Introduction

Patrick: [00:02:06] My guest this week is Anu Hariharan. Anu is a partner at Y-Combinator’s Continuity Fund where she focuses on growth investing. Before YC, Anu was an Investment Partner at Andreesen Horowitz where she worked with portfolio companies Airbnb, Instacart, Medium and Udacity. In this conversation, we discuss growth stage businesses and their business models, how her background as an engineer impacts her investing style, the most interesting international markets for tech start-ups, and how much opportunity there still is for investing in tech and e-commerce startups. This conversation left me thinking about how much digital transformation there still is in front us and the exciting opportunities ahead. Enjoy this great conversation with Anu Hariharan.

Investing Based on Business Model vs Industry

Patrick: [00:02:47] So, Anu, thank you so much for doing this with me. I thought an interesting place to begin would be with your framework for how you think about growth stage businesses through their business model rather than through their industry, and why that lens is appropriate for investing.

Anu: [00:03:00] First of all, thank you so much for having me, Patrick. I'm a huge fan of the podcast. Specifically on business models, I in fact gave a presentation in YC for the YC batch a couple of years ago. We focus primarily on business models, because when you break down tech startups, even by industry, certain business models really overlap. So, let's start with what they are. For example, if you're looking at marketplaces, whether you're a B2B marketplace or a consumer marketplace, at the highest level, you have gross merchandise volume, the dollar value of the transaction that's flowing through; you have a take rate that translates to net revenue. Then, you have a bunch of costs associated with servicing that revenue, especially all variable costs, and then you deduct that to get the contribution margin and then you have EBITDA margin.

The reason we focus on business models by function versus industry is that, as tech is pervasive across industries and software is eating the world, you do see that whether it's a B2B or a B2C, if your business model is a marketplace, your metrics are similar. If your business model is advertising (like LinkedIn has an advertising angle, but it's arguably on the B2B side and Facebook is on the consumer side of advertising), advertising as a business model, you monetize in a similar way.

When we broke it down, I think there are about 9 to 10 business models that really are the most common across tech businesses. At growth stage investing, especially at the stage we invest in with the B or the C, we're sort of testing on: okay, if you have product market fit, which is to say you have, if you're a consumer business, you hopefully have like a million plus users, or at least a few hundred thousands. But what truly is your business model at scale? What's going to drive your revenue? What are all the costs associated with the revenue? And what is the scale efficiency or the leverage that you're going to get as you scale? And is that underlying business model solid? So that's what we are looking for in a business model.

Winners Don't Necessarily Take All

Patrick: [00:04:59] You mentioned something before we hit record that I found totally fascinating, especially at the stage that you're starting to look at companies at this sort of B or C kind of growth stage. Which I take to mean, they've got a product that's working, there's less product risk than you might take earlier on. You mentioned one area specifically which is food delivery and DoorDash, where I know you've been involved. But you've made a point, which is interesting, which is the consensus is all of these markets are winner-take-all or most, and that there's effectively a pie to go win. And if you race to scale, which most growth companies are trying to do, you can have this outsized outcome. And I think you have a slightly different perspective on this winner-take-all or most idea, and I'd love you to explore that for us.

Anu: [00:05:39] So especially in Internet eCommerce businesses, and I think there was this very strong prevalent trend that most businesses would be winner-take-most. Let's talk about food delivery in particular. If you looked at food delivery a decade ago, in the most narrow sense, you'd think of it as how many people are ordering for delivery today and is that transitioning online? That may have been a small market, maybe $10 to $15 billion. But, if you look at the takeout market, it was a $70 billion market. Most of the takeout market was literally someone placing an order on the phone, either you go pick up, and that was the takeout market.

And then there was this whole restaurant market, which was hundreds of billions of dollars with the majority of restaurants not doing any delivery. DoorDash. DoorDash is a very fascinating company. It's a 2012 YC company. And, one of the things that's very interesting being inside YC is, we have this unique preview into knowing what the company and the founders' insights were from day one, because they have to apply. If you literally read DoorDash's application from 2012, they were actually called PaloAltoDelivery.com. I think they still own the domain. If you actually type the domain, it takes you to DoorDash. And that's because all four founders were from Stanford. And their insight was, "Hey, I cannot order any food from any restaurant in Palo Alto. No one delivers."

Well, there was so much talk about Grubhub. But Grubhub and Seamless, even after their merger, more than 60% of their business came from two cities, Chicago and New York. They literally built this delivery business in Palo Alto, because they were students at Stanford and they wanted the restaurants in Palo Alto to deliver food to them whenever they ordered. And their insight was two things. They said most people are going to chase cities because they think that's where people eat out. But actually, customers in suburban areas really want delivery and access to delivery but no one's doing it. So that was post insight which was, "Start with the suburbs because maybe their retention will be better because there's not many people focusing on the suburbs."

The second thing related they found out was a lot of the delivery drivers that were driving for restaurants in these suburbs didn't have much work. Because the delivery volume was lower, very few restaurants did it. And so they realized that if we could recruit those delivery drivers and increase their average pay per hour because we could give them more deliveries or more hours, we'll be able to retain and acquire drivers at a better rate than Uber could. And those two insights alone is really what helped DoorDash differentiate and scale.

So in 2016, I remember this very vividly because that's when Uber Eats launched. And Uber is such a big ride sharing business, and I think most investors thought that food delivery is a zero-sum game, and Uber Eats is going to be the market leader in the US. Fast forward four years, today undoubtedly, DoorDash is the market leader and there's still room for two or three players.

And so what was different? One, the insight they had. They started from the suburbs, the order values in suburbs are higher. It's actually they're way more cost-efficient to acquire that driver, like 10 times better than what a ride-sharing company does in terms of acquisition of driver, and you see that in their retention cohorts too. And then the other thing they focused on was merchant strategy. I think from very early on Tony's insight was, selection is the most important. 70% of customers on DoorDash really stay on DoorDash for selection. They not only acquired a wide variety of merchants, but also provided them with workflow integration tools, in terms of when you get orders, how do we queue up everybody? The ones that you get via DoorDash versus other tools or through phone calls? How do we help you with accounting? How do we help you with invoice processing?

