It is not often you get to interview a rising entrepreneur before they blow up in Silicon Valley. Fortunately we had the opportunity to catch Patrick Collison early on his way to stardom. A YC alumni and Irish hero, Patrick Collison of Stripe is out to tackle one of the hardest problems on the internet: efficient online payments.
Watch the full Startup Grind interview to learn more about how he plans to do it, and what made him and his brother John Collison take the plunge into one of the most difficult problems on the web.
Can't get enough of Patrick? Watch our more recent 2015 Stripe interview, and read Patrick and John Collison's Stripe founding story here.
Derek Andersen (Startup Grind) and Patrick Collison (Stripe)
[00:00:00.0] Hailing from the great country of Ireland, at 16 he was awarded one of the most prestigious science awards in the country. He attended MIT for a very short time then joined a Y Combinator company called Automatic which was sold less than a year later for $5,000,000. He has since founded Stripe and he is with us here tonight, so everybody let’s give a big round of applause for Patrick Collision. Here we go.
[00:00:39.9] That’s the introduction you get at work every morning, right? When you come in?
[00:00:45.1] Patrick Collison: Not quite. I'm sure. Someday, someday.
[00:00:49.4] So, like I said, we are, and I specifically am so excited to be able to have you here tonight. We had Steve Blank here a couple of weeks ago, and I think Stripe is one of these companies that we are really going to be hearing a lot about over the next couple of years, so I just really appreciate you coming.
[00:01:08.0] Patrick Collison: Thank you
Before Stripe: Patrick Collison's Early Years
[00:01:09.2] I'd love to just start and talk a little bit about your background. Tell us a little bit about your family and where you are from, and how you grew up.
[00:01:09.2] I'd love to just start and talk a little bit about your background. Tell us a little bit about your family and where you are from, and how you grew up.
[00:01:17.5] Patrick Collison: So, I grew up in Ireland. I guess it was a pretty standard Irish upbringing. I spent most of my life living in this really rural part of Ireland. It was surrounded by farms and fields and stuff. And I mean this was kind of problematic. I remember we didn't get a computer for a long time and you know when we got a computer it took a long time before we got Internet access. I remember renting or borrowing books from the library, and I'd read about the Internet because we couldn't actually get the Internet. So for a period of about two years I was reading these amazing books about the Internet and I found it so futuristic. Then we got the internet, but the phone line didn’t work very well and I remember going to these bandwidth test sites and I got .5K per second, and so the workflow was have a book with you while you used the internet because you could read a page of your book, look up to see if the page had loaded, and the chances are it hadn’t so you went back to your book. Then eventually for the last three years we managed to get this satellite internet connection because that was basically the only way you could connect in this really rural part of Ireland for the internet, and that finally worked well. That was 5 / 12K broadband but the latency was horrible because it would go out to space and then back down to Germany so, you have some like H session and then you would press a key and then two seconds later it appears. But at least the bandwidth was good. Most of my childhood was, I played with Lego, when I was really young, and then I jumped straight to computers. Most of my childhood was either misspent on Lego or computers.
[00:02:49.6] When did you start really using computers and programming?
[00:02:52.5] Patrick Collison: I picked up PHP like most of the world of teens, and I started that when I was 13.
[00:03:01.1] Tell me a little bit about your parents, Lilly and Denis, is that right?
[00:03:03.7] Patrick Collison: That's right.
[00:03:05.4] What were their professions? What did they do.
[00:03:07.8] Patrick Collison: My mom originally trained as a Microbiologist and, I guess, Ireland is, or at least was at the time I grew up, pretty traditional. I was the first child, and so when I was born, Ireland being such a Catholic country she was expected to leave her job and go raise her new kid. She has this wonderful letter from her employer at the time, congratulating her on her fine career and wishing her all the best raising her kid in the future before she had given any notice or anything like that. It was just assumed she would leave. She did anyway I think in part because she was so pissed off by that letter, and stayed at home with me. About three weeks after I was born she started a company, and so I guess I was that boring!
[00:04:02.6] Totally normal, recovery period that most mothers....
[00:04:03.8] Patrick Collison: I think that this just means that how un stimulating I must have been as a baby. So she started this corporate training company, and basically like ISO 9000 or Sig Sigma or all this kind of stuff. That company is still going today, and very much not a silicon valley start up. It has been profitable every year since it was founded, and now I guess it’s the same age as me, so I suppose 23 years later, I think it employs 12 or 13 people. That's my mum. My dad, dad trained as an electrical engineer and both my parents had very rural Irish upbringing. They grew up on farms. Dad was one of the first kids from the family to go to college and he went to work for Dell.
[00:04:53.8] In Ireland?
[00:04:53.8] Patrick Collison: In Ireland, yes. Then he decided, when I think I was about 5 or so, he was going to go and buy a hotel on the shores of this lake in the middle of Ireland. Which is how we ended up in this dire internet less language situation.
[00:05:09.7] It wasn't one of the upgraded features of the hotel I guess, there was no internet access there?
[00:05:12.5] Patrick Collison: Yes, it came later. It was great. There was none of these stupid pay $20 a night for the Wi-Fi things. There was just no internet.
[00:05:20.3] You have two brothers, your brother John who works with you at Stripe, co-founded the company with you, and your brother Tommy. Just talk about your relationship with each other. Are you very similar or are you very different?
[00:05:31.9] Patrick Collison: Yes, John and I are pretty close in age. We are two years apart, John is two years younger. We were, I guess we were similar ish growing up. I got more into programming, through school and John was more into Math and chess and stuff. But I guess we were both pretty nerdy, we both read a lot and we were pretty competitive actually. I think it was kind of helpful that I eventually went away to college in the US. I spent some time away from home, and I think that sort of allowed us to develop apart a little bit and then Tommy is five years younger than John and he's the one who is actually going to be famous and well known in several years. Maybe in ten years time people will....
[00:06:19.6] I think he's already famous.
[00:06:21.5] Patrick Collison: People will stumble across this interview because they are googling Tommy Collision, and then this turns up or something. He's had more Twitter followers than me, since he was like 13 or something, so mostly it's just embarrassing.
