Lecture 11 - Hiring and Culture, Part 2 (Patrick and John Collison, Ben Silbermann)

1. SASam Altman0:01

   Part 2 of Culture and Team, and we have Ben Silbermann, the founder of Pinterest, and John and Patrick Collison, the founders of Stripe. Um, ye- founders that have obviously sort of some of the best in the world at thinking about culture and how they build teams. So there's three areas that we're gonna cover today. One will just be sort of general thoughts on culture as a follow-up to the last lecture, and then we're really gonna dig into what happens at the founding of these companies and, and building out the early team, and then how that changes and evolves, uh, as these guys have scaled their companies up to, you know, 100 plus... I don't even know how many people you have now, but quite a lot. These very large organizations and how you adapt these principles of culture. Um, but to start off, I just wanna, uh, ask a very open-ended question, which is, w- what are the core pieces of culture that you found to be most important in building out your companies?

2. BSBen Silbermann0:52

   I'll start with that. Sure. What are the, what are the most important parts? Yeah, it's on. It's recording. Oh, it's on. Um, yeah, I mean, I think, I think for us, like, we think about it on a few dimensions. Like, one is, like, who do we hire, and what do those people value? Um, two is what do we do every day? Like, why do we do it? Um, three is what do we choose to communicate? Uh, and then I think the fourth is what we choose to celebrate, and I guess the converse of that is, like, what you choose to punish. But, uh, in general, I think running a company based on what you celebrate is more, more exciting than what you punish. But I think those four things, um, kinda make up the bulk of it for us.

3. JCJohn Collison1:33

   One thing that we've placed a hu- a large emphasis on as Stripe has grown and probably more than other companies is, uh, transparency internally. Uh, and I think it's something that's been really valuable for Stripe, uh, and also a little bit misunderstood. Uh, all the things people talk about, like, you know, hiring really great people or giving them a huge amount of leverage, uh, transparency for us plays into that. Uh, we think that, you know, if you are aligned at a, a high level about what Stripe is doing, if everyone really believes in the mission, and then if everyone has, uh, really good access to information and kind of has a good picture of the current state of Stripe, uh, then that gets you a, a huge amount of the, of, of the way there in terms of working productively together. Uh, and it, it kind of forgives a lot of the other things that tend to break as you, as you grow a startup. A- and so we've, as we've grown, you know, we s- we started off two people. We're now over 170 people. We've put a lot of thought into the, the tooling that goes around, uh, transparency because, you know, at 170 people, there is so much information being produced that you can't just consume it all as a fire hose. Uh, and so how we, you know, use Slack, how we use email, things like that, we can go into it more later. Uh, but I think that's one of the core things that's helped us work well.

4. PCPatrick Collison2:55

   Um, I think culture to some degree is basically kind of the, the resolution to a bandwidth problem, uh, in the sense that, uh, y- you know, when, when you start out working on something, you're sort of coding all the time, but you can't code all of the things that you think the product might need or the company might need or whatever, and so you, you decide to work with, with more coders, right? Um, and, uh, a- and, and so, you know, the, the organization gets larger, and maybe in some idealized world, I, I don't think this is actually true, but kind of ideally you could be involved in every single decision, in every single sort of m- moment of the company, in everything that happens. But obviously you can't, or maybe you can at two people, but you certainly can't at, at even, like, five or 10. Kind of that, that point comes very quickly. Then by the time you're 50, it's completely hopeless. And so culture is kind of how you, uh, uh, kind of what the, the, the strands are that you sort of want to have, the, the invariance that you want to kinda maintain, uh, a- as you can get specifically involved in sort of fewer and fewer decisions over time. Um, and, uh, and I think when you think about it that way, it, it, you know, may- maybe its kind of importance becomes sort of self-evident, right? Because a- a- again, like, the fraction of things you can be involved in directly is, like, diminishing, I mean, almost exponentially, sort of assuming your, your, your headcount growth is sort of on a curve that looks like, you know, one, one of the, the great companies. Um, and, uh, yeah, that, that's, that's super important. And, and it kind of, it, it manifests itself in a, a, you know, bunch of, you know, different ways. Like, for example, uh, in hiring, uh, I think a large part of the reason why the, maybe the first 10 people you hire, why, kind of, those deci- decisions are so important is because you're not just hiring those first 10 people. You're actually kind of hiring 100 people because you should think of kind of each one of those people as bringing along sort of another 10 people with them, uh, and sort of figuring out exactly what sort of, what 90 people, uh, you would like those first 10 people to bring along i- is obviously, uh, you know, it's, it's gonna be quite consequential for, for your, uh, for your company. Um, but really briefly, I think it's, it's largely about sort of, uh, abstraction.