So that stickiness of the integration workflow plus the core merchant focus, really helped them spike on selection. And the approach from suburban and then to cities really helped them tip. Fast forward to today, I don't think there is any doubt that there's room for two or three players. But I think having that insight that DoorDash did really helped. So you can see that there's room for multiple market players. It doesn't mean you can have 10, because there's still something called scale effects that you get, and you have a fixed operating cost base to support significant scale, but there's definitely room for at least two to four per market.

Patrick: [00:09:54] Can you say a bit more about maybe this -- stick with the DoorDash example -- I'm always interested when there's sort of a marketplace-like dynamic. You're effectively connecting supply and demand. How do you decide as you build out the business who to focus on? I find that of course customers matter here and of course the restaurants matter here. Any insights from watching a lot of marketplaces and maybe even DoorDash specifically, about how they choose to prioritize their different stakeholders?

Anu: [00:10:19] I did this piece of work while at Andreessen Horowitz on network effects. I remember I learned this from Chris Dixon. He always said, "In a marketplace, you got to figure out which is the hard side." And at different points of your company's evolution, one side is easy and another side might be hard. But it changes, the side might actually change. So the simplest way to explain this, the way I've thought about it is, I lived in New York for five years before coming to the Bay Area. If you look at New York bars, the dating scene was incredibly hard in New York. And some of the bars used to incentivize women at happy hour. There used to be like discounts for women to come to the bar. And that's because they said it was really easy to get the men into the bar, but it was really hard to get the women into the bar.

So at different points of your marketplace evolution, you need to figure out which side is hard, and do you need to subsidize? Let's look at Airbnb. You have the host and the guest. Initially, the guests were actually easier to get, because when they targeted events... So their famous story is the Rhode Island design conference, or there was a design conference going on in San Francisco and all the hotels were sorted out so, of course, some on Craigslist were looking for a place to say. But, it was really daunting for a host to put up their home to rent. So at that time, the subsidy was on the host side. In fact, they were incurring acquisition costs only on the host side. They were acquiring demand completely organically. They spent zero dollars for a long time to acquire demand.

And then at some point, it switches when your host, if you have a significant number of hosts in a certain city that's making a good revenue stream out of the business, you may have locked in supply, and now maybe the harder side for the next layer of scale. What I mean by that is say if you're going from 1 million to 4 million customers, maybe your demand side becomes harder. Then you've got to start offering subsidies on the demand side. But at every stage of the marketplace, you have to always ask yourself, "Which is the harder side today?"

Diligence in the Growth Stage

Patrick: [00:12:19] As you've explored the growth stage specifically, what are the key elements of diligence that happened for those companies that didn't happen or maybe were less important at say the Series A or seed earlier on in the business's life?

Anu: [00:12:33] I think the highest-level three criteria is pretty much the same, which is number one, team, number two, market opportunity for that business. And number three is performance to date. This is where I think a Series A investor would say “early signs of product market fit” because the company has been around for maybe 15 to 24 months, which is by the time we are investing, the company is usually three to four years old. And, so we call it, growth investors call it performance to date, which is “how have you really executed in the last four years, both from a top line as well as from a cost perspective?”

Let me touch on each. Number one, team. Team, a Series A investor would probably say, "What is the founder market unique insight that you have?" I use the analogy of the YC application on the early stage. It's literally they have nothing. They usually never launched. I mean, more than 60% of our companies haven't launched on the early stage. But what we are looking for is what unique insight you as founders have about the problem you're solving, and why you think that's a big opportunity. So that's what we're really trying to gather from the team, and second, the speed with which they move. From the time they had an idea, how fast have they been executing? Do they actually do an MVP in two months, or is it taking them six months to do an MVP, and why? That's what you're looking at the Series A.

In growth, what we are looking for from the team is clarity of thought--and I'll give an example for each--clarity of thought, your ability to learn and scale, and third is your grit and determination. So clarity of thought. B or the C level valuation is roughly like $200, $400 million. And so at that point, it's not necessarily the investor is looking for a 10X, but they really want to know, what is your clarity of thought on whether this company is at least a $3 billion or a $5 billion or a $10 billion company? I usually refer to the Brex founders for this example. The Brex founders went through YC in Winter '17, and second or third week into the batch, they had this idea of Brex, which is corporate credit card for startups.

But they took an accounting class in Stanford. They were actually Stanford freshmen. And I asked them, "Why are you doing this accounting class?" Henrique and Pedro, this was their second company, believe it or not. They had built Stripe for Brazil and sold it when they were 19. And they said, "Well, we just want to really build a case for why will Brex be a billion-dollar company at least. And we want to build a model that thoroughly reflects market share, how many customers we need to get, whether there's appetite for so many customers, what's the unit economics. So, we just wanted to take the accounting class to really make sure that whatever case we built was relevant." And so I said, "Okay, but what's the point?" And they were like, "Well, but the stakes are high. We're going to quit Stanford if we're going to build this. And so we just want to know, is this worth? Is this a big company? Is there a big company to be built here?" And that clarity of thought is rare. It honestly comes once in a decade.

And so I said, "Okay, but what's the point?" And they were like, "We're going to quit Stanford if we're going to build this. And so we just want to know is there a big company to be built here?"

That clarity of thought is rare. It honestly comes once in a decade.

To this day, maybe... Of course, they've deviated from plans and they've far exceeded plans in more cases, but like that product roadmap, even though they were starting with corporate credit for startups, and all the suite of products they thought through that they would build, is still pretty much spot on. That's rare. That's something we're looking for from founders. "Do you have the clarity of thought on how big this business can get?" The second example is your ability to learn and scale. Here, your first phase of the CEO is you're Do-It-Yourself CEO, which is you probably have a team of five to 10 people, it's all about getting to product market fit. It's all about tweaking that product, really figuring out “are there 1,000 customers that love you? Can you get that to 10,000?” And so on.

But the moment you start hitting 30 people, you now have to become the Company Builder and Chief. Most CEOs are, the median age of a YC CEO is like 27 or 29. So for most of them, they have not managed more than 20 or 25 people. And for some of them, it's their first job, so they don't have people management skills. Their ability to learn and scale as a leader is something I think YC, we have an unfair advantage there because we've known them for four years. And Faire, the company Faire, the CEO, Max Rhodes is an example I'll give you. Look at his resume. Tremendous background. He was the GM of Caviar, launched Square Cash. Faire is like a marketplace that connects makers and retailers, so of course he really knows the business building, but he's never managed many people.