Digging into Hard Problems
[00:06:34.4] So you grew up in this home, entrepreneurial home really, and by your mid teens you start to enter these science contests and things like that. Tell me how that started, what was the motivation behind that, what happened, and tell us how you did.
[00:06:49.3] Patrick Collison: So, this mostly came out of two things. One is this kind of science competition was very well known and fairly kind of prestigious within Ireland, it got a lot of media attention and everything. It was fairly obvious that students would love to take part in it. A lot of the winning projects that came out of it where roughly very cool. A couple of years, or like when I was in primary school, which I guess means ten or something like that. A project won that was a much faster version of RSA, with some fairly significant photography research. The girl who did that ended up going to work with Stephen Wolfram Corporate Research, and there were a bunch of pretty substantial mathematics breakthroughs. Euphemism, and generally stuff that just like, it didn't seem sort of anyway kind of, I think some of these kind of competitions for high school students and stuff have a tendency to seem kind of condescending or junior. You were patted on the head for making a volcano model or something, and there was no sense of that.
[00:08:01.4] Sounds like one of the US ones.
[00:08:03.1] Patrick Collison: I've never been to any of those, but
[00:08:05.0] You were doing encryptology, it's a word I can't really pronounce very well, and we're doing the volcano.
[00:08:11.9] Patrick Collison: We're screwed.
[00:08:16.1] Patrick Collison: I was very taken with that anyway, and so when I got into programming, fairly quickly I started wondering if I could do anything with this. After PHP I tried to learn JAVA and I found it, I remember spending about a year trying to learn JAVA and just finding it so miserable. I basically concluded that I should just do something else and not be a programmer. Then I stumbled across Python, from some essay that someone wrote, and I was reading the Python mailing list or something, people were always referencing LISP as the inspiration for a lot of things in Python. like math and reducing whatever. So I went and looked into LISP and LISP just seemed totally great. So I started writing this. I started trying to implement the MSN Messenger protocol, MSN Messenger was basically the aim of Europe and all my friends were on MSN Messenger. I actually think the earlier version of MSN Messenger protocol were fairly straightforward, so I started implementing this. Then I finally got to the point when you could actually send messages to people, and so the project, this prototype thing now was a success. Now I had to figure out what messages the program should actually send, and so that actually starts to develop pretty interesting problems, like the first version had a repertoire of pre-programmed responses, and pick those at random and then you realize that not everyone immediately realizes that they are talking to a program and they start to engage it in conversation. So I gave it a name, I called it ISAAC. Then started writing these heuristics to try to have it keep the conversation going for as long as possible. That just became the really totally fascinating thing. Im sure a lot of you guys know, one of the big challenges in artificial intelligence has been the Turin test, where you have some human who is able to correspond through some terminal with both another human and a computer program and the Turin test proposed by Alan Turin suggests, indicates that true artificial intelligence one would not be able to discern which was the program, and which was the human. So now with Instant Messaging you actually kind of......, I mean people didn't realize that they were playing the Turin test but you kind of had them think the Turin test and this massively parallel Turin test and they are running these programs that tried to learn from all the conversations that they were having and go as far as they could, and so built up this enormous data base from hundreds and thousands of conversations. And just basically went down that AI rabbit hole and so some work around that was what I first entered into that competition. That was the first time I really felt like I had done something remotely interesting in programming.
From Starting College to Starting a Business
[00:11:12.0] So in 2006 you applied to MIT. You attend MIT, you take your first semester and about that time you and your brother start this company. Tell us a little about that.
[00:11:23.1] Patrick Collison: Yes. I applied to MIT when I still had a year and a half left in high school, so I ended up starting a year early. I actually came out to Stamford, my first time ever in the US I came out to Stamford for the International Lisp Conference in 2005. It was like this really eye opening moment where there was this other world of American Universities basically. That was in June, or something, so I basically decided then I would just like to apply to American schools for the next year. Which, I guess, was kind of out of spec the worst that would happen was they would reject you and you could re-apply the next year or something. So I did that and it worked out and so I started at MIT, but I was kind of a year behind. I started when I was seventeen, and I sort of felt like I had this spare year. The Irish High School system is pretty bad in a lot of ways but one thing that is good about it is that you do three years then this big exam, then you have this optional gap year, before you do two years and then another exam. That gap year is supposed to be your time when you can explore other interests or do some work experience in a career that you think you might want to pursue or go travel the world or whatever it is. John my younger brother was currently in this year, and so we both kind of lined up some spare time and decided that we might as well try and start a company because it seemed kind of interesting.
[00:12:41.4] He was how old?
[00:12:41.4] Patrick Collison: He was sixteen. So we started this company and the basic idea was that we thought Ebay was really cool, in that by providing the credit issue to used items, and by exploiting the possibility for coordination, sort of making the world much more efficient. Obviously Ebay just sucks, I apologize to any Ebay.
[00:13:10.0] DEREK: It's ok, we're AOL
[00:13:13.2] Patrick Collison: I mean I guess the more charitable way of saying it is something like Ebay's rate of improvement is slow.
[00:13:24.8] For the Americans in the room it means it sucks.
[00:13:27.5] Patrick Collison: So we tried to build some better version of Ebay. We didn't really have any totally clear ideas of how to go about that, but we started working on some prototypes, and then we applied to Y Combinator and what actually ended up happening is we merged with another company started by two guys, Harjeer Kulveer Taggar who had built this student market place in the UK and basically also wanted to build a better version of Ebay. So the four of us came together in May 2007 to work on Automatic, and the goal was to build a better version of Ebay.
[00:14:02.7] And how did that, just tell us how that process came together. Was that something where after you met, because you applied to Y Combinator with this original idea and then was it something where Paul said hey you got to meet these other guys. How did it work?
[00:14:14.3] Patrick Collison: Yes, he basically suggested that we talk to them and try to figure out if there was any way in which we could work together and pretty rapidly after meeting we decided that it made sense to work together. They had spent a long time working on this. They had successfully launched this market place and they really felt that they had made a very good understanding of it, and we liked the idea of working with them. Actually now, Harjeet is a partner of Y Combinator.
