The Firms that Fueled Big Tech: Sebastian Mallaby on Venture Capital
"It seems to me that the creation of new companies based on applied science that deliver new products that people buy is a pretty good definition of something we want in society."
MONEY
Technological innovation and the startups that power it have been a key engine of U.S. economic growth and dynamism for decades now. The builders of the country’s biggest tech firms—Jeff Bezos, Steve Jobs, Elon Musk, Bill Gates, Larry Page, and Sergey Brin—have all become icons. Outside financial circles, however, the critical role that venture capitalists have played in the tech boom has been largely ignored. Yet VCs have backed, enabled and supported most of Silicon Valley’s most successful entrepreneurs, and it’s hard to imagine the U.S. tech sector succeeding at this level without VC support. In his new book, The Power Law, Sebastian Mallaby—an author, financial journalist, and fellow at the Council on Foreign Relations—tells the overlooked history of these investors and their impact. We spoke recently about the history of venture capital, its key contributions and missteps, and how the United States can continue to use markets to drive creativity and transformation.
Octavian Report: Explain where the name of your book, “The Power Law,” came from.
Sebastian Mallaby: The correct way of expressing the Power Law is to say it’s the 80:20 rule under which 80 percent of the bets by venture capital firms don’t do very much. Some will generate 20 percent of the returns, and many will go to zero. Meanwhile, 20 percent of the bets will generate 80 percent of the returns. Of course, there’s nothing magic about 80 and 20; in reality, the ratio is probably more like 90:10 or 95:5, because venture capital is a game of grand slams. Companies often fail, but a few earn more than 10 times their initial investment. That’s very different from a hedge-fund portfolio distribution, where if you had a position that went to zero, you’d be very unhappy indeed—but you also wouldn’t expect to make 10x or 20x or 30x on one position.
Another way to clarify the Power Law is to contrast it with a normal distribution on a chart, where if you plotted it, you’d see that familiar bell curve, with most of the observations clustering around the average—as opposed to the average being driven by the success of a few outliers on the right-side tail.
OR: How do you view the risk-reward calculus of venture capital versus other types of investment, such as hedge funds or private equity?
Mallaby: The data show that the elite 20 percent or so of venture partnerships have an extremely good risk-reward ratio, better than other asset classes, because they win a large chunk of that sector’s total returns. With median venture capital funds, however, things are different. Sometimes they outperform the public stock index and sometimes they don’t. So the risk-reward ratio for them is pretty lousy, particularly when you take account of the fact that they involve very long lockups in illiquid positions. If you’re an investor in a median fund, you’re not really compensated for that.
OR: How much of venture capital performance involves luck? Other than Sequoia, all the VC firms you write about seem to have a moment and then fade at some point. So how much of their success involves just being in the right place at the right time, as opposed to being really good talent-spotters or really good business-builders?
Mallaby: You can tell a plausible story whereby 100 venture capitalists open up for business and by the law of probability, a handful win big in the first year or two. If these funds have two lucky hits each, that starts to look like a trend. So the best entrepreneurs will go to them for capital first, because it looks good to have high-reputation VCs supporting you. The high-reputation VCs will thus get better deal-flow, and may be able to get into deals more cheaply.
These opportunities give the early winners big advantages. They can buy low, sell high. How could you mess that up? So to answer your question, initial luck followed by path dependency could explain what’s going on. This concept is fleshed out in some interesting social science research, which I quote in my conclusion. In one experiment using digital music, you create an online playlist of 12 songs. You then invite people to go on your website one-by-one, listen to the songs, and download the ones they like. Subsequent visitors can then see which songs have been downloaded a lot. So social proof starts to kick in and at the end of the experiment, you find that one song is far and away the most popular.
Does that mean the song is the best? Well, if you run the experiment again with the same 12 songs but a different bunch of volunteers, you’ll find that they again create a smash hit—but it’s a different one this time, because whichever song happens to be downloaded by the first five or 10 people creates a pattern that then becomes self-fulfilling. Now, is this true in venture capital? A little bit. When people try to measure this statistically, they find that, for example, an early IPO backed by a venture partnership will lead to a higher chance of that VC firm getting future IPOs. But the increase is only about 1.6 percent. So it’s not dispositive.