And he got this feedback very early on in his company evolution that he openly admits, his people were terrified of him. He realized that he wouldn't be able to successfully build a company if he doesn't improve on that lens. And so he went and hired an exec coach. He constantly sought feedback from everyone around the table, from his own leadership team. He's a phenomenal leader and scaling really well as a leader. People often look to Patrick and John [of Stripe], another YC company and saying they were one of the best founders that have built one of the best exec teams in the Valley, and honestly the entire company is great. I think Faire is the next Stripe in the caliber of people that they're going to build. That's the second thing we're testing.

The third is grit and determination, which is almost a given but so hard to test. But once you're in existence for at least four years, there's surely been at least two or three near death experiences in your company. And have you survived that? DoorDash has gone to hell and back, and I would never bet against Tony. I'd say the same about Rappi. It's a Latin American company which hasn't stuck up as DoorDash. I mean, odds were against them. They don't even have a strong venture capital market in LATAM, and here they are building a capital-intensive company. But my God, their grit and determination is unparalleled. And so you don't want to bet against those founders. So those are really the three things that we are looking at from a team perspective, which is more proof points and a different lens from what a venture investor looks at when you look at a growth stage company.

The Non-US Venture Investing Landscape

Patrick: [00:18:24] You mentioned LATAM, and it makes me think of non-US markets for these kinds of global technology companies. What other markets do you think are the most interesting maybe from a geographic standpoint right now?

Anu: [00:18:37] The world is increasingly becoming global. I mean, YC started a decade ago, and pretty much all the startups are US-based. Today, out of the 2,500 plus startups we funded, 27% of them have headquartered outside the US. More recently, in our batches at least 40% of our startups are international. I think if you play this a decade out, I think at least 50% of YC startups are going to be headquartered outside the US. Innovation is just happening everywhere. Specifically, markets I'm excited about: there's a lot being written about China. China is probably further ahead than other markets, but emerging markets I'd say, very bullish on India, Indonesia and Latin America.

What's going on in these three? India, I think the GDP per capita has always been really low, like in the order of $2,000 to $2,500, which is why the willingness to pay was not as high, and so you didn't see as many $30, $50 billion companies coming out of India. But that's going to change. This decade, 2020 to 2030, the GDP per capita is expected to go at least to $5,500 which is what China had more than a decade before. And that's what really propelled a lot more multi-unicorn companies that were built in the region. And India has the talent. I mean, if you measure the number of computer science grads, India produces as many or more than the United States. It's soon supposed to surpass that. Therefore, I do think it has the talent and it has the venture capital ecosystem. The GDP per capita is the one evolution we need, but it's happening this decade. So I'm very bullish there.

Indonesia, GDP per capita is already really high, the population is far lower, but you're seeing a lot more Southeast Asian founders, even founders from other regions moving to Indonesia to build businesses. And there is a huge investor ecosystem like some of our companies, Xendit, which is a YC company, which is the Stripe for Indonesia. It's really pioneering payments in Indonesia. You see a lot more new businesses starting in Indonesia, especially in the FinTech space. So that's another region.

Third is Latin America. What's fascinating about Latin America, and I learned there's something from Rappi. Actually working closely with Rappi. Latin America has 650 million people, 70% Internet penetration, GDP per capita same as China, $9,700. In fact, countries like Chile, Argentina, Uruguay it's closer to $12,000 to $15,000. And 200 cities with high people density. What's the eCommerce penetration? 4%. And the eCommerce penetration of China is 22%, and maybe because of COVID, it's probably higher. But pre-COVID it was 22% and pre-COVID in LATAM was 4%. So you just do the math, and I just think Latin America is going to be huge.

Latin America has 650 million people, 70% Internet penetration, GDP per capita same as China, $9,700. In fact, countries like Chile, Argentina, Uruguay it's closer to $12,000 to $15,000. And 200 cities with high people density. What's the eCommerce penetration? 4%.

Its geographic proximity to North America is going to help with the investor base and talent. It's really hard to build a company in India and find execs that have scale experience, but LATAM can at least get the people from the US to move to LATAM. I do think LATAM is going to see a huge wave of multi-billion dollar businesses, and probably a region that doesn't get as much attention as India or other parts of Southeast Asia.

Patrick: [00:21:46] What are both the reasons behind that and the risks associated with that sort of early investing? I mean, 4% is tiny. It seems like it'd be hard to not squint your eyes and look forward say 20 years. It'd be impossible to imagine that digital and Internet and ecommerce isn’t way more penetrated in Latin America on that sort of timeline. The opportunity seems obvious, but what's the catch? What are the risks and the difficulties and the reason it's gone slower?

Anu: [00:22:12] Look, it is fragmented across multiple countries. It's not one country. And so the regulatory environment is actually quite complex for a young startup starting out of one country. For example, if you look at the case of Rappi, they work with hundreds of thousands of delivery drivers and personal shoppers, but they have to navigate the Colombia landscape and the Brazil landscape, and Mexico - they're present in nine countries - and Argentina and Chile. The complexity really starts increasing.

[The catch in Latin America] is that it's fragmented across multiple countries. It's not one country. So, the regulatory environment is actually quite complex for a young startup starting out of one country. For example, if you look at the case of Rappi, they work with hundreds of thousands of delivery drivers and personal shoppers, but they have to navigate the Colombia landscape and the Brazil landscape, and Mexico - they're present in nine countries - and Argentina and Chile. The complexity really starts increasing.

Second complexity is in company building. They have probably more than 2,000 employees but that are spread across [LatAm] . There is this whole mantra to start in San Francisco, keep most of your employees in San Francisco. And you can also find all the other scale. Like that's simply not true there. And then you're also navigating with different time zones. But all of those are changing, and also investor interest before Rappi was very little. Rappi went through... I mean, we're talking about Rappi as though it's an established company—it's four years old.

So we went through YC in summer of 2016. I was actually at Andreessen Horowitz at the time, and we wrote the first seed check in Rappi. And it was fascinating because Andreessen Horowitz at the time, did not do any investments outside the US. Literally, our first investment was Rappi. I think that's because we could see through the vision of the founders' eyes how big that opportunity was, and we were already investors in Instacart at the time. People knew a lot about DoorDash, and there was this company that's building essentially the Instacart plus DoorDash of Latin America.