[00:14:43.5] So you went through the Y Combinator program because you moved here.
[00:14:48.7] Patrick Collison: That's right. We moved out here and we got this small, little apartment in San Francisco and went to work basically. I know that was my first time of spending any significant amount of time on the West Coast and learning how to start up to work. It was this incredibly foreshortened introduction to startup plans, it was like this rolling tour and came out and started working on the product, we raised some money from internet investors. We launched the product, it did reasonably well, people liked it, and then it was acquired. All of that happened from incorporation to acquisition in ten months, so Automatic never left sort of a huge mark on the world, but it was an incredible learning experience.
[00:15:33.0] I mean you guys were big news, you and your brother were so young, right? I mean you were all over the UK, there was just a ton of press. How did it feel coming out of that? Did you feel, how did it feel?
[00:15:49.1] Patrick Collison: To be honest I always felt it was a little bit misplaced in that, I mean there was a great learning experience for us. I don't think that Automatic ever... it was a great learning experience but it never left any huge mark the world and so I never have been all that enthusiastic with it being interesting because we were so young. I feel like, we were very lucky, circumstances aligned and I was out here and I knew Paul a little bit through LISP and stuff, where we mashed the right combinator, and things worked out, but basically I don't think we did anything all that special.
Riding the Wave of Early Success
[00:16:29.7] So you sold the company and did you move to Canada to actually be a part of the company.
[00:16:35.2] Patrick Collison: Yes, I moved and became the Director of Engineering for a year for the company, and spent about a year in Canada.
[00:16:43.3] So you spent a year in Canada. Your brother goes back to school, right?
[00:16:48.2] Patrick Collison: He went back and he finished school in Ireland and we were kind of lucky and this kind of gets back to the competitor aspect in that I actually had the urge to sort of graduate early to start college. I did the British leaving school exams A Levels and he did the Irish ones so we never competed head to head. Which was kind of lucky for me because he ended up getting the highest score in history in his exams. He just really sucks!
[00:17:22.6] So you go through this next year, you're working as Director of Engineering. He goes to Harvard, then are you guys talking this whole time. Are you saying we're going to, what's the next thing we're going to do, where was your head at that point? You're not going to stay in Canada for ever.
[00:17:40.4] Patrick Collison: Obviously we kept in touch and everything and then I decided to go back to MIT because I sort of, I felt like this, I was coming to the end of this one year experiment, from looking at starting a company and I guess neither of us were kind of sure that we didn't want to go down the academic rack like, and tried to become professors or something. So I went back to MIT, John was at Harvard, and we just had some side projects going on the side and I think the one that was the one that was the most fun to work on was, when the IPhone came out there was no STK initially.
But I was actually walking down the street in Vancouver with my friend and we realized that the IPhone was actually powerful enough and had enough storage space that you could store a complete copy of Wikipedia on the IPhone and so you could basically actually implement the Hitchhikers guide to the Galaxy. I think it was December of 2007, my friend managed to get a working Joe Broken IPhone tool chain, and this hacked version of GCC that could cross compile to arm with the IPhone and he just wrote some header files, guessing the method signatures of the UI Pin selector and he showed me at about 3:00 AM one morning a working Hello World App on the IPhones and this was when Apple still maintained that there were only going to be Well Apps and only the Apps that were built internally at Apple, they were the only apps that were allowed to run on the IPhone.
So this friend, Brian, who now actually works for Stripe, he got this tool train running so we decided to try to build this future spec galaxy app and this was just like a ton of time staring at arm disassembly and everything but finally it worked. We released it and it was terrible because we hosted the Wikipedia dumps ourselves and so it ended up getting a ton of downloads and each of them was 2 gigs and so I was basically hosting it on SliceHost and there was this overflow on Slicehost and you couldn’t show me how much bandwidth was actually being consumed. The project on the whole was always pretty fine idea. I think that I ended up getting a couple of hundred thousand installs on the IPhone again back before the Apps were and then it was imported to the $100 laptop the OLPC, because these were being shipped to everywhere, India and obviously most of these people don't actually have internet access, so this was how you could read the Wikipedia without it.
[00:20:17.6] That's really cool.
[00:20:18.2]Patrick Collison: So anyway, John and I were working on side projects like this and several other restarters came along.
[00:20:26.3] So just tell us a little about that: you’re trying to actually solve this problem yourself. Right? You're actually implementing merchant accounts, trying to set up payments on one of these projects, or how does it come about?
[00:20:40.0] Patrick Collison: Actually what happened was, one of the projects I worked on back a few years previously was just trying to build a better web framework in LISP. I guess I was right better web frameworks are very interesting but I was very wrong in thinking that like anyone who would use LISP, but I guess by 2009 or 10 or whenever it was, we had both built tons of web applications but we'd never charged for any of them. Like we had literally never charged for any of the software that we had built on the web and we started sort of wondering why that was? We looked back and realized that we had actually started applying for merchant accounts at various points and we had taken the first steps for integrating with PayPal which had never quite been the biggest priority at any one time. It was just the net results was we had never taken a single payment on the web. I was really taken with what was happening in other industries where there were these kind of varied developer or internet services coming along, building these really nice attractive areas over infrastructure making it really easy to go and do something. The obvious examples are things like Libode, or SliceHoster, Easy Tools, hosting also obviously telephony. Like a lot of these companies and so I think it was October 2009 or something we sort of had this idea that maybe it would be interesting to go Slice Host for payments and that’s how we thought of it. It was a slice host for payments. I guess a bunch of you probably use some of these hosting providers but at the time EasyToo was still sort of very bare bones and there wasn’t any much of a web interface and you had to use all of these command line tools to administer it. Whereas SliceHost was really nice slick gooey where you could start, click build server and 60 seconds later you would have a server. You sort of wanted like that magical instantaneous experience for payments.
[00:22:32.6] Sure. Paul Graham talked about using you guys as an example of really attacking very difficult problems and people that he uses comparison of how many recipe websites he’s got, how many pitches he gotten over the years versus no one ever pitched this idea and because it was so big you solved such a huge problem. Where did you even start? Where does something like that even begin.
