[0:00:00] For those of you who are in the minority and don't know who Sam is, he is the president of Y Combinator. He was the founder of Looped, which was a company that sold for 43 million. He was in the very first batch of Y Combinator and was also, when I went through Y Combinator a couple of years later, was the founder that Paul Graham would always speak about with the most awe.
[0:00:29] And at how old were you when you started Loop, Sam? I think 19. 19. So at 19, Sam was doing deals with U.S. mobile companies to put maps onto phones. This was pre-iPhone. And is now president of YC, an investor, and heavily involved in AI, amongst other things. That is all correct. Cool. So if you'd like to at least put your hands together to welcome Sam. Thank you.
[0:01:04] So Sam, I'd like to get straight into the nitty-gritty of startup company building stuff. YC is a very product-focused incubator, probably the most product-focused incubator around. The strap line being make something people want. The hacker dream of building a piece of software that you launch out through Product Hunt and Hacker News
[0:01:33] And that then grows into a large company like some of the YC companies like Dropbox and Stripe have done. To what extent is that a reality these days? And how frequently do you see that happening versus companies coming in and building sales teams from the get-go? Well, all of the really great companies, although they often figure out ways to make things go faster with sales teams and everything else, I think generally win with a great product.
[0:02:02] It's very difficult to point to a super successful startup that didn't have a product that people love so much they proactively told their friends about. And so I think no amount of growth hacking, no amount of sales funnel optimization will save you if you don't fundamentally build a great product.
[0:02:24] But because that's really hard, everybody looks for alternatives. And the advice no one wants to hear is go make your product better. People always want some sort of trick. They want to believe like, if I can just unlock the secret to email marketing or PR or something else, then I'll be great. And that stuff is important. You know, you do need to kind of jumpstart the whole effort to some degree. But I think it's so easy to blame.
[0:02:53] The cause of things not working on anything except a great product that people try every alternative. So I think everything you said is important. And we do try to teach founders about that. But the thing that we really try to do is find people who are going to build great products and then help them do that. And I think that is more important. That is, to a first approximation, you cannot succeed without a great product, no matter how good you are at everything else. And if you have a great product, you can still fail.
[0:03:22] But those other skills seem easier to tack on. So where have you seen the right balance between the artisanal development team who focus on what they view to be the great product, but perhaps don't get the traction and exposure to verify whether it really is a great product, and those that manage to take it all the way through? Because the danger, if you focus relentlessly on products, of course, you potentially never leave the house with it.
[0:03:52] So how have you seen successful teams drive through? Well, I think it's impossible to build a great product without a lot of contact with your customers. So I have never, I know it's theoretically possible, but I have never seen a team go lock themselves in there flat and come out two years later with a fully formed, incredible product that people love. I think you need the constant feedback cycle with...
[0:04:19] With customers, in fact, I think the best companies are often the ones that make the feedback cycle the quickest. They build something, they get it in front of their users, they watch how they use it, they listen to what they say, and then they iterate. And if you can get 10% better each cycle, and one company has a cycle time of a month, and one company has a cycle time of a day, you end up in wildly different places a year out. And so I think if you look at some of our most successful companies,
[0:04:47] Airbnb and Striper, too, that come to mind for being incredible at this. They took the cycle time back from build something, get feedback, fix it, down to incredibly short periods. And so I think it's impossible to build a great product without doing this. But the kind of canonical great founding team is one person who wants to build product, one person who wants to go off and talk to users. And that combination works really well. It can happen in the same person.
[0:05:17] But I think the company needs both of those things going in a cycle. Can you talk about some of the teams that have done that, perhaps going into Airbnb and Stripe? And even Stripe, I mean, how many years was it before Stripe publicly launched and started growing? Well, Stripe had users very quickly, but they were in private beta. Because the financial system was so regulated, they couldn't take on too many.
[0:05:48] Stripe officially launched several times. They kept doing it to broader and broader groups and getting more and more press. But probably what they would count as their official launch was a year, year and a half in. But they got users really quickly, one user and then three users. And what they would do, in fact, this became so popular that we named it at Y Combinator. We call it the Collison install. The brother's last name is Collison. They would find a company they thought should use their product. They would take their laptop out of their hands. They would install it.
[0:06:18] And so you have no excuse not to use this. We have it in your development environment now. Start charging. We got a bank account for you and everything. And then they would follow up with them every day, and they'd build things for them. And eventually, they got a bank to say, OK, you're doing this. We didn't think it was going to work. It's working. You can open this up publicly. But although they took a long time to launch, they got users very quickly, and they started iterating for those users very quickly. What do you think?