I think that there's more investor interest now, given there's a lot more startups in the ecosystem, and Rappi has sort of paved the way for generating that amount of investor interest. The second thing, people are realizing that in spite of that regulatory complexity and challenges, the market is very attractive. And it's not just from startups, look at Uber. I mean, for Uber and ride sharing, Brazil and Mexico are some of the most important markets outside the developed markets. And there's a lot of money, right? If the GDP per capita is really high, it's not that the order values are low.

If you look at Rappi's order value, that's something that's counterintuitive. It's close to $20. And that's not the case in China or India. India is like far less than $20, maybe $5. And US, the food delivery players blended will be around $30, $35. So $20 is not that far from $30 or $35. But the differences in cost of delivery, the cost of delivery is less than 10% of the order value in LATAM. Whereas in the US, 50% of your gross fees goes on delivery. It's unfortunate, but the socioeconomic status difference in LATAM, structurally helps support these companies. Because you have higher order values, at the same time your cost of delivery can be less than $1 per order. And the minimum wage per hour is also a lot less than the US.

Hard Tech vs Soft Tech

Patrick: [00:24:54] One of the things that was really interesting to see in companies that you've been involved with is Boom Supersonic. I bring it up because I don't even know if hardware is the right term. Physical technology, maybe. It's such an interesting looking company. How do you assess a company like that relative to a more traditional sort of software, heavy technology portfolio given the, I would assume, extreme capital intensity and extremely long lead times? What's different about that company and why do you find it interesting?

Anu: [00:25:23] Often at YC, I often say like, it goes back to the prior insight that I just shared with you, which is I don't think any other investor firm has that, which is we have this unique insight into how the founders are scaling and how big their vision or mission is. Blake is that kind of a founder. I think, Blake, even if you talked to him five years ago, he was absolutely laser-focused on building a supersonic jet. If you look at his background, nothing screams of that. He worked at Amazon, and then was at Groupon, and then back to Amazon. How do you think of an Amazon, Groupon person building a supersonic jet?

But then when you spend time with him, you see how much time he has spent outside of his core work learning about planes, and that's where he just puts up and he knows to fly, and he knows about all the jets, and he can tell you what the nuances of each jets are, what's the difference. His whole mission was, "Look, it's easy for me to build a software company. Everybody can, and there's lots of people starting. But people are not doing the hard things. Why is it that aviation has seen no innovation in decades? Not just one decade, in two or three decades. I mean, we already had the Concorde for crying out loud, and we didn't have any... They shut it down, and then even though there has been advancement in manufacturing, advancement in laser technology, advancement in supersonic, no one's investing in it."

The way from an investment perspective, we really look at these, what we call the hard tech bets. It's quite simple. We actually use a mental model. Which is you look at science risk, engineering risk, market/commercialization risk, funding risk. That's it. You break them down into four buckets.

The science risk for Boom was very low. The supersonic jet was already built. And so that is proven science, it flew. What most people don't realize, the reason they had to shut down Concorde was it was not economically viable the way they built it. And the cost per seat and the number of people in the plane didn't really support the cost to operate that plane.

That insight that Blake had from day one was very powerful. Science was we were like, "It's slow. Okay, we get it. You can build a better jet with a new technology and maybe even a better max speed over time, but it's largely proven and there was a plane that launched." Engineering risk is the bet we are taking, which is you have all this advancement but the advancement really matters only in the speed and pace with which you execute, and the less errors you can make.

So one of the biggest milestones is actually the assembly of the jet itself. You run all these software simulations to figure out is this how I need to design the part? What should be the specs of the part? But until you actually put it together, you don't know if those things work and like, okay, how fast can you iterate on the simulation and can you do the course correction so that your manufacturing is not delayed significantly? And that engineering risk, we were willing to take a bet on the Boom team, because that's their core strength. That's Blake's strength. His strength is in software and his ability to attract really strong software folks to help build that. So that's the engineering risk.

Market or commercialization risk, like I remember when he was going through the batch, one of our group partners told him, "Go prove to me. I know you are not going to build this, but go prove to me that you can get Richard Branson excited about this." And he did it. He didn't know Richard Branson at all. He just went, hunted him down, pitched Boom, and Richard was one of the first angel investors in Boom. Over the course of two years, he had more than a dozen, two dozen contracts from commercial airlines. I mean, Japan Airlines is an investor in Boom. But they all have contracts in place where they have an option to buy the first 10 jets when they produce.

We were like, "Okay. He's proven market and commercial risk, and funding risk." So when you look at funding risk for hard tech, you have to think different. We're not assuming that this is venture capital that's going to fund them forever and see them through. But if you actually spend a lot of time in aerospace, you realize that the pools of capital that they can tap into are plenty. You can start off with venture funds maybe, but then there are lots of suppliers that actually finance this program. For example for engine program, it's a known thing there, like GE and others who manufacture engines actually can finance up to $500 million engine program for any model that you're working on.

And then you also have these huge raw material suppliers who finance the manufacturing of the plane. And so when we did the math, we said, "Okay, even if they need it, they probably need a billion dollars to build the Overture which they call the commercial jet. Okay, what are the different pools of capital that they can tap into?" When we did all the risk assessment, it was literally market risk is low. If this thing flies, there will be enough appetite from airlines and consumers. Imagine if you could fly, do a round-the-world trip in a day. You don't need five seats. You could fit more people in the plane and make it economical.

And then there is funding risk is medium, but that depends on execution. There's lots of pools of capital. And engineering risk is high. Engineering risk and funding risk also falls on the founder, and that's where our conviction on the team comes. So that was our mental model on how we decided to lean in on Boom.

Anu's Background and the Importance of Upside

Patrick: [00:30:23] I know your background originally was in engineering, working for Qualcomm, that's what you studied. We were talking about that before we hit record. I would love to hear how that background plays into your style as an investor. Typically, engineers are unsatisfied until they understand deeply how something works. I would just be curious to know if that translates over. And if so, what are some anecdotes or examples of how that happens?

Anu: [00:30:47] If you asked me a decade ago, I don't think I would have ever even thought of being an investor. Not that I was never fascinated by it, but just didn't know much about it. At Qualcomm, I was tasked with deploying our 3G handsets with their video, telephony and video streaming solutions. This was pre-iPhone, so there was nothing called an app ecosystem at the time, all over Europe. Over my first two years, people really knew me as an awesome debugger. That really was my title, even though my official title was senior software engineer. But there was a joke in the company that if there was a bug anywhere in any part of the world, like deploy Anu, because she's going to go find out what's the root cause.