[00:22:59.5] Patrick Collison We didn't really know where to start and I definitely wouldn't pretend we did.
[00:23:06.6] I mean I would have started by like crying in the fetal position. I mean did you......it's just you and your brother right?
[00:23:14.0]Patrick Collison: Basically we faked it. What we did was, we decided to go on holiday. Just go hack somewhere. Maybe, I don't know how to pronounce his name, but there's this guy, Matchay in, I'm going to totally mash
[00:23:32.9] DEREK: He may be here, so...
Building Stripe on the Beach
[00:23:34.4] Patrick Collison: I'm going to totally mess up his surname but it's something like Tschoulosky something like that, I apologize Machay, if you ever see this. His website is IdleWords.com. and he is a really good writer and he wrote these series of blog posts about Buenos Aires and how it is like the best place ever to go and hack, because it is really cheap, and everything is open really late and there is Wi-Fi everywhere and the climate is really nice and everything. John and I were facing this winter in Cambridge in Massachusetts which is definitely not like that in basically any respect. So we decided to go to Buenos Aires and just hack all day in cafes and try to build the perfect gear. So we did that and it turned out to be basically exactly as he had described and so if ever you just want to go and work on something single mindedly for a month I cannot recommend Buenos Aires more highly. We spent about $10 a day and the weather was gorgeous and bizarrely all of the cafes have Wi-Fi for no reason that I can discern, much more than here. All the restaurants open really late in that we were turned away from restaurants at nine o’clock in the evening because it was too early. All the bars are open until about 5:00 AM. You start going to bars at 2:00 AM and nobody gets up before midday so basically it is like an entire city on a hacker schedule. And we were like the worst tourists ever in that I still have not seen a single sight or anything of Buenos Aires. We just got up and went to a cafe, and hacked all day. At the end of that we had the first prototype of Stripe.
[00:25:02.8] After a month.
[00:25:04.7] Patrick Collison: Yes after a month, more or less we had the first production user actually about a week after we went to Buenos Aires. What we did was called up a friend who worked at a pan processing company and said is it ok if we just send a couple of accounts your way and we sort of built this really nice API and interface for setting up accounts or whatever. When you clicked credit account or whatever rather than an account actually being created in the financial infrastructure or however that worked, we had no idea. We just went and called our friend. Which on a scale led to a couple of users and so we
[00:25:43.3] and you set it all up manually behind the scenes
[00:25:44.5] Patrick Collison: Totally manual. At this point we had literally two users and both of them were good friends of ours. Paul Buckai has this great story about building GMail. He kind of repurposed the Google Groups Code which predates GMail to build GMail and so he built the first working prototype in an afternoon and it loaded all of his email into like the Google groups interface and he showed his friends and they were clicking through it and saying this is great you know, but it would be really awesome if instead of viewing your email I could view mine. It was like this very kind of direct user requester and development and it was similar to this with Stripe in that we sort of built the API that allows you to charge your credit cards, and so went on user requests from there. Ross Bruchet was actually the very first user of Stripe. He was at a company called North at the time and it was like this API is great but it would be really awesome if you actually transferred the money into my bank account after I charged credit cards. And similar requests like this.
[00:27:01.5] So it didn't actually go anywhere. Because I thought you were setting this up on the back end frame. It's all fuzzy now at this point.
[00:27:07.5] Patrick Collison: It was very much driven by basically what things they bugged us to build and so we launched this prototype and again I said
[00:27:25.1] You're not back from Argentina?
[00:27:26.3] Patrick Collison: Yes, we got back from Argentina at the end of January, the project existed until fairly close formed like Stripe is today. This really simple version. You can go to a website, fill in really minimum admin information. You can start accepting credit cards immediately on the internet and the fundamental idea is that it should be possible to launch a website and accept credit cards within 30 minutes.
Before Stripe it was like four or five days.
[00:27:55.3] Yes. A week. Five to seven business days or something.
[00:27:57.9] Patrick Collison: I mean that's when it sort of got hard, right. Because figuring out what the software side of Stripe should look like. I mean you can kind of imagine, but then how do you possibly reconcile all of that financial infrastructure and merchant accounts and gateways and logs and PCI and they are just the words we knew. So we went back to school and finished the next semester and decided to work on it over the next summer. That’s when we started working full time on Stripe. About a year and a half ago.
[00:28:29.6] For some background I’m sure a lot of people in this room have dealt with this but one of my first jobs, my first start up job in college was signing up or servicing merchant accounts and it was really...It was $2000 upfront to set it up and it was another $100 a month and then the fees kicked in. So for a small business it was insane. Absolutely insane.
[00:28:59.7] Patrick Collison: The fee schedule for many of our competitors runs to more than one page. I mean, I think every now and again you stumble across an industry that, the online payments industry was just like an unusually compelling example of an entire industry that is going to have its lunch eaten. Whether that is Stripe or something else, this is not how the world should work and I think it will be very good for the world when fee schedules no longer run to two pages.
Getting Involved with Paul Graham
[00:29:30.8] So you come back, you finish your semester at school, you come back here at the beginning of summer 2010 and it's still just you and your brother. What is the next series of events. You reach out to Paul and that team or are you still just building on your own.
[00:29:47.6] Patrick Collison: I guess two things happened. One is I guess like a lot of startups we had at this point we looked at Stripe as very much a side project in that we were also working on these IPhone Apps and Stripe was another interesting thing we had going on the side. That changed over that summer and what sort of changed was we thought more about the problem of transactions and the internet and more about the internet in general. What kind of things are holding the internet back and it struck us that, when you think of Google as really solving the problem of search the internet and you think of searching the web, it's very clear that we should have a good solution to search on the web. Google does a reasonable job. Maybe you can do better than Google but there is a pretty solid start at least. Similarly when you think of social and the web again, say Facebook. Perhaps we can do something better but pretty good job.
Then if you look at payments or even more generally like the economic foundation of the web it's really shaky and just like nothing there. PayPal I think was at one point a trajectory to build this but they have kind of... they never really got there. I mean it took us a long time to realize this, it took us six months but I think that is really a big deal in that the web, and the internet are still very young. Of all the things we buy and all the transactions we engage in it's still only a minuscule fraction of those happen on the internet. When you think about it it's just really hard to buy something on the internet today. It's crazy how manual it is, punching your credit card number in, digit by digit. Go to these multi save check out processes, it's incredibly hard to start accepting payment online and even if you start accepting payments online it's really hard to transact across borders.
