try to have a
49:42kind of yes and culture and I really
49:46think there aren't that many as just as
49:51an empirical matter and there aren't
49:53that many large organizations that I
49:55think sort of really do this
49:56successfully but I think that sort of
50:00the the most successful of larger
50:03organizations somehow do succeed and
50:06sort of this iterative repeated repeated
50:10process of kind of augmentation
50:13you know Amazon and Google being have
50:14the most prominent examples for you know
50:17obviously even Google launched there was
50:18no Gmail or YouTube or Android or Google
50:20Maps or whatever and you know Amazon
50:23sort of is kind of an even more
50:26conspicuous example in some sense this
50:28kind of repeated attach of successful
50:29ancillary businesses and so I think kind
50:31of yeah there's a very natural
50:34temptation as you grow to I think become
50:36increasingly close to this I think is
50:37very important to avoid cool so I have a
50:40couple more questions and then maybe
50:41we'll have time to take questions from
50:42the audience one is are looking back
50:46is there something did you have a strong
50:49opinion of you know how startups should
50:51be run as a CEO that have just
50:53completely reversed because you're now
50:57to see you it's a good question and
51:08another way to potentially put it is
51:10like what are what are things that you
51:13did were you're like I know for sure
51:14this should have been done and they just
51:16turned out to be a mistake well I
51:25already mentioned the set of the
51:27consensus orientation when they were
51:29going through right now which is a quite
51:32sort of significant divergence is a you
51:35know we were sort of fairly centralized
51:37and up until sort of relative recently I
51:40mean we had some remote employees from
51:42pretty early stage but so by and large
51:44stripe was kind of concentrated in San
51:45Francisco a.m. last week we announced
51:49our fourth global hub and basically
51:53we've decided to do is we're going to
51:54have kind of major sort of product and
51:56engineering sort of efforts and teams
51:58and functions and all the rest in San
52:01Francisco also in Seattle also in Dublin
52:05and also in Singapore and and sort of
52:09these nonsensical locations they're not
52:11gonna be sort of I know operations
52:14offices or satellite offices or you know
52:17places where we do have some
52:18localization and local market adaptation
52:20we want to have kind of completely de
52:22novo new products that sort of become
52:24super successful started out of these
52:26offices and that's not the sort of
52:28standard obviously kind of Silicon
52:30Valley pattern and we're sort of Apple
52:32and Amazon and Facebook and Google and
52:33all these companies tend to be sort of
52:35highly concentrated in these sort of
52:37very I don't know so monolithic
52:39corporate headquarters and our thesis is
52:43that it's kind of several folds you know
52:45partly that kind of the availability of
52:46talent is becoming sort of far more
52:48geographically dispersed partly that
52:50sort of you know the Bay Area is going
52:52an increasingly untenable expensive
52:54location to locate and partly that sort
52:56of we really want stripe to be global
52:59infrastructure that work
53:00kind of just as well in instead of you
53:03know Asian markets or in Latin American
53:04markets or whatever as it does for
53:06businesses in the US and have the era of
53:08the internet being asleep predominantly
53:10North American or North America and
53:11Western European or whatever the the
53:13days of being kind of such a phenomenon
53:14are over and so I think that's sort of a
53:17fairly substantial break with kind of
53:19you know descriptive best practice of
53:22the past I mean you know we obviously
53:25think it'll work and we wouldn't do it
53:26if we didn't but it is also kind of them
53:28on some level risky like we don't have
53:31good you know prior examples to point to
53:33and you know I mean there are some great
53:35companies in Singapore like like grab
53:38and carousel but aren't really any
53:39examples yet a sort of American
53:41companies establishing kind of major
53:42private engineering hubs they're you
53:44know we're pretty optimistic it can be
53:45done but but you know kind of if it
53:48works you know we'll be the first or one
53:52of the first cool last question so in a
53:56hundred years from now what is stripe
53:57gonna be what do you imagine it to be
53:59well we're only seven years old so
54:01that's a difficult question to answer
54:05we're trying to build this kind of this
54:08economic infrastructure for the
54:10International is sort of platform for
54:12globalization and kind of technological
54:17progress in the sense that like it ought
54:21to be just as easy easy to start a
54:23company in in in in sort of you know
54:26Nigeria as it is in New York and it
54:30should be just as easy for somebody in
54:31Brazil to buy something so from any of
54:33these companies as it is for for again
54:35somebody in in the Western world and and
54:42it just seems so crazy to me that sort
54:43of that hasn't happened yet like I it
54:47sort of feels that stripe should have
54:48happened ten or twenty years before it
54:50did and and just kind of I don't so much
54:54think the were sort of pursuing this
54:55kind of unprecedented kind of path
54:57breaking I am sort of a inconceivable
55:03idea so much as kind of correcting a
55:05deficiency and instead of as a rip in
55:10the fabric of Internet infrastructure
55:12and and so anyway I think we still have
55:15you know at least five years to go
55:17instead of sort of correcting this
55:20inadequacy yeah so what happens after
55:23that I'm not sure in some sense in the
55:27future all transactions should be
55:29digital and they could very well it's
55:31just the all going through strife I mean
55:33right now like in the u.s. someone's
55:35telling me like 80% of all Americans
55:38have done some transaction Thursday
55:41right yeah that's right so in the last
55:4212 months more than 80 percent of
55:45American adults have bought something
55:47from a from a stripe business at least
55:49one thing from a stripe business which
55:52is cool and you know it's not just the
55:53case in the u.s. like in Singapore I was
55:55last week you know that that figure is
55:57about 70 percent right but I guess we
56:00don't think about it well I think about
56:02it more in terms of the things that are
56:05possible or get started as an its kind
56:08of we think a lot about the rate of sort
56:10of new firm creation what companies are
56:13getting started how successful of those
56:15companies which markets can they serve
56:17and so on and you know every now and
56:19again we look back and we look at sort
56:20of the those kinds of market coverage
56:22that's but I guess that's not really
56:23about sort of them what motivates us is
56:26it's more than sort of are these
56:28businesses that should exist and or
56:30should be able to kind of offer the
56:32present services you know in these
56:34places where they currently can't and
56:37that's I think I mean that's the kind of
56:39thinking that we use to kind of inform
56:40the product and that's it is kind of the
56:42the core on a locomotion day-to-day and
56:46then maybe the outcome of that is is
56:48that all these people get to benefit
56:50from it all right thank you so much
56:52Patrick this is great thank you
![How To Succeed As A NEW & YOUNG Realtor [Deals Every Month + Luxury Listings]](/_next/image?url=https%3A%2F%2Fi.ytimg.com%2Fvi%2FrtrRghgTayc%2Fhqdefault.jpg&w=3840&q=75)