[0:06:46] has characterized the false positives that you've seen. So in so many of these things, you always end up reflecting on the outsiders like Stripe and Dropbox and so on. But what were the companies that early on had the same characteristics but didn't execute in the right way? And how did you see that play out? I mean, most companies are false positives. One of the things that's a little bit depressing about startups is you can do everything right.
[0:07:17] But pick the wrong market or someone else just does something a little bit differently and it works better than you. And you can be a team that was as smart and as hardworking as the one that won. And for factors where you got a little bit unlucky or things that were a little bit out of your control, you can still fail. So we try not to overcorrect for false positives. I think one thing that investors get really wrong is they see a pattern fail once and then they are like, you know what, I'm never investing in another home sharing company again.
[0:07:47] And that's really dangerous. So we talk a lot about false negatives. Those are terrible. We talk a little bit about false positives. But I think it's very dangerous to talk a lot about false positives because a thing that didn't work in the past can always work again in the future. And a person that's failed their first time can succeed wildly their second time. This is not that well known, but many of our most successful founders,
[0:08:17] We're coming through Y Combinator for the second time on an idea that VCs as a whole had decided was dead. So Drew Houston first applied to YC with a very bad idea. Didn't get in, thankfully for him, because it would have been bad for him to focus on that. And applied with an idea, file sharing, that all VCs had decided was long over. And it would be very easy to not fund that company, and that would have been really bad.
[0:08:44] Many of our other successful companies have a similar story. So taking Dropbox, which happened to be in the same batch I was in, although Dropbox got retrofitted to the lean startup movement, but having seen it on the inside, they were basically right with their initial hypothesis, the product that the company wanted, and they continued executing on that and executed on it really, really well and built a great team around it. But they didn't actually fundamentally in the...
[0:09:13] The first couple of years pivot much at all. They just kept going. To what extent have you seen the really successful companies be a result of dramatic pivots versus actually keeping refining the original hypothesis? Dramatic pivots almost never work. There are exceptions, and the exceptions usually come when the pivot is what the founder wanted to build in the first place. I think that Instagram is the most common exception of this.
[0:09:43] An example of this where they started off building this thing that was like a location-based social network. It wasn't working. And then Kevin Systrom went on vacation. He was really bummed. And he's like, I'm just going to build something for fun. And he still is a very passionate photographer. At the time, iPhone cameras weren't very good. And so he's like, well, how can I make iPhone cameras more satisfying to use given this limitation?
[0:10:06] And he came up with the idea of filters as a way of addressing the problem with low-quality iPhone images and still making them sort of aesthetically pleasing. And then obviously you want to share them. So that was a pivot that worked, but it came out of the project that the founder always wanted to build. The kind of pivot where you say, damn it, our startup's not working. We have all these investors. We have all these employees. We have to pick something else and stand up at a whiteboard and just pick a made-up idea.
[0:10:35] They must have worked sometimes, but those are really rare, certainly in the Y Combinator experience. None of our most successful companies have that history. The thing that does work, I think, really well are sort of successive approximation pivots. So Stripe started off wanting to be a bank and then realized that that actually wasn't the problem. They could get a bank account, but they couldn't take payments. And that was really the thing.
[0:11:04] If it's on the same vector of the mission, but you find out that there's a better way to do this thing that you wanted to do with the original idea, those pivots work all the time. So going back to the... At the same time as you want to write off the false positives and write off the concepts, because timing is everything for a start, are there things that you see... You've got so much exposure to startup patterns.
[0:11:32] Are there things that you see as obvious failure patterns in companies that inset themselves in those YC companies that don't tend to get talked about more broadly? Sure. Some of these may get talked about broadly, but I'll just sort of say the most common problems. A huge one is the team falls apart. I think during Y Combinator, during the three months of our program, the number one cause of failure is the team falls apart.
[0:12:02] And that happens a lot. And the reason that it usually happens is one person on the team kind of gives up hope. And usually what happens is there's one person who's really determined and is going to see this through no matter what. And the other person kind of got pulled into it, but it's just doesn't want to do something that's going to be really hard. And when it becomes clear that it's going to be really hard, that person leaves. And it's usually in a fairly dramatic way and sometimes a huge fight over the carcass of the company.