Another test is just to look at which were the top venture partnerships in the 1970s versus the 1980s versus the first half of the 1990s. If you split it up by periods like that, you do see a decent amount of churn in which are the dominant VCs. And if there is this all that churn, and path dependency doesn’t have such a big impact, you can conclude that the kind of the companies that do perform well over a long period, such as Sequoia or Benchmark, probably do have skill.
After doing this research, I then went and talked to people at the successful firms. Some of the stories I got seemed cute and were unpersuasive as to where their good returns came from. But in other cases, I got stories that made me feel like I was talking to people who were just better at the work than the other people I’d interviewed. At the beginning of my research, I took the luck hypothesis so seriously that I almost didn’t write the book. But when I started to see the skill, and to read deeper into the social-science literature, I came to the view that there is a real difference between top firms and the rest. Part of my motive for writing the book was to illustrate that skill.
OR: If you were a venture capitalist, what playbook would you follow?
Mallaby: There are various debates about what’s the best methodology. One question is, do you pick the jockey or the horse? In other words, do you focus more on the character and quality of the individual who you’re backing, or more on the market that that person is going after? Another is whether you should try to generate returns by picking the right deal, or by working hard on the company after you’ve invested in it. In the end, however, all these sorts of debates turn out to be silly, because most venture capitalists pick both strategies.
One outlier is Peter Thiel, who set up Founders Fund; he argues that with early-stage investments, you should focus on choosing the right deal and then basically check out. Thiel has done very well, and his point is not stupid. For a company to do really well, it has to be founded by somebody who has a really idiosyncratic vision and pursues it with maximum energy. If venture capitalists get too involved, they’re going to dilute that special quality. They may save the startup from errors, but they’ll also prevent it from achieving great success. And since it’s great success that matters, you should let the eccentrics have their way and not mess with them.
OR: Does VC have a problem with herd mentalities?
Mallaby: I think there are two issues here. There is definitely a herd mentality on Sand Hill Road [where many Silicon Valley VC firms are based]. In fact, if you wanted to invent a hypothetical machine for generating bubbles, you would put all the investors on one road, make sure there are only two restaurants on that road, make sure that you cannot short the instrument—which is the case with private venture deals—and ensure that you cannot even speak negatively about deals because you’re hoping to be syndicated into other deals by other people, and so to excommunicate yourself from the network would be suicide.
That’s what today’s venture-capital industry actually looks like, and it’s a recipe for generating bubbles. The only thing that stops the bubbles from being even more outrageous than they already are is public markets, because they set the exit price for VC deals, and in public markets there’s less herding, you can go short, and you can trash the other guys’ investments. Public markets create incentives for contrarians, which makes those markets better at price discovery, and that ripples down into the private-deal ecosystem and sometimes disciplines the prices that VCs pay.
Another aspect of the question about herding involves the fact that while many individuals miss out on VC deals like Apple, the overall system, which did fund Apple, nonetheless often gets it right. The explanation for how this works involves a subtle point about why Silicon Valley has done so well at innovation. It’s because you’ve got a cluster of entrepreneurs and talent and engineers and people who understand how to scale marketing operations at a startup. You’ve got lawyers who do startups. You’ve got real estate agents who are used to renting to startups. You’ve got everything.
And the network is more than the sum of its parts. One of the ways it functions is that there are some deals that appeal to particular venture capitalists, and some that just don’t—even though they are also perfectly good deals. The reason for this phenomenon, as one venture capitalist explained it to me, is that, as a specific VC, you can only understand some technologies and relate to some individual personalities. And if you do get involved with a startup, you’re suddenly in a three-legged race with your leg tied to the other guy’s leg—so you’d better get along.
That explains what happened with Apple. At the time they went looking for funding, Steve Jobs was a pretty strange character: he’d dropped out of college before dropping out was fashionable, he didn’t wear socks half the time—or even shoes—and he had a fruitarian diet. A lot of the more traditional VCs saw his scraggly beard, thought he looked like Ho Chi Minh, and ran away. But then Mike Markkula, who was more of an angel investor than a VC, happened to visit Apple’s famous garage and see the circuit board that [Apple cofounder] Steve Wozniak was working on. And Markkula knew enough about circuit boards to say, “Okay, that is one cool design. So I’m going to look past Steve Jobs’ off-putting arrogance, his youth, his lack of experience, and the fact that Steve Wozniak tends to look at the floor when he’s talking to you, and I’m going to get involved anyway because this is a very clever circuit design.” In other words, the investment ended up suiting one person—Markkula—and that’s all Apple needed.