If you have a credit card, that's great if you’re not from a US user, Good luck buying from somebody in Russia or China or Brazil or wherever it is. Similarly if you are a consumer, if you are an American you can buy from other American web sites but if you a Guatemalan or Indonesian you are restricted to a very small little market. I don't think this is going to remain the case. I think that there will be, and there should be a really robust and reliable simple straight forward powerful economic layer for the internet. Whether or not Stripe builds that I think that we realize the problems we face is pretty interesting and meaningful. We are kind of coming at it very much from this fairly humble beginning and building sites for payments we at least became convinced that there is something there. That this might be a somewhat rich seam.
[00:32:44.2] What is most interesting too is that you didn't start out, you know, you and John didn't sit down and say "You know what? We're going to build a payment platform" Right. It actually sounds like you were doing some of these other side projects of work that were more interesting to you because you were spending more time and they were more fun, and then as you continue to think about it, as it evolved over time. Over a span of six or eight months, it was like Wow! This thing we created has real potential. This kind of leads into something I want to talk about because we constantly hear, Paul constantly talks about working on big problems. You also talk about work on something you are passionate about and I wonder if payment something you are passionate about. I mean is it
[00:33:25.9] Patrick Collison: Absolutely. I think that that it is the fundamental enabler, as in your enabling people to make a livelihood. Like get into business. I think that there are very few things as fundamentally kind of interesting as that, and I think that payments at its core is pretty interesting and meaningful but especially on the internet. Like in a lot of ways, many classes of payments are resolved problems and so that there are tons of people that are confusion with, what’s the next way of buying a cup of coffee at Starbucks, and Starbucks mobile app that allows you to pay for your coffee and Google are doing Google wallet and all that NFC stuff and probably in ten years I will not hand my credit card to the cashier I'll do something else. Fundamentally I'll buy my cup of coffee at Starbucks whereas on the Internet a lot of the transactions that are going to have to happen they are still to be defined. We don’t really know what is possible, what things can be built, and so temporary make the experience of transacting on the internet better all the new stuff will be created. If we build a better way of buying a cup of coffee at Starbucks the world will still look basically the same. Whereas on the Internet when you solve this problem I think we will actually have a pretty different Internet. So I think it's not immediately obvious to me what are the more interesting problems solving on the Internet would be.
[00:34:51.7] In the end you've made this comment that ten years from now the things that will be accepting payments, it's new things. It's not the majority of the payments that are coming, it's not going to be on what we using today it's going to be on something new. So you are basically creating the infrastructure, creating the platform for that. Is that right?
[00:35:08.9] So you come back to Silicon Valley, you finally say ok we are going to focus on this thing. Then what happened, what happened after that?
[00:35:18.8] Patrick Collison: By and large we set up keyboards and typed and 99.9% of the time was spent doing that. I guess by the time the summer started we had a waiting list, I can't remember how many, like some number of thousand people.
[00:35:38.9] This is just spreading, I mean how do you get a thousand people?
[00:35:39.7] Patrick Collison: This is just by word of mouth.
[00:35:41.6] Word of mouth. You have a website that has a signup page, Put your email here, and we'll email you when it's ready.
Growing and Scaling Stripe
[00:35:46.5] Patrick Collison: Friends told other friends, this kind of stuff. We couldn’t have them use Stripe because I mean the problem is I told you how the account payment process actually happened in the bank, and so we sort of convinced ourselves that there might be something interesting here, but now we had to actually go try and build the infrastructure and make it work. And so we decided at the end of that summer that we would go and take this seriously and we would go and build that infrastructure and figure out whatever it is that we need to learn and actually launch it properly. So we took some investment and Y Combinator, Peter Thiel, Elon Musk, Sequoia and Theresean Hardwoods invested in our c brand. .
[00:36:26.9] Anybody we've heard of?
[00:36:27.6] Patrick Collison: Then we basically spent the next fourteen months building the infrastructure that could actually make this work. Something from like August 2010 to the end of September 2011, it was building infrastructure. We finally launched it to the public I think it was 30th September 2011. Just over six months ago.
[00:36:51.8] Walking back to that point I'm raising, tell us how those types of conversations start. You already know Paul, he invested in you. Some of these were probably your previous investors. Like you go and you meet with Peter Thiel and what happens. Is it... walk us through your first meeting with Peter Thiel.
[00:37:08.6] Patrick Collison: I remember being very critical of PayPal and then halfway through the meeting, maybe not quite the best idea,
[00:37:24.9] Wrong tactic.
[00:37:24.9] Patrick Collison: Mostly I remember him thinking that a lot of the things we were doing that I thought were a bad idea, and him thinking they were a good idea, and sort of vice a versa. And that being pretty interesting and so at the time we wanted to have this forcing function that would encourage us to make the product really good. So we charged 5% plus 30 cents for every transaction, just like way more than M & L charges to process online payments and so our thinking was, we don't want anyone to use us because of the costs. So rather than try to be even competitive why don't we just price it way higher than everyone else, and until people are willing to use us, like pay that large premium for the product that wasn’t sufficiently better. We didn't really think of it as a viable long term strategy and I remember Peter arguing that it quite possibly could be. That could be a really interesting business model.
[00:38:24.5] Basically you would provide a better customer service, better experience and you could charge a premium for it.
[00:38:29.3] Patrick Collison: Like a fundamentally different business model compared to every other internet company. I guess I always remember Peter being like smart and decisive within two hours of meeting he had agreed to invest. We had tons of other meetings and lots of other people told us that this is not even possible. Or there were a million reasons why we would fail, whereas Peter pretty liked it and ok I will invest. That I think really changed our projector.
[00:38:56.0] So you also took Y Combinator money but you didn't go through the program again. You had been through it right?