[0:12:32] We ran an experiment once where we let people add a co-founder that applied as a solo founder after YC started. We don't do this anymore. And in one batch, there were like, let's say, 12 solo founders and nine of them added a co-founder before YC started. And all nine of those, the team fell apart and the company failed within the first year. So one thing we've learned more generally, although that's a stark example, is that founding teams that haven't known each other for a long time,
[0:13:02] And have a long history together don't usually work. And in fact, to the degree we let ourselves pattern match it all, founding teams that haven't known each other is something that we're very, very nervous about. And that's a major cause of failure. A more general indication of what I mentioned earlier, and I think probably the easiest way to trick us, or our biggest systemic flaw, is we are bad at evaluating determination of founders.
[0:13:33] And we say this all the time, but it's really true, that determination is the most important predictor of success for startup founders, more important than intelligence, market, product, anything else. And it's relatively easy to evaluate intelligence. Markets are harder. Products are really hard.
[0:13:52] But determination, I think, is the hardest of all. To talk to someone, especially someone who is younger and does not have a lot of evidence in their career of being determined they can point to and to try to evaluate that is really, really hard. And that is probably our biggest cause of error still in our selection process. We have all kinds of crazy ideas about how to get better at that. None of them have worked so far. I'm sure we'll figure it out someday. I think another thing that...
[0:14:21] What goes wrong, and this sounds really obvious, is that people never figure out how to make a product users love. And sometimes this is because there is no product in the space that anyone wants. But more often, it's that the founders focus on the wrong thing. So I think one of the hardest problems any startup faces is of the 100 things you could do this week, what is the most important number one thing to do? That's a hard question. And then can you make yourself focus on nothing but that?
[0:14:50] That's a really hard question. And most startups just don't do either of those well. Almost all of the startups that make really great products do both of those things really well. They figure out the exact most important thing to work on. They don't work on anything else or very nearly nothing else. And a lot of startups kind of get caught up in all of these other things that are urgent but less important. And they lose sight of the thing they really have to do. And those startups...
[0:15:19] can look along the way like they're being very successful. They're getting great press, they're raising money, they're off meeting customers. They're doing all of these things that to an untrained observer or someone watching a movie about startups or something would look like they were doing all the right things. And then it doesn't work at all. One of the things that I've observed in spending time in San Francisco and London is that startups in London will tend to pursue larger customers earlier.
[0:15:49] Then startups in San Francisco. Because there's less funding than there is in San Francisco. And that quite often leads to... A large customer forever presents a challenge to a startup because they're obviously a source of revenue, but they're also a source of demands. And when do you think... To what extent do you think that good startups can kill their potential to...
[0:16:15] Early on, by going with a large customer too early, and what's the right way to scale through customer size? Yeah, this is a tough one because I really do fundamentally believe you're better off starting with smaller customers. But I totally understand your point that in London that's harder to do if you need to raise money. You know, I've noticed that whenever I'm on a long trip out of the States like this and I get home and I get off the plane and I'm walking through SFO and...
[0:16:47] I feel better. And I was trying to articulate what about it I really like. And the thing that I think is magic about Silicon Valley for all of our many, many flaws is that there is a belief that you can just sort of feel in the entire area and most of the people you come in contact with that anything is possible.
[0:17:12] No idea is stupid. However improbable something sounds, it might be the next Google. And so you better not ignore it, even if the company only has two small customers you've never heard of. And we can build general AI and electric cars and spaceships and nuclear fusion reactors and Facebook and Google and anything else. And in an environment like that, people aren't that skeptical because they've been burned by it in the past. And there's this wonderful optimism about the future.
[0:17:39] And not being dismissive. And in most of the rest of the world, if you start a startup, for whatever reason, and I think it's different in different places, most people will think about why you're going to fail. And then when they talk to other people about your startup, they feel cooler if they say you suck more. And in an environment like that, there is a huge premium on big customers. Because the constant sort of
[0:18:09] Vacuum pressure is against you. And I think you should try to resist that as much as you can, even if it means it's harder to raise money, even if it means that you meet a lot of people who tell you you suck. Now, there are practical limits, and if you can't raise money without a big customer and you need money, you might have to do it. But there's all this talk about, like, what is the secret to Silicon Valley? Why can no one else replicate it? And this is it.
[0:18:40] If you tell people in Silicon Valley that you're going to change the world and you have one tiny customer, they might still believe you. And that has been the path of most great companies. So do you continue, I mean, I just want to go a bit deeper into it because it is a really strong difference between Silicon Valley and the rest of the world, not London, but the rest of the world in general. That path and that belief that keep focusing on the simple things.
[0:19:10] And the bigger things will follow. Do you see that continue to play out? I mean, one of my favorite YC companies is Weebly. They started small, raised a few hundred thousand dollars, and then raised their next round at a valuation of half a billion or so. And just built great company and great product all the way up. Do you see that continuing to play out in the companies around you? Yeah. That's happening more and more, actually.
