Raising over $300m, perfection vs. iteration & starting at eleven w/ Vercel’s Guillermo Rauch | Ep 7

Yeah, well, thanks for having me and great questions to start off. I think at the very basic level, I just always love to build and learn from people that are building. Even someone that is just getting started with their journey, there's something to learn from what they're doing. Maybe they might be using Vercel and I can get their feedback. There's this insight that Patrick from Stripe shared recently that I loved, which is why does it make so much sense to focus on selling to startups? And there's many reasons for that. But for example, Jessica mentioned you might love innovation and you want to be a differentiator of it. But something that I particularly love about startups is they really push you. They push your product to its limits. You have no patience. They're quick to switch if something doesn't go great. And so even though the platform itself over longer periods of time will make most of its revenue from large enterprises, and this is the kind of common journey of most enterprise software businesses, the ideal customer to actually kick the tires on your product is the startup, is the entrepreneur. So selfishly, I love making the product better for Vercel, and so I take in every opportunity that I can to hear folks that are using it or in the early innings of building something. And so much of what we want to do is enable folks to go from idea to app. So whatever idea you have, I want to know if Vercel is a good fit for you. And like I said, I just love the game too, right? For the love of the game. And I love being in the arena and learning new things from founders and hearing about their enthusiasm and how they're thinking about building software when there's no baggage, if that makes sense. Like every generation of software finds new fat to trim and ways to speed up and accelerate. So yeah, for many of those reasons, I find myself spending a lot of time with founders.

Yeah, I think a huge part of building a successful company is always going to be recruiting and ironing out your people skills. So if you think about it, when you make an angel investment or a pre-seed investment, it's so much about the people. It's so much like a job interview almost, right? Like you're trying to assess is this someone that I would love to spend time working with? Is it someone that's going to inspire, push me, that I can learn from? So I have a couple mental frameworks for this. Like one is I'm a very busy person and I have way too many ideas to pursue. So there's two things I could do. One is just write them down and then wait until someone does it or wait until someone comes with a similar enough idea and we can totally mind meld. And that gives me a ton of energy because I can start sharing insights and things. Think of it as a multi-threaded processor and my main thread is Vercel, right? And I have all these other lower priority threads where when the computer is idling, maybe I delegate some work to the other thread. And so I'm cooking on ideas or things that are possible or things that I'm learning about that are getting traction, things that are adjacent or maybe things that are possible to build on top of Vercel. So if I had infinite time, many of the things that entrepreneurs that come to pitch me and I invest are doing, I would be pursuing myself. And that has served me really, really well as a sort of heuristic. It's not perfect, but I think it's a great heuristic for at least this is going to be an investment that I'm going to put my energy towards and help people out with because I have a passion for it. So that gives me energy rather than takes energy away. And then the other one I think also is fundamentally, as I mentioned, you're thinking about companies that one day could become partners of yours, right? And so you're kind of trying to iron out those skills for could this be the right person to go pursue that, very much like you would be doing when you're hiring someone.

Yeah, something that seems very clear to me that gets as close as possible to a perfect characteristic for success, but I'll tell you what the gotcha is, is how fast is the founding team iterating and moving in the early days. If I look back on my most successful investments, it's insane how fast like every investor update was so meaningful and juicy and I was getting all this information about the external world. In fact, even if there wasn't a return on investment sometimes, just what you can learn through the right team that is out there exploring new marketing, new technology, and new technique is just so valuable that I optimize for learning at the end of the day. And so what is not clear though about this heuristic is that are the teams that are moving so fast the ones that have already kind of gotten into that product-market fit train, right? And therefore the market is like this dance as an entrepreneur. You're not just doing all of the dancing yourself. You need a dance partner that you carry but then carries you. And in that sense, you kind of start getting that traction and there's so much going on because these companies are starting to get customers and customers inform roadmap and customers inspire innovation. So much of what people consider a good idea, it's just customers gave you the good idea. Correct. So I'll give you an example. One of our most successful technologies, ISR, is this technology that sort of blends the benefits of a static website and static generation and the benefits of dynamic rendering at scale for the largest websites and web applications on the planet. And that idea, I actually give credit to HashiCorp. They were a customer of Vercel. They were moving off of this sort of more static platform and they knew that we were doing a ton of stuff with server rendering and moving compute to the cloud other than the device. And then the Hashi team was like, oh look, but I want some of the benefits that I had in that other platform. I want you to innovate and solve this problem for me. And so we created that technology. But probably in the annals of Vercel and the web industry, it would be easy to mistake that as, oh yeah, the founding team at Vercel were totally cracked and whatever. Instead, what really happened is market provides the best insights, information, the most demanding customers, the ones you want to partner with. And it goes back to why partner with startups, right? And so yeah, going back to those early signals, I think teams that are moving really, really fast in the early days. I think another heuristic would be transparency. So teams that really come to you with good problems and good facts come to mind. But yeah, there's a ton. And the thing I always share about startups is that I've just been in the game enough that there's all kinds of outliers that break all the rules and still succeed. So don't get too attached to the rules.