[00:39:00.3] Patrick Collison: We had been through it. I guess Automatic had gone through Y Combinator and Stripe sort of already existed, so I guess for no particularly good reason we didn't end up going through the program again. Like I think it would be very hard to go back and look at that position again. I think it would be 50/50 as to whether we should but we didn't go through it that time.
The Leverage of Y Combinator
[00:39:23.4] Can you talk about what are some, I mean Y Combinator it's clear they provide financing, they're great there, but as somebody that's now been a part of two companies that have been invested through their program what are some of the tangible benefits that people get, or intangible benefits, that they get going through Y Combinator.
[00:39:45.3] Patrick Collison: So here’s a bunch of them. I think that for us, one thing that was really valuable was having all these alumni just being a part of this network of startups, I mean we were building a service that initially users included a bunch of other start ups, and so having all kinds of people who we knew were building start ups, and being able to say to them, Oh you guys should use Stripe instead of telling launch just to help us because we were a fellow Y Combinator company. I mean that was just super valuable for us. Other Y Combinator users were our users and I think our first ten or twenty users were all YC Companies and so that was really helpful. I guess I presume many if not most of the people here have start ups themselves and it turns out start ups are really hard, and just having a bunch of other people around you who are doing the same thing, and get that, is really nice and valuable. Actually I think, Y Combinator talked quite a bit about, more than most, similar incubators or whatever you call that class of C Funding, I think that they talk more than most about their values and what they do and they have really detailed document on the YC site describing just what YC is what you do. They talk of some of the benefits, they have a ton of data, they see hundreds of companies and so they can apply aggregated other companies experiences to your experience and help with data advice in that way. Or the alumni network, or whatever it is. I feel that YC can't actually, it's difficult for them to talk about it, the biggest advantage of YC, is that simply that the Paul and their partners is really smart. I feel like they can write this document, that you should do YC because we're really smart. I think it's actually the best reason to do YC, in that Paul initially, and the other partners, and the other partners that have joined, really really really good. I mean I guess with Stripe we are lucky we got to work with, lots of good investors, but like there's no question of YC are like at the very top tier.
[00:41:58.0] And when you say good and smart are you talking about product feedback or execution or technology or what exactly do you mean by that?
[00:42:07.4] Patrick Collison: About a month ago, or something, there was there beam going around Facebook of like Well my mom thinks I do, and my friends think I do, and for all these different professions and I feel like the world thinks startup founders are thinking up all these strategic thoughts and reclining in their chair and thinking how to deal with the next big threat or whatever it is. What the next giant breakthrough will be. In fact most of your time is spent, this particular thing is broken and the person who is going to fix it, like had to go home and I want you to fix it and this investor has this question you need to produce, this document. All these distractions, when I think that basically it all turns out that startup is dealing with distractions in serial. Sometimes the issues are actually huge, that you are trying to close this round, this big deal, some recalcitrant investor, this stressful thing happening or it's I don't know how to talk to this customer, or whatever it is. I really kind of spans the gamut. I think Paul and the other partners are really good at getting that. They have done companies and they have seen this happen literally thousands of times in other companies and so whether the issue is closing this round, doing this big deal, this acquisition or whatever. Or its just well, how should the signup button work? They have pretty good advice. Thats has made a big difference to Stripe.
[00:43:39.3] Talk about just a little bit, and then we'll get back to Stripe. Talk a little bit about Paul Graham. You know Paul Graham probably as well as any of the YC founders do. What makes Paul Grahams specifically so good, what makes him unique? How has he helped you in your startups.
[00:43:58.7] Patrick Collison: He kind of reminds me a bit of Peter Thiel actually. He just makes these surprising connections, and comes up with these surprising ideas, things that I have not thought of. I guess I feel like lots of investors are pretty good at the quotidian things. Helping you close other deals or you have a question that how shall I grant options, what lawyers should I use. I mean there are a lot of investors that I think can give you pretty good advice there, like usually I find the conversation that are most kind of useful from Paul are thinking about something totally new area or something and go to Paul and I come back with ten ideas that I had not thought of. Some of them are really outlandish and terrible ideas and some of them are good. I think that both Paul Graham and Peter Thiel both have this property where they sort of look at the world sideways. They see it slightly differently to everybody else and I’m not sure if thats kind of an acquired thing or it's a kind of consequence thing.
[00:45:15.9] Is there anything that he's not very good at? You can say no!
[00:45:17.2] Patrick Collison: He's not very good at feigning interest in something that he doesn’t care about. He'll get really taken about some particular thing you could do with the business. You will be like, and then you'll move on to some other topic and for example say you're talking about how you should issue options and now you're thinking about how you should issue options and Paul is ten minutes back in the conversation and thinking, you know what you should really do and it’s like this giant take over the world plan and topic you have never even thought of.
Progress at the 6 Month Mark
[00:46:08.1] Great. You launched six months ago and tell us, at this point, how big is the company. You guys were based down here, you've now moved to the city. It's more than just you and your brother, actually and the guy from Trading North joined your company right. Your first customer actually joined your company.
[00:46:27.0] Patrick Collison: Actually several of our customers have joined Stripe. One of the things that turned out to be really nice building a service product start ups is that you meet all these people that you would love to work with and so we repeatedly have the experience of somebody build something along with Stripe and sort of over the course of them building on Stripe we got the chance to interact with them a lot and we're really impressed. Building this business with Stripe is great but we'd love to have you actually build Stripe stuff. When we launched we were eight people or so, now I think we are twenty three. Obviously, I think it was eight people when we launched, four or five were Stripe users, and thats how we came across them, and I guess of the twenty something the vast majority are engineers. Mostly, when we launched, I don't think Stripe is actually all that incredible an idea, I think that it's kind of an idea that's been in the air for a while. Finally we'll make payments, not like really suck on the internet, and so I think that the reaction to Stripe was great, payments don't suck anymore. Since we launched it's very much been just trying to keep up with the growth where it turns out there are all these people that want to use payments that doesn’t suck and so there have been tons of sites going live on Stripe every day and obviously we have just been trying to keep up with that.
[00:48:15.3] So you grew one to a thousand based just on word of mouth, how have you grown? Has it been the same?