[0:19:43] As startups have gotten easier, as there's more knowledge available, as there's better tools and you can do more with less, this thing that used to happen, and it used to be very rare, but happened with Weebly, where a company like the founders still own, or the founders and employees together still own like 92% of the company six years in, and the company's worth billions of dollars. That is becoming more common and more doable.
[0:20:10] The good news about Silicon Valley is that capital is cheap. The bad news is that everything else is expensive. So it's really hard to hire people. Office rents are crazy. And so some people have said, you know what? Although this capital sounds cheap, it's a bad deal because anything I spend it on is going to just be so expensive. So I'm going to raise only a small amount of money, not have a big team, and get a lot done. WhatsApp is a very famous example of this. And I think you see more and more companies doing things like that.
[0:20:38] So let's drill into that and look at an example. Let's take an example company. I mean, I think there are the companies that have huge natural inherent growth, like I think WhatsApp did. Facebook had growth right from the get-go. But then there's obviously a lot of middle ground companies that aren't obviously growing and aren't obviously locking down huge customers at the same time as well. That's the more challenging path to take because it's not necessarily obvious that
[0:21:09] If you're selling stuff at $10 a month, that there are going to be a million customers to pay for it. What do you see the right kind of strategy going through that and getting more customers and knocking those milestones out? You know, I think people just need to be a little bit more patient and a little bit less entitled. The people make fun of this for millennials about, like, I need to be successful tomorrow. But you see it with a lot of people starting startups.
[0:21:39] Giant company. They never grew at like thousands of percent like most people want. They just doubled every year for 30 something, 40 years. And that turned out to work really well. Reddit, which is a company I'm on the board of, just kind of doubled every year for a really long time. And the thing that people forget is like in the world of startups where, oh, you only grew 2x last year. What a failure. You know, like if you double.
[0:22:08] Every year for 10 years, I think 2 to the 10 is like 1,024. It's like a big, like you grow 1,000x. Or maybe I'm off by a little bit on the math. But it's like, this is a big number. And so I think if you are building something that is going to have defensible long-term terminal value, which most things don't, but if you are building something in that direction where you're going to have a real monopoly or lock-in of some sort, it's okay if it's like Reddit-style double every year for a long time.
[0:22:38] Not as only okay, it's great. People just need to remember that that's great. The hard part, of course, is if you're not, I mean, if you're on that doubling trajectory, then it's great. But it's when you're on the flatline trajectory, pre-doubling, what are some of the companies that you've seen go through that have, I mean, I guess Airbnb to an extent, but perhaps some of the less visible ones that people don't know that have flatlined for a long time and then found their groove and picked up?
[0:23:06] I mean, Airbnb, since you just mentioned it, they talk about the 1,000 days in the darkness. This is longer than any rational founder would accept eating dollar store cereal every day with max stock credit cards, but they did it. And it took them 1,000 days of flatlining before it really started to take off. And this part always gets edited out from the story. But it took them a while to get the product right.
[0:23:30] And even then, when they start growing and you have one user and you double, now you have two users. It takes a while. And one of the things that we were able to help them with at Y Combinator was how to get out of that flat line. But most of any successful company, when you look back at their growth graph six years in, looks flat for the first two years. That's just how any exponential curve works. But most of them are even flatter than that. And this is why determination is so important.
[0:23:59] It is so hard to keep going through that. It's so easy and so tempting to give up when people are telling you you suck and you have no users and you have no revenue and nothing's working. And sometimes the right answer is to give up. But what the companies have had in common is even if they were mostly flatlining, they had a few users who really loved them. Even though Airbnb didn't register on a graph, the seven users they had
[0:24:29] Like, talked constantly about how much they loved Airbnb, how important it was to their lives. Same thing with the early users of a Stripe or Instacart in particular. I remember that really well. Instacart is this sort of grocery delivery thing. And the grocery store they started delivering from was right near my house. And so people in my neighborhood started using it. And including Jessica Livingston, one of the co-founders of Y Combinator.
[0:24:56] I remember her saying something in particular, like, if you take this away from me, I will have a meltdown. And so I think the right indicator is not, are you growing like that? And you have to be brutally honest with yourself here, but it's like, do your small number of users desperately love what you're doing? And if so, you can often, usually even, get through the flatline and figure out a way to make it grow. And if not, I think you're very likely working on the wrong thing. Have you seen, I'm not intending to be,
[0:25:27] Perversely contrarian, but I think you can always learn from the corners that you don't look at. What are the companies that you've seen have a small number of users that were passionate about them and still didn't make it through? This is always hard to remember because you kind of question your own memory. Like, did the users really like it that much or not? And I also don't like to point out specific companies that have failed and talked badly about those founders, so I will decline to answer that one on two counts.