Yeah, I think number one is cultural. So if you build a culture where customer is king and customer is never wrong, I actually frequently you'll find myself repeating, if you were in our Vercel Slack, I joke that I've never met a customer who's wrong. I just have not. Maybe they're wrong in maybe they try to solution for you, right? Or maybe they troll you, but there's always some kernel of truth in what basically every customer that I've talked to has shared with us, even if it's not the right channel, presentation, like even if it's not the most optimal message, whatever. And so it starts with that culture. And I do think also customer prioritization is quite important, right? So when companies like this that we knew would be the most demanding ones or had a lot on the line, had a lot to lose when partnering with us, when they bring you an insight, obviously it holds a lot of weight, it holds a lot of value. So again, we had the cultural engine to respond very quickly to that insight. Whereas I think, and even myself many years ago, I could have been like, I don't know, maybe this customer just doesn't get it, they're not a good customer, they're wrong, whatever. I'm supposed to dig in deep, which is actually one of our values, dig in deep and getting to that fundamental problem that they were trying to solve, right? It's very sort of a defensive mechanism to essentially say when a customer doesn't like something or is complaining in the wrong way, that they have a hidden motivation, that they are basically bad. One million per one million per, you're always trying to protect your ego. Like your startup is your baby, right? And which comes with a lot of bonuses. It has a lot of pluses that you as a founder become very inextricably linked in your identity to your product, your creation. And I think actually now that you're prompting this, I do think we early on created a culture of this attachment with the product. It's actually kind of strange. Obviously we're obsessed with the product and at the same time we try really hard to be our harshest critics. And some of the things we say about the product sound like, I don't know, it's just like, oh, you're tearing it apart. It's like a Shark Tank but for products, you know what I mean? And I think ultimately that's the right stance to have. Like don't get high on your own supply, don't drink your own Kool-Aid when it comes to product. And the customer who's actually worried about something else, if you're a vendor, if you're a technology provider, you're just helping their business be better. You're not their core thing. You're just somebody else, right? And so customers have the least patience and the least ego attachment to your product. And so their perception of reality is much more accurate than my own team's perception of reality.

Great question. One of the things that every SaaS company, every infrastructure company should be thinking about is you're trying really, really hard to build just one software product, right? When you go down the path of customizing software on a per-customer basis, again, there's no hard rules in this game, but generally speaking, you want the economies of scale of running, updating, patching, and evolving one piece of software that multiple people buy. And so when Hashi had that problem, and then I have another story about TripAdvisor which led to another innovation. When Hashi had that problem, my nose immediately told me, whoa, if we solve this for them, right, this very same thing we're building for them, everybody in the world will benefit from. And it turned out to be true. The same thing happened with TripAdvisor where they loved, this is actually so funny because it's like a chain of innovations. They had ISR and they were telling us that's super cool. So they were using it for, we run CruiseCritic.com and you can go and check out your cruises and whatnot. And they're like, awesome, you've now nailed the speed of, we have lots and lots and lots of cruise ships that you can buy and packages and inventory and whatnot and our team can push content, update those over time and stock changes over time. Vercel is updating those pages. But what about when I want to run an experiment? I want to run two versions of a certain product page and I want to determine that perhaps on a per-country basis. So we need to run compute close to the visitor as well as compute close to the content database. Oh, interesting. And so that's where our investments into edge compute began. And that conversation with TripAdvisor led directly to the creation of edge middleware. I see. Which now, because what I was really captivated by when they came to us with that requirement was every customer in the world seems to have a slightly different preference on how they want to customize the workload. So some people were saying I can't do business in a certain country due to OFAC regulations, so please at the edge let me do some configuration over here so that instead of rendering my current website, I can say sorry we can't operate in your country. Maybe another customer like luxury brands that we host were saying China is a huge emerging market for us, help us deal with that, which is a completely different sort of inventory product approach. And so I kept hearing all these requirements. I was like, if I implement every single one of those requirements as a configuration flag, I'm going to die a death by a thousand cuts, right? I'm going to have a PM each responsible for like a little product or flag or feature configuration. Instead, we gave the customer the power of code so they can run any compute that they want and customize their platform to their liking. And interestingly enough, this is a lesson that pattern matched to my very first, one of my very first angel investments, which was Auth0. They had found a similar problem. Customers were coming to them with all kinds of endless customization needs for authorization rules and they could have created a rule engine with a very complex UI and lots of check boxes and whatever. And I remember the CEO telling me our competitor tried that. Instead, Auth0 gave people the power to put any code in there, right? So it goes back to why it's so interesting to talk to a lot of people. Maybe some idea sounds irrelevant to your business at the time but then it finds an application. So patterns like that are really useful to learn and it came really handy for us because with edge middleware we're able to sort of evolve the business to the next level and bring in a lot more workloads and enterprise complexity that prior to that we couldn't really handle.