5. SASam Altman4:59

   S- so one thing that a lot of speakers in this class have touched on is how, uh, hiring those first 10 employees, if you don't get that right, the company basically, uh, will never recover. But no one's talked about how to do that. So, so what, what have the three of you looked for, um, when you've hired th- these initial employees to get the culture of the company right? How, how have you found them, and what have you looked for?

6. BSBen Silbermann5:23

   Uh, sure. So I guess this answer is different for every, every company, and I, I'll say for, for us it was very inductive. So I literally looked for people that I wanted to work with and that I thought were talented. Um, I think I've read all these books about culture because when I don't know how to do something, I first go read things, and everyone has all these frameworks. And I think one bi- big misconception, um, that someone said once is that people think culture is like architecture when it's a lot more like gardening. Uh, you know, you plant some seeds, and then you pull out weeds that aren't working, and they sort of expand. So when we first hired people, um, we hired people that were like ourselves, um, and, uh, I often looked at, like, three or four different things that I, I, I really valued in people. You know, I looked for people that worked hard and seemed high integrity and low ego. Um-

7. PCPatrick Collison6:08

   Um, I looked for people that were creative, and that usually meant they were really curious and they had all these different interests. Um, some of our first employees, uh, are probably some of the quirkiest people I've ever met. Um, they were engineers, but they also had all these crazy hobbies. Like, one guy had made his own board game, uh, with this elaborate set of rules. Another guy was really into magic tricks, uh, and he had coded not only, like, this magic trick on the iPhone, but he had shot the production video and the preview. Um, and I think that, that, that quirkiness has actually been a little bit of a calling card, and we find that really creative, quirky people that are excited about many disciplines and are extraordinary at one, um, tend to build really great products. They tend to be great at collaborating. Um, and the last thing is, you know, we really looked for people that wanted to... They just wanted to build something great. And, and they weren't, they weren't arrogant about it, but they just felt like it'd be really cool to take a risk and build something bigger than themselves. Um, and that at the beginning is very, very easy to select for, um, if you were in our situation. We had this horrible office, like, nobody got paid. Uh, so there was no external reason other than being excited about building something, uh, to join. In fact, there was every reason not to. Um, and that's something looking back I really, really value, because we always knew people were joining for the purest reasons, and in fact, were willing to forgo other great job opportunities, market salary, uh, a clean office, uh, you know, good equipment, just for the chance to work. So, um, to this day, I think a lot of those traits have been seeded and are embedded in the folks that we look at now.

8. JCJohn Collison7:36

   Um, yeah, the, the first 10 hires are really hard, uh, because, you know, y- you're, you're making these first 10 hires at a point where no one's heard of this company, no one really wants to work for it. You're just these, like, two weird people working on this weird idea.

9. PCPatrick Collison7:51

   Like, like, their friends are telling them not to join. Uh, for, for, for our second employee, uh, uh, he was like... I think maybe he'd accepted the offer or he was just about to, and his best friends, like, took him out the night before. It was like a full-on assault for, uh, you know, why you should not join this company, why this is, like, ruining your life, basically. Um, and, uh, and so anyway, the, the guy subsequently, you know, or continued to join. A- and actually, one of those friends also now works at Stripe.

10. JCJohn Collison8:17

    [laughs]

11. SPSpeaker8:17

    [laughs]

12. PCPatrick Collison8:17

    But, uh, but, but, uh, th- this is what you're up against.