[00:48:22.1] Patrick Collison: It's all been word of mouth so far, or word of mouth with people writing about us on blogs or the thing that we really like and is really kind of meaningful for us that happens, people write about their experience of integrating Stripe and because that's the goal. We just want to make is super simple for people to build a business on the web. That is the thing we are always asking ourselves, how could we make it easier for somebody to start building a business on the internet. We've done no marketing so far. We started running some limited ads on like stack over phones a week ago but up to that it was all either word of mouth or stuff that our users wrote about Stripe. Again I think we intentionally held off marketing for a long time because it’s that forcing function again. I think if our users aren’t talking about Stripe well then we should make Stripe more awesome until they do. I think that marketing and sales teams and that kind of stuff can be really useful excellence when you are really confident that you have the product right. I think that... I came across this quote yesterday from Eugene Kleiner how even turkeys can fly in a high wind. I think it was something like that with start ups. With enough sales and marketing you can make even a really crappy product successful and so I think, that by not doing any marketing for a long time it's making sure that we are liked. Not just like this floundering turkey. I mean now we have some nice suggestions out that people like the product and stuff. Now we are much more ok with, I guess there’s two sides. On one side you want to make sure you're not the turkey but the when you are ok I’m not the turkey the other trap you can fall into, I think alot of engineers, and I see this trait in myself, have this I don't like this solipsism or this kind of self-centeredness or like I’m going to build my beautiful code over here in the corner and I will wait for the world to discover me and I’m just going to focus on all these technical problems. I think you have to go and engage with the world and talk to them, tell them about what you are doing and why its meaningful and everything. Yet you don't want to do that, or there is some value in holding off doing that initially or not pushing it too hard because its really nice to wait and get to the point where people are actually starting to discover you but once you have got to that point, and you think that you actually have a thing, then I think you have to go and tell the world and really bring it to people. I think that over the coming years we will try to do that more but there is some value for holding off for a while.
Taking on the World of Payments
[00:51:17.8] So you are currently in how many countries are you in. Which countries are you in?
[00:51:22.2] Patrick Collison: Well you can look at it in two ways. If you have a Visa or MasterCard or whatever brand of credit card, it will work with a Stripe user. We have purchase users from all most every country in the world, however, if you want to accept payments with Stripe, and sign up with Stripe you have to be based in the US today. That is the biggest bug with Stripe right now. The whole promise of the internet is that we will break down borders and there will be this incredible shiny global real time exchange of information. We are trying to build the compliment economic layer but it's frankly just broken right now and you can only accept payments if you happen to be physically resident in the US. That makes no sense on the internet and so we are working on fixing it. I think that, at least when we launched, that the majority of people working at Stripe were born outside of the US and grew up not being US citizens.
[00:52:20.8] It was just you and your brother at the front of the company is that right?
[00:52:23.0] Patrick Collison: When no when we launched there was eight people or nine or whatever it was, even then there was like a ton of non US citizens. Or at least people who didn't start out as US citizens and I mean you build a startup here and you always care, you always want your mum to be able to use it. Whenever I talk to my mom about Stripe she's like can I use Stripe yet? No mom, and so we are working on it and I think
[00:52:55.1] You have a big announcement here right? Exclusive!
[00:52:56.5] Patrick Collison: I wish, I wish. I think now at twenty something people at Stripe I think that 50 percent is non US and so.
[00:53:04.4]You're cranking on it. We will get to the audience's questions here in just a minute. I just want to ask you one more thing. Paul talks a lot about going after real problems. Difficult problems and I reach out to your younger, soon to be very famous brother, Tommy, and he's sixteen and I got this email back from him and I leaned over to Russ who is sitting across from me, Jese like I want this guy to write my obituary like today. He's just amazing. Tell me something about what makes Patrick unique and it really feeds into this and he just said Patrick is a son of... no he basically said you have an unique outlook on life that very little is truly impossible. It's deep. Really really deep. Speaking to an audience of entrepreneurs and founders who like myself have built mobile games and social apps and where does the motivation come or how do you start to work on a really really hard problem, and how do you keep at it. You have obviously hit these very difficult steps along the way and you guys have stayed with it.
[00:54:28.8] Patrick Collison: I think that the biggest thing that I didn't realize about quote on quote hard problems when I started out is that the harder in a sort of specific way. You pick some giant jobs, like say its fix health care, and like fixing health care is really hard. There are all these obstacles and bureaucracies’ in your way and it's not even clear how you should fix health care and all that stuff but it's all these hard problems are easier in a way in such as its much easier to inspire other people to work with you and it's much easier to inspire yourself to work on this problem. In that you wake up in the morning. I don't want to pick any particular example because it's very easy to sort of wrongly generate things that look trivial, but if the thing is actually trivial, sometimes you wonder why am I doing this? Is this really what I thought I would be doing at whatever stage in my life. Whereas if you are doing something really hard and really meaningful I think that that pitches on some metaphysical level and makes it all easier. You're trying to convince someone that you really admire to work with you and if you are working on something that actually is trivial you can be faintly embarrassed and do you want to come and work on my thing but you can’t really put your heart and soul into it whereas if you really are working on something that your heart and soul is in fact in then it makes it easier. I think that people are pretty good at dealing with the kind of mechanical challenges of a hard problem and the ones that really cause you to lie awake at night are the, for the lack of a better term, the metaphysical philosophical ones or whatever, working on hard problems almost definitely solves at least that second class. You have picked the thing that you really think is meaningful so I think that hard problems is actually a bad term for it because it’s like hard easy problems, I think that helps.
[00:56:40.6] Great. Can we give Patrick a round of applause?
From the Audience
[00:59:51.0] Audience: How does Stripe feel about patents?
[00:59:52.1] Patrick Collison: I guess like almost everybody else in the industry we are pretty opposed to software patents. I think that the real problem with patents is the game theory of it. No matter what your belief on software patents you are severely disadvantaging yourself or you the company by not pursuing software patents because kind of everyone else will and it's sort of this arms race. So I think that to break that equilibrium we really need to change the law or change the structure in some clever way. For example like Twitter is doing. If you actually go and read the legalize I’m not convinced that the Twitter approach the right solution. For example that allows authentic use of patent even if cases where the defendant is believed to be a future patent as opposed to someone who has actually taken a lawsuit. I am actually pretty, there are lots of things that I despair about but software patents isn’t one of them because I think that we are living in this stupid time right now where we are being stupid about software patents. I think we will fix it within five years because software patents are a netpad idea and I don't think we'll get there terribly quickly but I think we will get there.