It might have actually been even earlier. The very first thing that I did on the internet that made me money was I would make websites that drove traffic in things that I was interested in, like no joke, like Dragon Ball Z. And I would put banners to MercadoLibre, which today trades on NASDAQ and it's like the eBay and Amazon of Latin America. And when they were getting started and they themselves were a startup, they launched a referral program. And so I was part of that referral program and I knew how to create websites and I knew how to drive traffic to the websites. And then people would click on that banner, sign up for MercadoLibre, and they would, I think it was five pesos per sign up. And so with that program working, I actually made like, I think everything is so blurry at this point, but I made a good amount of money, like hundreds of pesos. Wow. Enough that my mom was shocked. Enough that even building up to that figure, my brother and I, my mom, we all comment on this, like it was the big topic in the household and we all believed it was fake, like it couldn't be true, right? I think the whole thing was bizarre, right? Like the internet is so early in Argentina at this point. Kids that make websites plus referral programs of this again rising star of the internet, but nobody used it at the time. It was so new. So yeah, that was the very, very, very earliest that I remember.

Yeah, so at that point I think it was mostly the novelty. And I was so early and truly so young that I don't think I could even rationalize that well what was going on. But the very next instance of it was I was spending a lot of time teaching people online about Linux, about system administration, about programming, and about like, I just had a passion for sharing and disseminating knowledge. And someone on an internet forum commented and said, you realize you could make money of all that knowledge that you have, that you're just writing tutorials all day long. And I was still, of course, living with my parents at this point. I was like 12, 13, 14 years old when I began my freelancing career. And I think, you know, it's interesting, we're talking about feedback in the market and product-market fit. What I really loved about the first forays I had into freelancing as a kid was that I wouldn't just get the money for the job that I would do, I would get a rating. So this website ScriptLance, I had my stars, like five stars reviews page, right? And I just loved getting good feedback of like I did a great job. And I think that in many ways was just as impactful to me as the money that I was making that was helping my family and I was buying stuff for myself and whatever. So the market provides this awesome feedback loop of like, oh I don't like that, yes I like that, oh I'm churning, your product is no good, oh you did an excellent job, I'm expanding, I'm coming back. And so it completed my journey in many ways. I think I was missing that other element of confronting the real world. And funny enough, during Q&A with the whole company today, I was sharing this insight that I think as engineers become more savvy with the help of AI, they become more complete in their skill set. For many engineers, the next frontier is becoming full stack, not just product engineers. And again, platforms like Vercel are helping you not have to provision so much of this infrastructure and whatever, so you can just create entire products that could reach billions of people with just product engineering skill sets, right? But what about the business skill sets, right? I believe that if you're engineering in a vacuum and you're not thinking about the business, then you're also not going to create the best product. So part of this full stack product engineering vision of the future is an engineer that is thinking a lot about growth metrics, that is thinking about comps, that is thinking about promoting the product. Even if you don't do that, right, you have to understand it's not something that you just simply delegate. Oh, that's the marketing team's problem, that's the sales team's problem, because otherwise you're not getting the right inputs into the thing that you're creating. So I think the product engineers of the future are full stack marketing, sales, comps, as well as product design and engineering.