[0:25:58] Okay, fair enough. You put a tweet out, I don't know how recently it was, but I really liked it. You said that scaling before product market fit slows the company down. Scaling after product market fit speeds it up. I thought that was really interesting. Why do you... Do you want to just talk a little bit more about that? The basic idea is that a company can be like a little jet ski or an aircraft carrier.
[0:26:29] Depending on how many employees it has. And there's things that jet skis are good at and there's things that aircraft carriers are good at. And if you haven't yet figured out the right product, then you would like to be a jet ski, right? And the overhead that you put on the company is bad because you want a little boat that can turn fast and keep trying new things and zigzag around until it works. And aircraft carriers are terrible at that.
[0:26:56] So once you hire, let's say, 50 or 100 people, and the product is not working, now you're trying to turn an aircraft carrier and then turn it again and again and again until you get to the thing that works. And just the organizational inertia usually makes that impossible. There is a reason that companies that scale up almost never successfully pivot once they've scaled up. It's hard to turn the company.
[0:27:26] Till you have really got the company doing the thing you want, until you don't have to make any more turns, you would like to be like a jet ski. On the other hand, you can't land an airplane on a jet ski, right? There are things that aircraft carriers can do. And so there's some point where you're like, I am sure this is the direction we're going. We're going to go really fast. And we are optimizing for might and just raw power. And then, yeah, you want a big organization.
[0:27:54] As you're going through that product market fit phase, there's a constant tension between the things that customers want that are, let's say, more horse-shaped car features and the direction that you want to take the product that can potentially lap those features. Have you seen people create the right trade-off between those two demands, between answering the short-term requirements and saying, hang fire?
[0:28:25] You're going to see in the course of the future. One of the most difficult skills to learn as a founder is when to ignore data and when to not listen to your customers. And I don't think anyone is born with a good intuition on that, or at least I was not. But you develop it as you operate. And I think it just takes years of experience to know when to trust yourself and when you have an explanation that is likely to be true.
[0:28:53] About why to ignore your customers. This comes up for me at Y Combinator all the time. You know, I mostly try to do what founders tell us they want. But some of the most important things I've done are when I have high conviction that they're wrong. And that skill, I think, is just one that gets built up via experience. And you'll be miscalibrated initially and you'll learn some painful lessons on both sides and you kind of successfully iterate. So just a case of trying it and getting a feel for what level of gut is right.
[0:29:24] Many people never get good at it. It's a really hard skill to learn, but the way I got better at it at least was, yeah, trial and error. What have you seen? I think there's a not super often talked about contrast, huge contrast between what's an optimal strategy for an entrepreneur versus what's an optimal strategy for an investor. And so much gets written about the optimal strategies for investors, and Paul wrote that.
[0:29:52] Essay on Black Swan Farming that really sums it up in that you're wanting to optimize this small number of extremely high return companies. But on the flip side of the coin, obviously, as an entrepreneur, when you can't process 80 companies in a lifetime, you have to optimize for the success in the short term as a founder. What have you seen being, I mean, you look at things through the eye of an investor, obviously.
[0:30:18] But what have you seen be the optimal strategy and the optimal balance for the entrepreneur in maximizing the probability of success? You know, this is really different for different people, and I think it's impossible to give a one-size-fits-all answer. There are some people who, they don't care how many times they fail, but there's problems they're going to work on, and that's it. And if one of those works, it's going to really work, and if none of them work, they'll be okay, because at least they worked on the things they want to work on. And there are some founders who...
[0:30:47] If they're perfectly honest with themselves, one, financial security, and they don't care about making a billion dollars, but they want to be damn sure they make $10 million at some point in their lives. And that's the most important thing to them. And maybe if they do that, then they'll go take a risk on their second startup. And I don't think any of these things are wrong. I just think founders have very different primary motivations and thus very different optimal strategies. And I think the question is,
[0:31:16] What you want and just being brutally honest with yourself. You don't have to tell anybody else. You should probably tell your investors at some point. But you should answer this question for yourself and think about what you want and then pursue that strategy. And I've seen people at the extreme opposite ends of the spectrum and people somewhere in the middle. And I think people are just honestly quite different on this topic. Have you seen investment structures that actually...