And that's just because I actually thrive when no one knows the answer to a problem, and it just keeps me up. It's just something that I've always liked and it's somewhat nerdy. They used to send me to Italy a lot at the time. I mean, we were deploying 3G all over with all the operators, and there was this weird bug at one of the cell zones during transition, and I had to like keep driving with my form factors to A, replicate the bug and then debug it, and then fix the code. It was often done in a time pressure setting. So fast forward, I often say that process of finding root cause is really what excites me about a diligence process.

And so when I was at BCG, post business school, I actually spent five years in the private equity practice. The private equity practice was doing due diligence for several private equity firms across several industries. My first case was in mortgages, and it was literally to figure out a mortgage book and whether this PE firm needs to acquire that. Two, I also looked at evaluating a diamond mine for one of the private equity funds. And so I knew nothing about diamond mining, for example. I was like, "Okay. How do you replicate this process? We need to figure out whether we need to buy this mine. We need to just understand how the price of diamond evolves and what's the role this diamond mine will play in the supply scale and evolution because that influences."

To build the supply curve of diamonds, I mean believe me it's so hard. You realize that it's such an opaque industry. It's an oligopoly between De Beers and Alrosa, but there's this long tail of diamond mines, maybe like thousand of them, and you have to estimate their supply production capabilities. And the only way to do that is like we were doing calls to Botswana. We even had somebody from BCG Africa to visit the mines and talk to some of the workers that really understand what the supply production is. We had to come up with this probabilistic curve.

The way I used the diligence skill or the debugging skill, if you will, was literally, in private equity side, it was that diligence root cause process, because private equity funds are all about downside risk protection. It's like, "Okay. I know what I can do with this mine but tell me what my downside risk is." And so we would really unearth to the extent possible, like what's the true curve and what's the true impact from a downside that you will see as a firm?

And then fast forward to Andreessen Horowitz, which is such a different nature of investing. I feel like you can be a great tech investor only if you're an optimist, not if you are a pessimist. And one of the reasons why I feel like private equity worlds and venture worlds never overlap for a reason. When I was at Andreessen Horowitz, I was hired for the diligence skill set, by the way. They loved my diligence experience and the fact that I could go deep on any sector pretty much, and figure out how to sort of build an investment case around it.

I feel like you can be a great tech investor only if you're an optimist, not if you are a pessimist. And one of the reasons why I feel like private equity worlds and venture worlds never overlap for a reason.

But two months into my job, Marc Andreessen told me this, he said, "Your bias is going to be to say no, because you use your process on private equity to figure out downside risk." He said, "Here's my challenge to you. You use the same process to figure out the upside potential." And the reason is he said, "Look, as venture investors, you will lose money, but you lose only 1X the money. If you turn down a 10X investment, you lost an opportunity in a company that could potentially return the fund. And that's grand slam." And he was like, "You have got to reduce the number of grand slam misses we have." It's a very simple insight in hindsight, but that just changed my way of thinking.

So now in growth stage diligence like when I work with my team, I always tell them, "Look, don't spend 30%, 60% of your time writing the bear case. Spend 70% of your time telling me what's your 20% upside probability case, because that's what most investors are not going to get." The example I use here is Stripe. I mean, unbelievable founding team. What they've accomplished is stellar. Patrick and John have done a phenomenal job of building the team and the payments processing business, which if you look at it, you can say easily as a Series A, Series B, investor like, "What's your moat here? You don't have any network effects. You're just acquiring all these companies, one by one."

But if you talk to Patrick and John, you will see their crystal-clear clarity of thought on product roadmaps, the suite of products they plan. So when it came to upside potential, I remember this very vividly three or four years ago, a lot of investors underwrote Stripe to like maybe a $20 billion company because they could see it as a core payments processing company. But if you looked at their product roadmaps and even believed that let's say two of their products worked with a 5% attach rate, what would that do to their upside case? Like what happens if Stripe did things right? You could start seeing them being a $50, $100 billion company. I think as growth investors, that should be your secret sauce. It's very easy to model a base case and a downside case, but where you're going to get wrong is the upside case. And you turn down investments because you estimated the upside very low.

Patrick: [00:36:09] That's fascinating. One of the descriptions I've seen in one of your various bios was this interest in network effects, which we've talked about, and also in what you called growth programs. I think that piggybacks off that last question quite cleanly, which is in the upside case, obviously there's going to be tremendous growth. And what I've learned, certainly watching tons of companies and building one is that growth is a very careful, painstaking process. It can be incredibly expensive, it can be very difficult, but properly thought through and set up, if a company gets the right flywheel going, it can be super powerful. Say a bit about everything you've learned watching these companies and the different kinds of growth strategies they've employed.

Anu: [00:36:47] I think this is something that honestly YC is really good at. People often say, "Why are you so focused on growth?" But for a startup, growth is really an indicator of everything that's going on, especially in the early stages. Is there demand for your product? Growth is the only way to signal that. But it's important to make sure that this is organic growth, not paid growth. So that's the most important difference.

We have stats that are like, "Oh, if you're a nearly startup, like is your week-on-week growth at least 10%?" And that's because it's going to start compounding a lot. Compounding is such a great factor, but if you're able to do that organically, that means you've built something that people really want. So at the early stages, it's really organic growth. But even at the growth stage, we are looking for, "Hey, is more than 50% of your growth really organic?" Often you'll see that the difference between the startups that really scale very well and the ones that struggle to scale is that organic-driven growth.

Often you'll see that the difference between the startups that really scale very well and the ones that struggle to scale is that organic-driven growth.

I'll give you the example of here, Monzo. Monzo is a challenger bank in the UK. Tom Blomfield, the founder of Monzo literally just built this for himself. He'll say that, "Hey, I just wanted something to be the operating center for my finances. Why is it that I have all my money in a banking account? I have stocks, maybe in an e-trade account. I then use Venmo and then I use Mint and I have all these apps. In today's world, my bank should be smart and my bank should be able to do anything I want." And that's literally what he built. The best way to describe Monzo, I often say, is like the Venmo plus Mint, but also does a lot of other things.