[01:01:59.6] Audience: So what specifically does Stripe do different from PayPal?
[01:02:06.0] Patrick Collison: It's a little bit hard to answer that question directly because PayPal is not just like one model of their product it’s like this large cluster of products flying in loose formation. The core PayPal product people think about is this thing that appears on a website, you click pay by PayPal, you go to the PayPal website you enter your credentials there and you offroad around the PayPal fortress for a while and then you hopefully end up back. Stripe is better than that PayPal experience because the checkout experience happens right there on the page, people building a site on the internet, just want to have a payment experience that is just a good as the best sites in the internet. You want your payment form just
[01:02:46.9] Or feels like my payment, if it’s my products on my website it feels like it’s my payment process.
[01:02:54.6] Patrick Collison: and you want to be able to integrate it into the flow of your site in a way that is natural and makes sense, and not have to throw people over onto these other websites and hope they come back. The other part of it is, with PayPal there is this adverse selection come negative signaling problem where you have to look a little bit more like a mom and pop or smaller operations who use PayPal, just because all the best companies insist on going getting a merchant account and they are willing to invest the weeks that it will take to do that. We are better than that particular PayPal product, we provide the first class clearance but there are other PayPal products, it’s hard to go through them one by one but I think at the middle level we feel that Stripe is a more compelling product than PayPal because, not so much because of a featureless comparison although we can do that featureless comparison. I think that Stripe does well in that but just, if you compare the IPhone and Android on a feature checklist they look basically the same product and yet if you use an iPhone and then Android phone for five minutes it's like night and day. They are such different products. They have such radically differently quality of execution and its user experience. I think there are valid reasons to choose Android, things like the App review process and all that stuff, but they are very different things. I am not claiming that Stripe and PayPal is the IPhone and the Android but there’s something philosophical in these products, the IPhone and the Android just differ and it's hard to put your finger exactly on what it is but it’s there. So with Stripe we have a very firm belief what this product is and how the payment should work. It should be a really simple API, it should be integrated onto your website, everything should be really fast and beautiful and simple. With PayPal they have tons of different products and certain permutations and different pricing schedules and the interface is a little bit dated and all these things but it is not one single thing that separates them its just underline philosophy.
I’m not sure if that answered the question.
[01:05:15.3] A better comparison would be maybe Stripe is like an IPhone and PayPal is Blackberry. Kind of like that. That's how I look at it. I mean that's my two cents.
[01:05:36.1] What are your thoughts on Shopify, or enabling payments like Gumroad.
[01:05:43.6] Patrick Collison: I don't know Shopify products real well. I’ve looked at their website but I’m not all that familiar with the product. I think that the, no, I’m going to try and dodge the question and talk about the general concept here because I’m just not all that familiar with Chirpily. At a middle level I guess Chirpily are building or helping build or facilitate more light weight forms of commerce. In that commerce is not the going to Amazon and putting a thing in your cart, and basically like a recreation of a supermarket experience online but it can be something much more casual and buying something at a stand or something and I think on a philosophical level that is a really promising direction but I think the internet of ten years’ time when we have had this sort of better economic foundation or whatever it looks like. It's all sort of making much more three hundred dollar purchases so I think we will be making much more casual purchases, and giving value to way more people and just much lower activation energy. And in the sense that Chirpify is doing things like that by integrating them into Twitter that is a really promising avenue.
[01:07:21.8] What are some of the solutions you have in place to deal with fraud?
[01:07:26.8] Patrick Collison: I think this question is really interesting because when we all think about online payments one of the first things you think about is fraud. I mean everyone has heard the horror stories from PayPal and how they almost went out of business because of the Russian mob and all these crazy stories and I think that a lot of the subtlety of the stories have been lost a little bit in that. For example with PayPal even at their time of their greatest losses they actually had a pretty good idea of where the fraud was coming from and even how they could mitigate it. They made the determinations in hindsight correctly that they could afford the losses and the additional growth they got by taking on these somewhat suspect users was just worth it. Eventually it got better at detecting a fraud and it became less of a problem and if you look at PayPal’s fraud losses today they are really low. I do think still, and alongside PayPal have process have very small factor of all online transactions. A vast majority are processed by this kind of legacy of infrastructure merchant accounts and gateways and everything else and these companies they don't have these insanely brilliant mathematicians and programmers and ML people and big gatewaypeople sort of curating these models they apply a fairly simplistic approach to fraud and it turns out that that actually works So I think fraud is not the kind of existing essentially challenging problem for an online company that people tend to extrapolate from the PayPal experience. I think that one of the mistakes that PayPal have made on fraud is that they just don't, I feel they look at it from a very mathematical standpoint. They think oh well if I shut down this user there is X probability I am wrong, they are a false positive and the loss in making an incorrect decision is the loss of the profit I would make on this user. When of course that is not the loss, the loss is that you make is the giant reputation loss from someone talking about PayPal and telling their friends and sharing the experience of how PayPal, literally in many cases took away their livelihoods. I think that that is short sighted or mis calibrated mathematical model is actually very dangerous way to model a product. We are really conscious of that at Stripe and we can do some things to convince you that we are trustworthy and safe and everything else and we are audited to the most stringent security levels possible and we have hired people from other companies who have spent tons of time dealing with fraud and know the industry really well and this kind of stuff but I think ultimately what it boils down to is we do not intend to have these stories appear of like how Stripe took away my livelihood. I think that alot of it comes down to that. We simply were not willing to make those mistakes, and that we are not willing to look at the individual user loss as being the loss of that profit. I think that it is much larger than that. Coupled with that we are very emphatic in the view of the downside of being wrong on fraud but the slightly more encouraging side is that it is pretty hard but it is not quite as hard as everyone assumes so hopefully we will do ok.