[0:31:44] Cater for the different ends of the spectrum. So much of the investment structure is catered towards the large outcome, large exit. Have you seen founders go in and say, listen, I do want to make a $10 million exit, and I plan to do that in the next eight years. And here's how you can make money even at a reasonable valuation through that. Yeah, look, I think that doesn't work for the Series A model, at least in the US. But many seed investors kind of specialize in that now and are happy with it.
[0:32:11] And how does the structure of the terms look different from regular documents? Well, the main thing is that you just raise much less money at a much lower valuation. There are some structural differences. Like if you're planning to do something like this, you know, maybe like dividends really matter or something. But it really is just like setting the expectations appropriately of your investors up front. And, you know, it's sort of a risk reward and time frame for the company that matches up with what they're expecting and what you want.
[0:32:43] I want to touch a little bit on machine learning and AI. There is a definite degree to which, certainly around here, it's the gold thread of the emperor's new clothes at the moment. You say that you're weaving it in, makes the startup hot regardless of what's going on. To what extent do you think that machine learning and AI are...
[0:33:04] Are going to be viewed in the same way that mobile is now viewed now, in that you had to be a part of it, and you almost certainly had to embrace it into your company. And to what extent is it use case specific in what you're seeing? You know, I also have a similar reaction. Anytime someone tells me we're going to be the machine learning for X startup, I sort of immediately think fraud. However...
[0:33:28] I think that it is the best idea to work on. But it's the companies that really, truly deeply understand machine learning and are applying it to a specific industry try to keep that relatively secret because they know how powerful it is. And so they don't usually lead with, we're an AI company. They're like, oh, yeah, we're going to optimize shipping because that's a good industry. And then you look at it, and they have three brilliant deep learning people on the team. So I think the companies that talk about it the most
[0:33:56] It's the most emperor's new clothes. That said, I think it is the single best thesis right now in starting a company is take machine learning and apply it to any specific vertical you want. To what extent, I mean, my first, I started a PhD in computer vision and I would feel pretty hesitant about getting stuck into machine learning. What do you see, do you see companies that don't currently have a history of it moving into it and making the most of it effectively?
[0:34:26] Or is it getting stuck in a tar pit for many? No, I mean, I think that the big companies like Google and Facebook are doing the best work and a lot of other companies are figuring out where it makes sense to add it. But again, it's so easy to fall into like the buzzword tar pit there that the people who are doing it really well, I think, are wary of making too much hay about it. And for people here who are running companies or...
[0:34:54] Funding companies that are going into machine learning and AI stuff. Is there an extent to which they're now developing a set of technicians of AI who will apply it and machine learning will apply the TensorFlow and the other pre-built technologies that are out there versus investing in actual core?
[0:35:18] Core design of machine learning systems. You know, I suspect that the answer will be really different for different companies. Cool. I think, shall we switch to some questions from the audience? Great. So if you'd like to stick up your hands. Do we have a roaming mic? Great. I guess on the point about focusing each week narrowly on the most important thing, two questions around that. Which methods?
[0:35:55] Have you seen being the best? Like, you know, you sit down Monday morning, you go through the whole list and then that's it. And second part of it, how do you merge that with, let's say we have a monthly goal and these can be like sales, customers. So how do you bring that all together? So I think the best possible way to do that is an external advisor you really trust who has a lot of data in their heads about startups saying, you know, here's our goal three months out or this month. Here's the current state of the startup.
[0:36:23] What could we do that would have the biggest effect towards our longer term goal? And that, honestly, is mostly what Y Combinator does. So trying to find someone who can give you an external read with a lot of data, I think, is really valuable. If you don't have that, I would try to do it yourself. I would say, OK, what's the most broken in the company that's preventing us? What is the thing most likely to make us miss our monthly sales goal, if that's even the right goal?
[0:36:51] Again, one of the reasons external people are good is that so often companies are bad at setting the right goal. Maybe you have a sales goal, but what you really should have is a product goal. But then say, what is most likely to make us miss that? And you can try these things, like do it every Monday morning. But in practice, startups are always so chaotic that
[0:37:18] I think you should just do it whenever it comes up and then wait until you feel like you've addressed it. Hi, Sam. I'm doing a lot of hiring at the moment and so I'm thinking and talking a lot about employee options. I'd love to know your thoughts on, I guess, any advice for how to think about employee options. What's really tough is how much to give and also how to think about it versus other of your senior team. And what have you seen people do badly? The common mistake people make is they don't give near enough.
[0:37:51] Equity to employees. And I think most people are willing to be very generous giving equity to investors and very stingy with employees, and it should be the opposite. So my very rough math is that the average company should give like 10% of the equity to the first 10 employees, which most companies do much less than that. Our best companies have certainly given more equity out. Beyond the first 10 employees, you just run into a problem with...