They have four and a half million customers. It still shocks people, shocks people, that it's completely product-driven growth. They have close to 8% of the UK market, and it's completely product-driven growth. And their NPS which they publicly report is greater than 70. Name one bank that has NPS greater than 70. It's a bank. One of the best global product engagement metrics. And so we often say, when you're looking at growth, that's what I'm looking for. I'm looking for as an investor which is, show me product-driven growth, because that's the best kind of growth.

If you can show that, then there's nothing that stops you even if you end up spending paid marketing towards the end, like it's going to benefit your organic growth a lot, because people are already talking about you. So word of mouth is a mechanism of organic growth. So in Monzo's case, a lot of people talk about Monzo. They have a very strong community. They also have the Venmo feature which is peer to peer payments, and therefore, an average Monzo user has 16 friends on the Monzo app. And so there's lots of ways in product where you can drive product-driven growth.

Another example of that would be Faire. Faire is a B2B marketplace that connects makers with retailers. It's very counterintuitive to think how will they have any sort of organic growth? Because you're talking about brands, Allbirds or Glossier, or the long tail of Swell water bottle. All these brands that are trying to sell their products into local independent retail, so that there's products that are available at local retail as well. How does Faire drive growth? Faire is simple. They have something called the Faire Elevate product.

What they did is they essentially built an order management system for these brands, because no one caters to these small brands. So they went to the brands and said, "Hey, let me help you build an order management system where you can receive orders and invoice retailers. And by the way, I know that retailers are doing bulk ordering because they're not just buying one water bottle, they're buying bulk water bottles for their store." They were like, "We'll help you offer the retailer free 30 day returns if they're ordering for the first time, or net 60 day financing, we can help them with the working capital for 60 days, depending on you have to be a good performing retailer." And brands loved it. Each brand on average is bringing five retailers to Faire. Faire doesn't spend anything on retailers to acquire them. That's a great viral loop.

If you can find either product-driven growth or some form of viral loop in a two-sided network where there's a reference mechanism, that's gold. That really helps you set up for scale. And so that's what we're looking for when we are diligencing that distribution advantage.

Patrick: [00:40:54] You mentioned earlier the kind of price range typically that a growth stage company will raise equity at. I feel like a really under-discussed item in the venture world is valuation. Obviously, the entry price is a huge determinant over the outcome, and there's no company that can't be ruined by a high enough price. I'm curious how you think through that, and whether or not, now this starts to bleed into your old private equity roots, where you're underwriting a specific range of outcomes. Or if it remains at the growth stage, more of a clearing price that's just a reflection of supply and demand for the company shares.

Anu: [00:41:28] This was something I struggled with a lot when I first came to the Valley because I'm such an analytically-driven and nerd engineer. And to come and suddenly feel like the whole science behind valuation is art. The early stage was rough. But I think look, at the earliest... But now that I've spent close to six years, I think I have a good understanding for how valuation methodology is done. So at the earliest stages, it's really art, it's not science. I think it comes down to if you're raising a Series A, typically the lead investor is taking about 20% to 25% ownership.

So if you say I'm raising 10 million, it's 10 divided by 0.25, that's your valuation. It's as simple as that. And the way you come up with that range of, "Do I want 20% ownership or 25, or sometimes 16?" depends on how strong of a product market fit do you have? There's also a function of supply and demand, as we said. Maybe how competitive the deal is and how it's getting marked up. And also the potential for that company. I think at the end of the day, a Series A investor is still investing because they're hoping for a 10X.

But if it's a grand slam, if it's going to return more than the fund, well, it's okay if you owned only 15% at the start and even got diluted to say 11. It's going to return more than the fund. I think as a Series A investor, that's sort of your mental model. And Series B is also a little more art. So while the ownership is not 25%, people are looking for 10%, maybe 7% to 10%, really hot consumer companies like Uber, Instacart, or DoorDash, they probably would have been sub 10% ownership for the Series B investor. But if it's B2B, it's usually at least 10%. That's the Series A and B. It's literally what's the amount you're raising divided by the ownership percentage that the lead wants.

Series C and onwards, it's different. I don't think it's ownership-driven. It's more... I think as a growth investor, it's very important to focus on dollar return, because you're not playing for 10Xs now. But if you do believe, say you're investing like a 3 billion valuation, and you think this company can be 30 billion or even 20 billion, or even 15 billion. So it's about, "Okay, if I'm putting 100 million of 3 billion, and it's going to be at 15 billion, you're going to make $500 million." A little less than that, because you'll have dilution. But it's about the dollar scale of return. So then you got to compare it to your growth fund, which is like, "Okay. How big is my growth fund and is this a meaningful check size?"

I think as a growth investor, it's very important to focus on dollar return, because you're not playing for 10Xs now.

I like to use this analogy that I actually learned from the DST team, which I think they do a phenomenal job of this is the growth stage. Are you betting the house? Because if you're not betting the house, you're probably not... don't have that conviction. If you're willing to write a $50, $100 million check and let's say it's 10% or 15% of your fund, depending on the size of the fund, you are going all in on that investment. And you're saying that even if it's a 5X, "Look, I'm going to return $500 million." And so the dollar return becomes very important. So how do you then look at valuation?

The valuation is a mix in the growth stage. We actually build a full-fledged model, full five-year model, we use certain valuation comps, not current valuation comps. We actually look at if this company was to go public, what would be the median comps of five or 10-year median multiple? And we look at different multiples. And what will this company be really valued at exit, and therefore, what's the entry price we like now? We usually come up with a threshold of this is the max we're willing to go, just to see, understand that dollar return math. But I don't think you ever turn down a company, because this is the max valuation you want to go to, because then that's your conviction.

And this goes back to my upside case, which is like, "Hey, if you have a 10% probability that this upside case is going to materialize, you're turning this company down because of some strict math and your base case, and you really don't have conviction on the upside case." I think we use the valuation methodology to make sure like we're not setting up the company for failure down the line.

What's most important is you don't want them want to set up too high a price to prepare them for a down round, because it affects the morale of the company. It's not a public company and it really impacts their hiring, all of that. And so you want to make sure you're pricing it somewhat appropriately and not too artifact, but the mental model is more: what's the true exit valuation for this company? What's the range of outcomes? And so therefore, what's my entry price? And how does that meet the typical dollar returns that I'm expecting?