OR: You point out in the book that unicorns have bad corporate governance and that they’ve turned into huge wasters of capital. Have VCs created an ecosystem that is developing more WeWorks, as opposed to Apples or Googles?
Mallaby: In the late 1990s, Amazon went public with a valuation of less than $500 million. So much of its growth took place after the IPO. Today, by contrast, when big companies go public, their valuation can be as high as $100 billion—200 times the size of Amazon’s. What changed is that you’ve now got different people involved in the public markets. Yuri Milner is the guy who brought the new model to the United States when he invested in Facebook in 2009. Tiger Global had created the model abroad, and then quickly followed Yuri Milner in the United States, as did other hedge funds.
OR: And for these investors, a critical factor is their check size, right?
Mallaby: Exactly. They get into the deal partly because they can write very big checks, which is flattering to the founder. Essentially the conversation they have is, “You’ve created an absolutely genius company, which is already worth north of a billion dollars. You are changing the history of Silicon Valley, and far be it from us investors to tell you what to do. We want a board seat, but we trust you absolutely to do the right thing and we will vote with you on any board decision.” That’s a big difference from the earliest venture capitalists, who also wanted a board seat and shares, but voted as they saw fit—and would challenge the founder if they thought he or she was going in the wrong direction. I think the newer governance model, of excessive deference to the founder, is damaging, because we’re all human and even the most brilliant among us can benefit from having checks and balances.
Especially since the qualities that are needed for building a startup from zero to a billion dollars are not the same skills that you need to go from a billion to 100 billion. Some founders are able to do both very well, but not everybody can. So I see zero downside in having a board that has teeth and that defers to the founders when they’re doing a great job, but asks difficult questions and forces course-corrections when needed. Had the investors been in charge at WeWork, they might have stepped in before things got out of hand. I think they would have done that at Uber. I have a strong view that you should either govern companies via public markets—where poor management will be penalized by people shorting the stock and making the company a candidate for a takeover— or you should have governance in the traditional VC way, where the investor goes on the board and exercises real oversight. Or you should use the private-equity model, where the firm buys the whole company and runs it. But what you don’t want is this messy middle ground where the biggest investors explicitly don’t want to have any governance.
OR: I think you point out in the book that four out of the six founders of PayPal had built bombs in high school. What do you make of that? Are there an unusual number of jerks in Silicon Valley, or is that just part of the mythology?
Mallaby: You have to be very driven to become an entrepreneur. You need to be single minded, you need to be willing to bet five years of your life on a project that, when you begin, will seem outlandish to most people, and you have to bust your way through the fact that in the early days, it’s going to be very difficult to persuade anybody to come and work for you. It’s going to be tough to find your first customer. You may be messed around by a big company that suddenly decides to get into your market and compete with you. You may have to pivot and you’ll probably fail. Because most startups fail, right? All this is grueling, and not for normal people.
So I think that the Valley does attract extreme characters. And I think that venture capital is a less-pronounced case of the same phenomenon, because a decent number of VCs are led by former entrepreneurs. They may have mellowed, but they’ve still got some of that DNA.
OR: There are a few people conspicuously absent from your book or relegated to supporting roles. One is Elon Musk, another is Bill Gates, and another is Jeff Bezos. Also, you talk about Genentech, but life sciences seem not to be the focus of Silicon Valley. Why do you think that is?
Mallaby: My book is largely focused on Silicon Valley, but I think Seattle is an interesting parallel story to the one that I tell. Although Amazon was a classic venture-backed company at the beginning, it didn’t need money thereafter, because it did an IPO early. Bill Gates, meanwhile, founded his company without a venture capitalist and only grudgingly allowed one to invest a bit later on because he thought the connections would be helpful. So both of the Seattle megahits were not strongly venture-derived.