[0:38:21] Maths, and you have to do it differently. But I think you should err on the side of more equity going to employees. In terms of senior team or not, people have very different philosophies. It used to be in the US, at least, that the custom was that VPs and up got a ton of equity. Engineers got almost no equity. And now that's flipped. Now you'll often see a star engineer get more equity than the VP of operations or VP of sales or something like that.
[0:38:51] I think a good framework for it that people will still be unhappy about, but is at least fair, is to think about how much expected value that person will add in the next four years to the market cap of the company. And even if someone's more junior than somebody else, they should get more equity if they're likely to add more value. And I think that's the fairest overall framework. So just following on from that, Sam, you've got some fairly non-standards.
[0:39:19] Certainly, perhaps generally by the Valley, opinions on vesting structure and exercise period for options. Could you talk a little bit about what your thoughts are on that and why? The main thing, I think, is that people who work at your company and invest equity own that equity, whether or not they leave. And there are all these tricks of the law in the US where you can make it really hard for them to own that. But I think the right way to think of this is it's compensation for the work they did. And even if they leave, they should still get to keep it.
[0:39:50] I also think the vesting schedule is a little bit screwed up in that it incents people to stay a year and then go somewhere else and collect a bunch of different lottery tickets. And I think you really want to reward the employees who stay a long time heavily because that tends to be how the companies that really win big win. So those are, I guess, the main two things. If I were starting a startup today, I would do like a five- or six-year vest weighted to the back of it.
[0:40:21] So maybe like 5% or 10% invested in the first year, and a lot of it invested in years 5 and 6. Oh, so this is like, you sure you want this? This is extreme detail of US tax law. No, we have this, it's a quirky tax issue. It's not the exercise price. The short version of this is, in the US, unlike most other countries,
[0:40:55] If you exercise an option, you have to pay for the option, which you can usually afford, but you are then taxed on the spread of the option price and the current fair market value at a very high rate. We have this thing called alternative minimum tax that you usually can't afford. Hi there. So I have two questions for you. So the first is, if you look at the YC application, and you look back to what you just said about Silicon Valley, I think YC is pretty clear that they would like those folks applying to sort of commit that they'll re-domicile their company as a U.S. company.
[0:41:26] Presumably sort of coming and giving talks to the UK and whatnot is to some extent trying to kind of build the YC awareness. Can you talk a little bit about, you know, potentially the idea of inculcating YC abroad as opposed to kind of replicating or deepening it in Silicon Valley is my first question. And then my second question is you talked a little bit about the startups as your clients, if you will. You know, what would be the top five things you would recommend to WIRA to kind of think about kind of?
[0:41:55] Replicating. More than half of our startups that we fund, or maybe just about half, come from outside the US. They're non-US startups. And we've thought about opening offices elsewhere in the world. And we may someday do it. In fact, we're considering one location in particular quite seriously right now. One thing we have found is that
[0:42:23] The best founders are willing to move for three months. It's not that long, right? You come for three months, you go back. And we don't mean it as a filter of seriousness, but it has worked that way. The number of companies that we have wanted to fund but have been unable to because they didn't want to move to the US for three months is low. Sorry, go ahead. It's the re-domiciling the actor. Oh, you know, I think...
[0:42:54] To a first approximation, every entrepreneur that I have seen optimize for taxes on anything, this usually comes in a conversation about government tax credits for engineers, has not been very good. So the best founders are...
[0:43:12] Trying to make their company several orders of magnitude bigger, not save 10% on taxes. So we will very occasionally, if there's a good reason, let someone stay as a foreign corporation. But honestly, most other investors won't. And most investors that will pay the highest prices and dilute you the less are in the US or invest the most. And so we used to let companies stay foreign. And we were doing a disservice to them. So we'll still do it.
[0:43:40] If it's really important for some particular reason, and that comes up maybe three times a year, something like financial service law or something. But I generally think that most startups in the world today, and this could change next year or the year after, but most startups in the world today should be U.S. Delaware C-Corps, no matter where they are in the world. And there's a new service called Stripe Atlas that makes that really easy. But we asked, you know, speaking of asking our clients,
[0:44:08] We asked a whole bunch of our foreign founders, like, are you happy you stayed a foreign company? Do you wish you had converted? And almost everybody, and certainly everybody good, wished they had converted. The people that stay for tax credits, I think, are just focused on the wrong thing. So even if we do open other offices around the world someday, at least looking at the world today, we would still recommend people be USC Corps.