Present and Future Opportunities in the Digital Transformation

Patrick: [00:45:47] You mentioned earlier, a big part of this is obviously the overall opportunity for companies like this to become big businesses, not necessarily just the one that you're evaluating, but just the ability for the market to withhold and build some of these huge outcomes to deliver the returns. That begs the question like how far into this technology deployment era are we? You mentioned the 4% penetration of eCommerce in Latin America, obviously super early, China much further ahead. In the US somewhere in between and accelerated because of COVID. What is your take on this critical, fundamental question of how far along we are in this transformation and how much more opportunity for digital transformation still exists?

Anu: [00:46:29] I actually think there's way more opportunity ahead of us. Let me compare it with a little few numbers. Pre-COVID I had looked at this. The total global market cap was $85 trillion. Internet economy enabled businesses was less than 10%, so roughly 8 trillion. Even if you assume a 10% CAGR and play this out. Let's say in 2045, I think I had seen estimates that if you assume the global market cap is going to be around 450 trillion, Internet economy should surely be at least 15% of that, even if less like $60 trillion economy. Guess what? From $8 trillion to $60 trillion. I'm willing to bet all day long that we are still very, very nascent. Even in the most developed markets.

Let me make it further specific and real for people. Let's look at the US economy. Pre-COVID, our Internet penetration was up 20% in 2019 and I think April reports 27%. A lot has been written about consumer eCommerce penetration. Not much has been said about B2B wholesale eCommerce penetration. B2B wholesale in the US is a $16 trillion market. Less than 8% of it is online. Less than 8%. 49% of B2B wholesale eCommerce transactions happen via phone and fax. And Faire which is one of the marketplaces that's working on it is still just attacking a small sliver of retail. The retail market that they serve is a $670 billion market. You have so many more verticals here.

Think of aerospace, chemicals, industrials. You're just going to see an explosion of vertical players in B2B wholesale eCommerce. B2C consumer eCommerce itself is still sub 30%. So therefore, just the Internet economy we're just still scratching the surface. We just have years to compound, and I think we're still in the early stages of the Internet economy.

Patrick: [00:48:15] I love the tiny untapped B2B wholesale. That's such an interesting example. Everyone always talks about eCommerce because it's, I guess, the most obvious for the average person. What other areas like that do you find especially interesting where digital and tech have maybe begun to make a difference, but are still in the early days, if any?

Anu: [00:48:34] Yeah. I think second area is fintech. Obviously at YC we see a lot of financial services startups. I mean, we ourselves have funded Stripe, Brex, Monzo, LendUp, to name a few. And then Grow recently. But I think there are few trends that are going on in fintech. One, the payments landscape is exploding. If you believe the Internet economy is at least a $60 trillion economy, we've got a lot more work to do on payments, and so there'll be more there, both on the consumer side.

But also as I mentioned, B2B is a big side. Brex started as corporate credit card for startup, but that's just scratching the surface. They're well beyond startups now. They also launched Brex Cash last year. There are a whole suite of products they're working on. So B2B wholesale payments will also be a big area. The third area I'm seeing is in digital and challenger banks. We saw with Chime recently. I just think that if you were to build a bank from today, it just won't look like your prior bank. Like as I said, your phone should be your bank, and it has to be really smart in helping you manage money. Why do I need 10 apps? Make me really smart in managing my own money.

I think you're going to see a whole suite of subscription challenger banks. They're also changing the business model, not just tech. They're saying, "Look, instead of me charging you fees for every single thing, let's just make it clear. Here's your subscription model, here's a set of services you get." And that really excites millennials. They don't want to be charged for every transaction and they don't even know what to expect. Versus you just say, "Hey, here's what you get for $7 a month. Here's what you get for $15 a month." So they're changing the business model too.

And then the third lens in the financial services I'm seeing more in COVID is look, savings yields are at an all-time low globally. So you could have argued in all of the recessions that each country came at a different lens versus now it's just across the board. And so a lot of the users got flat footed. A lot of them are looking for other ways to invest. And that's why Robinhood is doing extremely well. But Robinhood is still, they will undoubtedly be a large player. But they're a small player in a large pie because it's for the Do-It-Yourself market. So we're seeing Robinhood everywhere globally.

Robinhood is will undoubtedly be a large player. But, they're a small player in a large pie because it's for the Do-It-Yourself market. So we're seeing Robinhood everywhere globally.

We're seeing Robinhood for India. We made an investment in Grow. We're seeing Robinhood for Indonesia, Robinhood for LATAM. You'll also see this extend to active asset managers. We're seeing a lot of startups attacking that space. I think in the next decade, you're going to see at least three to five $100 billion fintech companies, at least 10, I'm willing to bet. Maybe 10 $50 billion companies, and two dozen $10 billion companies. That's going to happen in fintech.

And then the third area I'd say, which has been fascinating is EdTech. Historically especially in the US, EdTech businesses have not been huge, therefore, not a lot of investor appetite has gone behind EdTech. But there are two things that are changing. One thing from the lens of Lambda school. I think there's a huge shortage of software developers just in the United States. I mean, there are different stats. Like I think, Okta reported recently that maybe there's a need for 200,000 developers, but I've seen others say, we need 500,000 developers in the US every year. And we have less than 50,000 computer science graduates, less than 50,000 computer science graduates.

And historically, we have tried to fill it through immigration and other means. And given the geopolitical climate, immigration is the question mark right now. We don't know how it gets impacted. But the short story is there's a huge shortfall. And if there's a shortfall, there comes Lambda school. They just, through online resources and a nine-month intensive program, are able to get you to the stage of starting your journey as a software developer. And for many people, it's increasing the salary from $15-25,000 a year to $75,000 a year. That's game changing. That's one trend we're seeing.

Outschool is another YC company that's done a phenomenal job especially during COVID with all these Zoom classes, where public school depending on state, city, councils, you don't have enough budget to support teachers in certain areas. And here comes Outschool. They just have the marketplace of best teachers. And if you've gotten over the inertia to take a class online, Outschool will play a role even post-COVID. Because after you attend public school, if your child has deep interest in science, you can just sign up an Outschool class with three other students and pick the teacher that you want, and they can go deeper. I think you're going to see interesting times in EdTech as well that's going to actually explode in the next 10 years. It'll be fascinating to see, what are some interesting business models that come out of that.