13. JCJohn Collison8:20

    Yeah. Um, and I mean, it's also hard because no batch of 10 people will have as great an influence on the, uh, on, on the company as those first 10 people. Um, and I think everyone's impression of recruiting is, you know, you, you open LinkedIn, it's sort of like ordering off the dollar menu. You're just like, "I want that one and that one and that one." Uh, and now you have some hires. Whereas, uh, at least for us, uh, it was very much over, uh, a very long time period, uh, talking people we knew, uh, or friends of friends into joining. Uh, we, we didn't have, you know, huge networks. Uh, Pa- Patrick and I were both in college at the time, so there were no people that we'd, you know, we'd really worked with to draw on. Uh, and so a lot of those early Stripes were people we had heard of, friends of friends. Uh, and the, the other interesting thing they all had in common is that they were all sort of, um, early in their career or undervalued in some way. 'Cause when you think about it, if someone is a, a known spectacular quantity, then, you know, they're probably working in a job and very happy with that. Uh, and so we had to, to try and find people who were, uh... You know, in the case of our, uh, designer that we hired, he was 18 and in high school and in Sweden at the time. Uh, in the case of our, our CDO, he was, uh, in, uh, he was in college at the time. You know, a lot of these people, they were, they were early on in their careers. Uh, and the only way we could... You know, you, you, you can, you can relax one constraint. You can, um, you know, relax the fact that they're talented or relax, you know, that it's apparent that they're talented. Uh, and we, w- you know, not consciously, but we, we relaxed the latter.

14. PCPatrick Collison9:57

    Yeah. I, I think sort of f- f- finding kind of the people who, uh, uh, are, are kind of... I guess you have to kind of think like a value investor, right? You're, you're, you're looking for the, for the, the human capital that's, uh, uh, that's significant to the value of the market. You know, you probably shouldn't look to hire, uh, your, your brilliant friends at Facebook and Google or whatever because they're already discovered. Uh, you know, if, if they're willing to join, that's great, but, uh, they're probably, uh, harder to convince. Um, uh, John and I spent a, a little, a, a while, uh, yesterday afternoon sort of trying to figure out in retrospect, um, what kind of traits, uh, our first 10 or so people had in common that we thought were significant. Um, you know, in general, sort of in speaking about culture, uh, you know, I sort of want to caveat everything we say with, you know, I, I... Paul Buchheit I think said that sort of, uh, advice is, you know, a very limited experience wildly, uh, uh, over-extrapolated. Um, and, uh, I think there's a lot of truth to that. But, uh, for, for our particular first 10 people, the, the things we sort of, uh, figured out that seemed to be important were, uh, they're all sort of very genuine and straight. Um, I think that actually matters quite a lot in that sort of they're people that, uh, that others want to work with. Um, that they're people that others trust. They sort of have an intellectual honesty in how they approach problems and, and so forth. Um, they were people who really liked getting things finished. Uh, there's a lot of people who are really excited about tons of things. Only a subset of those are actually excited about, like, completing things. Uh, you know, there's a lot of talk about, for example, uh, hiring people, you know, off their GitHub resumes or whatever. I actually think that doesn't quite ring kind of correct to me in the sense that, that that places a large premium on sort of lots of different things. I, I think it's actually a priori sort of much more interesting to, to, you know, w- work with someone who has spent two years sort of really investing in going deep in a particular area. Um, and then, uh, the, the third trait that they all seem to have in common is they just sort of cared a great deal. Like, it was offensive to them when something was just a little bit off. And, uh, it kind of... Again, in hindsight, there are all these, like, crazy things we used to do that, um, uh, I mean, do in fact seem crazy. Like, we probably shouldn't have done them, but, uh, every- everyone was almost like, well, w- was borderline insane in sort of how much they cared about tiny details. Like, we used to, um-Uh, like, uh, every single API request that ever generated an error went to all of our inboxes and phoned all of us. [laughs] Uh, because it seemed terrible to, like, ever have an error that, that, you know, uh, uh, didn't go and sort of get a resolution, uh, from the user standpoint. Um, or we used to, like, copy everyone else on every outgoing email, and we'd, like, point out, you know, slight grammar or spelling mistakes to each other because it'd be terrible to ever send an email with a spelling mistake. Um, so, uh, anyway, those were the, the, the three traits we came up with anyway. Uh, genuine, um, uh, uh, caring a great deal, and, um, uh... Sorry, what was my second one? The other one. Uh, yeah. [laughs] Um-

15. BSBen Silbermann12:54

    Completing things.