Now, why is my story about software and not other kinds of companies? Because starting with the arrival of the Internet in the 1990s, the companies that really did super well were software companies. They didn’t need much money at the beginning because their product was basically just code, and they could generate absolutely massive market share very fast. If you look at a chart of venture capital returns by sector, you’ll see that in the last 25 years, software has overwhelmingly dominated in terms of value creation. In Mark Andreessen’s words, software has eaten the world, while both life sciences and hardware have generated much lower returns. That’s because in those sectors, you need to put in more capital at the beginning, and it takes longer to develop the product. More capital plus more time equals lower returns.
Now, some people have drawn the conclusion that venture capital is therefore only good at investing in software firms, but I strongly resist that conclusion. Because there is a pre-history to the software phase of venture capital, which involves companies like Fairchild Semiconductor, Apple, Genentech, 3Com (a networking company), Cisco (which does routers), and UUNET (which did the plumbing of the Internet). These are all hardware companies, or in one case a biotech company.
OR: Do you think we’re going to see more people like Musk doing the kind of things he’s doing in transportation?
Mallaby: Musk is an extreme character. To have founded either SpaceX or Tesla would be astonishing, and to have done both is more than astonishing. And he did it off the back of two prior startups, one of which made him a few million dollars and the other of which was PayPal, which made him significant money. Now it happened that he was obsessed with cars and space, so once he had enough money, he decided to take on these hardware projects, which were technically really, really hard. But he could do them because he had his own money, which meant he could operate outside the venture capital ecosystem. He did take VC investment in SpaceX, but that’s not how these companies were initiated.
So Musk is a strange, idiosyncratic man who’s done what he wants. The more interesting question is whether other hardware founders can get going in a more normal fashion without being multimillionaires before they begin. I think the answer is yes, because there are partnerships such as Lux Capital or Peter Thiel’s company that have had a lot of success with hard tech.
OR: What role did government play in creating or hindering the startup ecosystem, and then what do you think the right role for government is today?
Mallaby: This is one area where I disagree with some of the preexisting literature. The economist Mariana Mazzucato, for example, has argued that the ultimate venture capitalists are governments. Exhibit A in this type of argument is the Internet. The early Internet was invented and rolled out by DARPA, the Pentagon’s research-funding operation. At first, the Internet was a system for military communications that had high levels of redundancy, so that in the event of an attack by an enemy power, you could still communicate across the country. The project involved government scientists and government-backed university research programs. And in the early years, if you were not in the government, you were not allowed on the Internet, and if you were trying to do anything commercial, you were totally not allowed on the Internet.
But things started to shift before the Internet had a million users, because people started to recognize that this government-run Internet could be of massive interest to people outside the government. Some of those pushing for change were former government-backed scientists who didn’t want to be kicked off the email and messaging-board networks that the Internet offered just because they took a job somewhere else. Then there were those, like then Senator Al Gore, who saw the Internet’s potential for delivering content like online education or entertainment. And then there was a more practical vision, which was backed by a trio of venture capital companies: Axel Capital, Menlo Ventures, and New Enterprise Associates. And they got behind the effort of turning the existing telephone network into a data network. So yes, the government created the initial idea of a distributed network of communication. But the force that turned it into a product that effects everybody’s lives was venture capital.
Now, I’m not against government involvement in backing technology ecosystems; in fact, I’m very much in favor. But I would say that the key inputs that the governments can have are, first of all, backing fundamental science at universities. The second is backing scientific education at high school and at universities. And the third is to ensure a legal and tax regime that is friendly to venture capital, because venture capital can turn fundamental science into products that make us all live more productively.
OR: Do you think that Silicon Valley’s culture and process has been successfully internalized in China? Or do you think that as China moves away from free markets, it will lose some of what helped build a lot of the big tech companies there?
Mallaby: Starting in the late 1990s, U.S. venture capital moved into China and backed all of the famous early Chinese Internet companies, like Baidu, Alibaba, TenCent, NetEase, etc. In the process, U.S. VC firms taught China the template: you’ve got to embrace risk, and you’ve got to understand the Power Law, which means that many startups will fail but a few will hit the bigtime and pay you back for those failures. They taught China that when you’re starting VC firms, you set up limited partnerships and then partner with the companies you invest in. You go on their boards and you don’t give all the capital up front; you invest stage by stage, so that if the company fails, it will fail more cheaply. All these ideas were transferred to Chinese practitioners who totally understand them, so individual American VC investors are not needed anymore to make the system work.