[0:44:38] Top five things that founders tell us? To Wira. Well, I only have one, but it's a big one. The thing that YC has done that is the most... Actually, I'll give two. You're not going to like one of them, though. One thing that we have done that is really different from other startup accelerators is that we always do the right thing for the founder. Even if the founders come before our relationship with other VCs, they come before our own returns.
[0:45:15] We have this long-term greedy, short-term, super generous mindset where effectively all of our returns come from the $10-plus billion companies. So this thing that most accelerators or most investors do where they fight for an incremental gain and a $20 million outcome, we never, ever do that. We just give the stock back in those cases to the founders most of the time. If a founder violates our ethics policy, we just give the stock back. We don't want to deal with it.
[0:45:44] But we always, always do the right thing for the founder. This makes people feel incredible loyalty to YC. And fundamentally, what YC is, is this community and this network effect. And so we have taken, probably further than any investor I know, this idea of you do the right thing for the founder, even when it's not in your own interest, and eventually that will work out. Very few other accelerators I have ever met do that. And they fact like, you know, a few times, we don't.
[0:46:12] We haven't invested in too many companies that another accelerator has also invested in. But when we have and a situation like this comes up, they can't wrap their head around it. It's like they assume that someone's hacked our email. Then they think that's more likely than the fact that we're willfully giving our stock back. So extreme founder friendly is, I think, still one of the most misunderstood secrets about us. Another one is office space. So we don't provide shared office space.
[0:46:40] And I think that's really important. I think startups need their own space, their own culture. You need to develop this feeling that it's sort of you, especially against the world, not you, one of other many interchangeable startups. And most importantly, I think you need to be in a space, you basically need to be in a safe space for bad ideas. And if you're by yourself and you're doing this thing that sounds bad, maybe no one tells you it sounds bad for a while.
[0:47:06] Or no one makes you doubt your conviction. And in co-working spaces, that tends to happen. So some of the other accelerators I've visited have taken this to heart. They took their once open co-working space and they chopped it up into offices with separate doors. And they've had way better results since then. But I think that's a really big one. Another thing is good optimistic advisors are just hard to find.
[0:47:39] My limiting reagent and how quickly I can scale YC is how quickly I can find good partners. And it's really hard. And I don't think, like I should probably be spending half my time on that. And in fact, I do probably spend a third of my time on that. But this is like, you know, in a very real sense, like the product of an accelerator is its partners or the advisors. And trying to find good full-time committed people to that is hard and really important.
[0:48:10] Two more? Three more. Yeah, okay. I really like the jet ski aircraft carrier analogy. Would you say when trying to make that transition, you look for measurable metrics or it's more of an intangible feeling you get when talking to the people that are using the product? Yeah, I think you don't want to scale up until both the fact that users really love it, so you have great usage per user, and you've actually figured out the user growth problem. So you have great usage per user.
[0:48:45] And go to top line growth. And I think you want both of those together before you're like, all right, we're going to now scale up the company itself as fast as we can. I was really interested by your comment on when you're at an early stage, especially B2B, not necessarily pursuing the biggest customers, but pursuing the right customers, which is very different to what you see here in London, especially from advisors and mentors as well. So in your opinion, I guess, what is the right?
[0:49:25] I actually don't think size is the most important metric. I think the two things that really matter. One is how desperately does this customer need your product? And the more desperate they are, the way better. Because the more they'll put up with the problems of working with a startup. And two is how closely will they work with you? The important thing about your first customer is not...
[0:49:54] I was just wanting to ask what your process is currently for measuring determination in founders that you bring into the program. What is it specifically that you find hard about that? And also to what extent...
[0:50:26] Should founders themselves look for that same determination in employees versus having more of a cultural mix? You know, I think the thing I have found that is best, and I will preface this by saying I don't know anything that's really good, is to ask about things that that founder has done in his or her life that are the equivalent of running through brick walls. And depending on the life circumstances someone's born into, that can be, there's a big...
[0:51:01] So you try to adjust it for the hands someone was dealt. If the company's been around longer, you can talk about the challenge the company's overcome. But this is a case where I believe, generally, if you can properly adjust it for someone's life circumstances, that the past predicts the future. So I ask about questions like that. You also just get some sense for the force of will someone has when you talk to them. But that's harder. I think for early employees, you do want this as the company gets bigger.
[0:51:28] It's maybe even a bad thing. It's at least you don't need it in everybody. Cool. Thanks for having me. Listen, Sam, thank you so much. I think the number one thing I gauge on a successful speaker is the information density that they present at. And you present extremely high information density with very low gaps. So thank you very much for coming. Thanks again. And please put your hands together for Sam. Thanks, Sam.