16. PCPatrick Collison12:55

    Sorry. Uh, sorry, yes, uh, completing things, like the list of three items. [laughs]

17. BSBen Silbermann13:00

    Yeah, I mean, the only other thing I'd say, I, I just don't think there's any wrong place to find people. Um, so when I look back at our, our first few folks that we hired, they came from all over the place. Like, I put up ads on Craigslist. Uh, I went to random tech talks. Um, you know, we met people at barb- We used to throw weekly barbecues at the office. It was like bring your own... Like, bring your own food and drinks, and then we would just talk to folks. I think every time I ever went and got coffee in Palo Alto, like, one of you guys was, uh, recruiting, uh, at Coupa because their office was, like, strategically situated next to the best coffee shop. Um, but I, I think that the really good people generally, um, they're generally doing something else, and so you have to go seek them out rather than expecting that they're gonna seek you out. Uh, triple so, uh, when no one's ever heard of or is using, uh, the product that you're working on.

18. PCPatrick Collison13:48

    Yeah, uh, um, uh, it's probably really important to have a, a great elevator pitch, not even for investors, but just because everyone you, uh, run into right now is, you know, maybe six months, a year down the road, a potential recruit. Uh, and so the, the right time to have gotten them excited about your company, the right time for them to have started following us and, you know, be thinking about it as they think about what they're going to do next, um, is, you know, as soon as you can start. Uh, it's going to take a, a very long time to recruit people, so being able to consistently get people excited about what they're... What, what you're doing, uh, will pay back dividends later. M- Maybe this is, um, uh, a l- little bit tangential, but, uh, John and I were also chatting about it yesterday afternoon. Sort of, uh, like a, a bunch of our friends, uh, have sort of started companies right out of school, and we were sort of thinking about, uh, you know, what, what seems to go wrong in those companies. Uh, and I think something that, um, that maybe the m- most common, uh, failure mode seems to be sort of doing something kind of overly niche or overly sort of specific and bounded. I think maybe it comes from sort of, uh, like there's a major shift in time horizon a- as you go from classes to building a startup, right? A, a class kind of plays out on a quarter or a semester or whatever, um, whereas a startup is like a five- or ten-year thing. Um, and I think this is really problematic because, uh, it's actually quite hard to hire people for niche things, uh, in that if, if you tell somebody, "Look, we're going to build, uh, a rocket that goes to Mars," like, I mean, that sounds almost impossible, but it also sounds fucking awesome, right? Uh, and so it's actually pretty easy to convince people to work on it. Whereas if it's, "Well, you know, we're going to build this," uh, I don't want to single out any particular idea because probably sound like I'm picking some actual startups doing it. But, you know, if, if you pick something pretty narrow, something that maybe kind of inductively comes out of the kinds of problems you'd solve as part of a class project, uh, that's actually much harder to hire for.

19. SASam Altman15:31

    S- O- One specific question that has come up a lot, um, is how as a relatively inexperienced founder do you identify who the really great people are? So, you know, you meet people at these barbecues or through your friends or whatever, and maybe you've worked with them a little bit. But h- what specifically did, did you guys do in your processes to identify, like, you know what, this person's gonna be really great, or when did you really get it wrong? But what, what have you learned about how to identify raw talent if you can't just say, "Well, they work at Google or Facebook, they must be good"?