Patrick: [00:52:51] Do you have any observations on the changing nature of the US technology investing ecosystem? And I don't just mean growth equity, I mean, any observation around from seed all the way through the biggest public equity managers, that has changed across your career in technology investing specifically. Like what would be the most notable difference from day one to today?

Anu: [00:53:14] We often joke at YC that we have more investors than we have companies. That's actually true. Our Demo Day gets close to 2000 investors, and per batch we have only 250 companies. It's a good problem to have, but I just think you're seeing the number of investors increasing across every stage. Having said that, I think the misconception people have is there's money everywhere and how are you going to differentiate? I think at the growth stage, it's not true that there's that many investors.

So if you ask our founders, both Series B or Series C where I spent time with founders, our founders are a lot. They can't name more than 10 investors who can write $100 million plus check. They can't. And so it's the same usual suspects. You could come up maybe with a list of 10 to 15, and you'd really scrape that list, and they don't even know them, but it's the same. Yeah, the crossover funds have started investing in startups, yeah more private equity funds have started investing in startups, but the thing is, unfortunately, most of them don't understand the pain of a startup. They don't.

It's great when you get 14 term sheets, how do you win, but the company is not going to look up and to the right every year. It just doesn't happen. I mean, YC has so much data on this. I mean, we've funded 2500 companies, less than 1% of our companies have greater than a billion-dollar valuations. We have maybe 20 to 25 of them. And I can tell you pretty much, every single one of them has had at least one or two really hard years in their growth stage. Really hard years. You often see that these investors that are new to the growth stage are working with startups cannot comprehend that, because your full-time job is really to focus. If you're a public market investor, your full-time job is really to manage your portfolio and your day-to-day investments and the volatility associated with that.

[YC has] funded 2500 companies, less than 1% of our companies have greater than a billion-dollar valuations. We have maybe 20 to 25 of them. And I can tell you pretty much, every single one of them has had at least one or two really hard years in their growth stage. Really hard years.

And then you sometimes may be more afraid or chickened out by like, oh, the quarterly projection didn't really foot with this, or the market landscape is changing. Or like the most common thing that I mentioned at the start of this session was, you fear when a competitor or an incumbent especially launches something, because you immediately think, "Oh, it's a zero-sum game." I think founders are usually very attentive to who they are working with and which partner they want to bring. And especially in this lens, they want someone who really understands startups. I think you don't see a lot of growth stage investors who truly understand startups.

And pattern recognition is a dangerous thing. Because when you use pattern recognition, on one hand, you may say it helps you but actually the zero-sum game always comes to hurt you. Because if you followed the software market and you followed Microsoft, you would think zero-sum game is real. But that didn't play out in the Internet eCommerce world. I mean, even China, everyone thought Alibaba would be the only player. We have Alibaba, we have JD, we have Meituan, we have Pinduoduo. I mean, look at the commerce market that's supporting China. I think there are far fewer investors on the growth stage that truly understand startups, even though there might be a few more players that are willing to invest money. So that's one of the ones that I offer which is probably different from what the people think on perception is outside.

Second difference, I'd say SPACs. There's a lot of talk about SPACs now. I think there have been SPACs that have existed for years. It has happened in the past. But what's different about this time around is the underlying manager that's sponsoring the SPAC. I paid attention when Reid Hoffman launched the SPAC, and Micky Malka launched the SPAC. I mean, I work with both of them on two boards. I mean, Reid is on the Conway board and Micky is on... Three of the boards actually we share with Micky. These are both just phenomenal investors, incredible investor that a startup founder would love to have, and they truly understand the pains of scaling a startup.

So if you use from that lens, you do see, can a SPAC really play a role? And I do think there are certain technology companies that could benefit from an SPAC. So these fintech companies are extremely in regulated spaces. And they need a lot of regulatory capital and regulatory buffers, which I think people don't understand. And SPACs depending on the underlying sponsor could actually play a big difference for those companies. So I do think I can see the role of a SPAC there.

And then the other argument I've heard from especially like Brad at Altimeter is most founders don't know who your day one traders are. And I think that's fascinating. Because if you ask a founder, they always want to know who your investors are. Because they want to know, who are the true partners in your journey who are believers in your company. They don't want someone who's just flipping after day one. That could be potentially a promise that attracts. But I think the question remains to be seen, can the next Snowflake go on an SPAC? I think if that happens, there could be more SPACs.

Patrick: [00:57:46] I have to turn now to my traditional closing question for everybody, which is to ask you for the kindest thing that anyone's ever done for you.

Anu: [00:57:53] Well, there are a lot of people that have really helped me, but I think the kindest thing that comes to mind is Dr. Jeffrey Reed. He was my research advisor at Virginia Tech. I was doing my Master's in wireless communications. This was in 2002, right after 9/11. And the funding that most state universities got from the government was completely slashed. Even private funding was at an all-time low. So I had come with the hope of getting a research fellowship. The university and the research group had no money to be able to fund me.

I come from a Tier-3 town in India. My parents were very middle class, and my dad had pretty much taken an entire loan against all his assets and could pay only for a year of my tuition. Come summer, and I was working enough in Virginia Tech to cover all my living expenses. But as an international student, you can work only 20 hours and you have to work in campus. You can't work outside. So there's only so much I could do. And I remember very vividly, this was July, and my dad basically said, "Look, this is it. This is the last straw. Finish whatever credits you can, you're going to come back in August." And I was like, "Yeah, I get it."

And so I went to my advisor and said, "Look, I really can't continue, and I need to find a way to graduate. So I should finish whatever credits I can in the summer, and maybe I could do it remotely. Would you be open to doing remotely?" He asked me, "How much money do you need?" And I said, "Well, I haven't paid tuition this month. I need to pay $1,200." He took a checkbook, wrote a check, and gave it to me without any questions. And I think if I look back in life, if he hadn't done that, my life would have turned out very different.

Patrick: [00:59:22] I'm amazed by these little, tiny things, obviously not tiny, significant things that people do constantly all over the place. That's a very unique one. I haven't heard one quite like that yet, after doing more than 200 of these, so I love the story. And thank you so much for a really interesting hour and a half. I've learned so much. I love learning about the growth stage where I've spent a lot less time and just really appreciate all your insight.

Anu: [00:59:42] Great. Well, thank you, Patrick.

Created by @babyxcar and @ghickmantweets