But China now faces a trickier question, which stems from the fact that the government is clamping down on the portfolio companies that VC firms have backed and making arbitrary rules changes, such as suddenly deciding that online education companies are not okay or that firms like Didi, which listed in the U.S., should be punished. These changes are damaging the business climate. Venture capital is a long game—you invest in a startup and might have to wait 10 years before it has an exit and you get paid. If you’re not sure what the regulatory environment will be like 10 years from now, it’s harder to make intelligent bets. So while some investors will continue to invest, the uncertainty is going to have a dampening effect—although my guess is that in short term, you’re not going to see a lot of deceleration in Chinese venture activity or innovation, because it’s still a huge market that, for all its troubles, is going to continue to grow faster rate than the United States for the next five, 10 years.
OR: Do you think the United States is doing enough on AI?
Mallaby: It’s doing quite a lot. These days, when you type an email, the system auto-fills the rest of your sentence. There are all of these small improvements happening in software that are making our lives easier, but they’re so continuous and being delivered so seamlessly that we take them for granted. We don’t realize that AI is already here and is already changing the way we operate.
The big companies—Google, Facebook, Microsoft, Amazon, Apple—have enormous teams focused on AI. It’s less about startups, because AI is pretty capital intensive and the big guys have a head start. So I’m not too worried about the United States falling behind.
OR: Overall, do you think venture capital has been a positive or negative force for the world?
Mallaby: I’d say it’s been hugely positive. In the book’s conclusion, I go through reasons why people would disagree. The only one I agree with is that venture capital would be even better if it was more diverse. If you’re trying to invent the future for all of society, then your investment team should look a bit more like society. It’s just not acceptable that women and minorities are massively underrepresented at Silicon Valley venture partnerships. But beyond that, it seems to me that the creation of new companies based on applied science that deliver new products that people buy is a pretty good definition of something we want in society. Of course, there are certain products that may be net negative for society. But I think those are the minority of the products that VCs back. And if certain products are genuinely bad for society, then government should regulate them. But to be anti-venture capital is to get the wrong end of the stick.
One final point: there’s a sense out there right now that Big Tech is too big, that the platform companies with network advantages have too much market power. Well, if you don’t like the power of Big Tech, you should be in favor of small tech—meaning smaller companies backed by venture capital that challenge the power of the big platform incumbents.
This interview has been edited for length and clarity.
Someone needs to check the data behind the author's 80/20 rule. One place to check is the work done by my former colleague, Prof. Josh Lerner, at Harvard Business School. There are other sources that track these data. In fact, the success ratios are far lower when deals are measured over time, say 5 years. Moreover, ventures for products and services launched within existing large companies do not fair much better. With few exceptions, business concepts fail on account of market fundamentals, particularly absence of demand and paying customers, and effects of direct and indirect competition. Both of these factors lead to revenue shortfalls, Revenue is the sine qua non in business. Revenue pays for everything else. Without durable and profitable revenue streams, all fall down. Next, a large share of capital that has been flooding into the venture capital funds over the past 20 years has gone to financial engineering. Likewise, risk capital going into PE funds is used mainly for roll ups and financial engineering. As in the past, the primary source of seed capital continues to be sweat equity, family and friends, and angel investors well in advance of approaches to VC funds. The success rates at this stage are low. Investments by professionally managed funds and by individuals and groups of them that end up clinging to life in the Land of the Living Dead cannot be considered successes. All in, the odds against success in venture investing are overwhelming. A positive expected ROR is generated by the one or two winners. You get the idea. The name of the game is to crank up massive amounts of assets under management, clip off the 2 percent "management fee," and hope for the best from the 20 percent interest in capital gains from the fund. This incentive structure led to an interesting period in the late 1990s to 2000, when VC fund managers flipped early stage companies into the public market, including those with no revenues and no profits, and declared victory and clipped off 20 percent of the IPO gains. I recently threw out 4 bankers boxes of S-1s from IB s who specialized in this business, e.g., Roberston & Stepherns, et al. You get the idea.