20. BSBen Silbermann16:03

    Well, I mean, you'll never, like, 100% know, obviously, until you work with folks, which is why the flip side of it is, you know, if someone you hired just wasn't a good fit, you owe it to the company and to them, uh, to tell them how they can improve and if they're not working out, to fire them. But I think that, that generally the question of talent falls into two big buckets. Like, one is you have some sense of what makes them good at their job, uh, and there are some areas where you have taste in that area, um, and there are some where you don't. And, and the ones where you don't are actually much more difficult. Um, so what we would do is, is we would do a few things. Like, first, before talking to anyone, we try to get a sense for, like, what, what is really world-class, uh, in that discipline mean? And this becomes very, very important later when you're hiring things like head of finance, and you don't know anything about finance except what was contained in, like, a library book you got about, like, an introduction to finance or head of marketing. Um, so, uh, I always made it a habit of, like, talking to people that I knew de facto were world-class and then asking them specifically, "What are the key traits or characteristics that you look for? What are the questions that you ask, and how do you... And then how do you find them? And, and if you're looking for the next person that's as good as you, like, where are... Where is that person working right now, and, like, what's, what's her phone number?" Um, I think that, like, learning what good and bad is during the interview process is extremely expensive. It's an expensive use of your time, and it's an expensive use of everyone else's time, so pre-calibration of that really matters. Um, and then once you have someone in sort of the interview process, uh, you'll build the process over time, uh, to both, um, screen quality. And so, um, at Pinterest, you know, we have an evolving set of standard questions that we're always rotating through, and we're always measuring are these good indicators or bad indicators of quality? But the other thing that the questions are meant to do is they're supposed to give a sense for is this the right place for that person to come and work? Um, and this is to the point you guys made about being very transparent about what's gonna be easy or hard. Really great people wanna do things that are hard. They wanna, they wanna solve tough problems. And so, um, there was a, there was a certain brilliance in, in Google setting out these interview questions that were thought to be really difficult because then people who like solving really difficult problems, uh, they come out and, and seek those. Um, I think it's really important even as companies get bigger that you don't whitewash the risks. Um, I heard that, uh, PayPal, you would go in and after interviewing with, like, Peter Thiel and Max Levchin, then they would say, "And by the way, like, Visa and MasterCard wanna kill us, and we might be doing something that's illegal. Uh, but if you succeed, you'll redefine payments." Um-Or when they were recruiting for iPhone, they didn't even tell people what they were doing. They were like, "You won't see your families for, uh, three years."

21. SASam Altman18:36

    [laughs]

22. BSBen Silbermann18:36

    "Um, but when you're done, when you're done, your kids' kids will remember what you built." Um, and I think that's a really good thing in recruiting as well, that you're very, very transparent about why you think it's an amazing opportunity, but you lay out in gory detail, um, why it's gonna be hard, and then the right people select in or they select out, um, of that opportunity.

23. SASam Altman18:57

    Um-

24. PCPatrick Collison18:57

    Evidence suggests it's worth allowing people to see their kids, though. [laughs] Um, I, I feel like one thing you have to do, uh, as you try to identify talent is have the, uh, the confidence to, to interview for it in a way that works for you. Uh, I think, uh, you know, if you're, uh, say you're not the, you know, the world's best engineer and you, you're, you're trying to interview engineering candidates, uh, I think it's tempting to try to cargo cult what everyone else does, and you know, get them to code on whiteboards and do other engineering interviewee things. You know, in the case of Stripe, hiring our first engineer, we, we flew the guy out and we, uh, spent a weekend coding with him and, you know, looked over his shoulder and kind of... The, th- It was the only way we could really tell and get ourselves confident, uh, that that person was good. And I think you can actually extend that to, to other roles where, again, you're not an expert in that, you know, I'm no business development guru, but when we interview people for business development roles, we'll ask them to do a project, you know, where they, uh, talk about how they would improve an existing partnership that Stripe has or which new partnerships they would go out and do. And again, you know, e- even though it's not my domain area, I, I am actually confident enough that I can, I can judge those pretty well. And, and I think people often have this, uh, imposter syndrome when it comes to, to interviewing for roles. Yeah, I, I think a pretty specific kind of just tactical thing, um, to, to do for, again, the first 10 people is to work with them as much as you can, uh, before committing to hiring them. I mean, once you reach a certain scale, it's kind of impractical because it's a huge time commitment on their side and also, I mean, it just would be unscalably expensive from your side. Um, but it, it's really worth it to get the first 10 people right. And so for a majority, maybe even all of the first 10 people, uh, we, we, we worked with them in some capacity, usually for a week in advance, and it's, it's pretty hard to, to fake it for, for a week. Um, uh, pe- people... It, it tends to be pretty clear, uh, quite, quite, quite quickly. Um, another thing I thought of, uh, you know, to, to the question of sort of how, how do you know, like, who's great or who's good enough or whatever, and people always talk about sort of this notion of the 10X person at sort of whatever the skill set is. Uh, I don't really know what 10X means. I think maybe the, the slightly more kind of helpful, uh, or, like, intuitive version of that is, is this person probably, uh, the best out of all of their friends at what they do? [laughs] Um, uh, and so, you know, it's a little bit sensitive to, uh, how well they choose friends and how many they have, but, uh, uh, for, for me at least, I, I find that kind of a, a more helpful way to think about it. Like, is this the best engineer this engineer knows? Um, and, uh, and the other thing I think that's actually probably just worth mentioning in all those kind of first 10 people or even more generally on the culture and hiring topic, um, I think everyone, uh, sort of doesn't realize until they go through it themselves how important it is, in large part because, uh, the, like, in life and the media and everything, people focus way too much on founders, and that, like, here we are, and so kind of we're reinforcing the sort of structural narrative that, like, Stripe is about John and Patrick, and Pinterest is about Ben, and so forth. Uh, whereas, I mean, obviously sort of the vast majority of what our companies do, like 99.9%, is being done by people who are not us, right? Uh, and I think, I mean, obviously that's kind of, it, it, it's obvious when you say it, um, but it, it's sort of very much not just the, the macro narrative, and, you know, companies are abstract. You kind of need to associate them with specific people. But I think it's worth bearing in mind that, like, for, you know, Apple, you know, everything was, was, you know, Steve Jobs was, like, a tiny, tiny detail at the end, right? Or Google with Larry and so forth.

25. SASam Altman22:17

    So don't screw it up is what you're saying?

26. PCPatrick Collison22:19

    Something like that. [laughs]

27. BSBen Silbermann22:21

    I think, you know, the only other thing I'll mention is that I think referencing people is really important. Um, and, uh, referencing people is just what it sounds like. You're basically asking people who have real, material working experience, um, for their honest opinion. And, uh, we, we do that really aggressively, and what we're trying to figure out is, what's this person like to work with? Um, uh, we're not trying to validate, like, whether they told the truth on their resume, 'cause we assume that they're telling the truth. So very standard questions that, that I'll ask somebody, um, is in an interview, I might, I might, I might say, "Hey, you know, we both know, um, Jonathan in common. I'm gonna talk to him, uh, you know, in a couple weeks, 'cause we're both social friends. Like, if I asked him, um, what, what you're the best at, or what you would be most proud of, or what you were kind of working to improve, uh, what would, what would he or she say?" Um, because it sort of tests a level of self-awareness and creates a bit of social accountability. Uh, and then I'll try to ask something that makes the question, which is typically very soft, feel a little bit more quantitative and then calibrate that over time. So, uh, you know, in evaluating this person on this dimension, do you think this is the top 1% of people you've ever worked with, the top 5%, and the top 10%? And it forces a scarcity that gives a materially different, uh, reference, uh, than if you just say, "Hey, like, what's the best thing about John? Uh, he told me, you know, he's good at these things. Can you validate them?" You're like, "Yeah, sure," 'cause they, they don't gain anything. So I think that's, uh, just a tool that people should, should take seriously.

28. PCPatrick Collison23:44

    Yeah. Yeah, and, and re- referencing isn't obviously easy to begin with, uh, but it, it does actually prove really useful over time. Uh, and you just have to... People, I think, uh, especially for named references, people really want to be nice, so you have to do things like create an artificial scarcity by saying, you know, "Where would you rank this person amongst the people you've worked with in this role?" Uh, or, you know, there's lists out on the internet of how to reference, but, uh, you should aim to spend, you know, 15 minutes on the, on the phone with that person, not just let them say, "Yeah, this person's a, this person is awesome," and then hang up.

29. BSBen Silbermann24:13

    But also a tremendous source of new recruits.

30. PCPatrick Collison24:15

    Yeah. [laughs]

Episode duration: 50:36