下载掌阅APP，畅读海量书库
立即打开
In the interest of unlocking the somewhat opaque doors of venture capital, behind which are the inner workings, incentives, and decision-making processes of VCs, let me start by more properly introducing myself.
The first thing to know about me is that if I weren’t a venture capitalist, I would sing country music in Nashville. But lucky for everyone who is a real country music fan—and for my ability to support my family financially—I somehow found my way into the VC business! I live in Silicon Valley, not Tennessee, so the best I can do is wear cowboy boots to work and play the guitar in my spare time. Both of which I do, as often as possible.
Let me give you a little bit of context about what the tech and investment world was like when I was getting started in in the 1990s.
Some of the big tech names back then were E.piphany, NetIQ, VA Linux, Commerce One, Razorfish, and Ask.com. It’s possible you haven’t heard of any of these companies, but they—like me—were products of the 1999–2000 tech bubble that produced roughly nine hundred initial public offerings of venture-backed tech companies. It was a great time to be starting out in the tech industry, as there seemed to be no end to the promise of technology and to the amount of wealth creation that was available to everyone involved.
Netscape had gone public in 1995, a mere eighteen months after its founding, receiving a huge amount of media attention and heralding the beginning of the dot-com boom. Google wouldn’t be founded until 1998, but Silicon Valley was already fired up with dot-com fever. New internet startups were appearing daily. The tech world was abuzz.
Venture capitalists (VCs) were investing in new companies at an unprecedented pace relative to historical norms. About $36 billion went into new startups in 1999, which was approximately double what had been invested the prior year (although that’s now less than half of what was invested in 2017). Additionally, limited partners committed more than $100 billion of new capital to the venture capital industry in 2000, a record that hasn’t come close to being broken since! By comparison, limited partners committed about $33 billion in funding in 2017.
Startups were also getting to an IPO faster than ever during the dot-com bubble. On average, it was taking companies about four years from founding to go public, which was a huge acceleration of the historical trend of taking six and a half to seven years to IPO. Today, that time period often exceeds ten years, for reasons we’ll get into later in this book.
In addition to a record number of IPOs, the public markets were also exuberant. On March 10, 2000, the Nasdaq index, the barometer for technology stocks, peaked just above 5,000. More interesting, the price-to-earnings ratio (P/E ratio) of the companies listed in the Nasdaq index stood at 175. This means that stock market investors were valuing one dollar’s worth of a company’s earnings at $175.
While it’s generally the case that investors value a dollar of earnings today at some multiple greater than one because a company’s stock price is intended to reflect the present value of the cumulative cash flows of a business into the future, a 175 multiple is a historical anomaly. For comparison, the Nasdaq P/E ratio today is under 20, which is generally in line with the long-term historical trends for the index.
At the time, Cisco was anticipated by many to become the first $1 trillion market capitalization company. Alas, Cisco’s market cap peaked at about $555 billion in March 2000; today it stands around $200 billion. Early in 2018, Amazon became the first $1 trillion market cap company, albeit for a brief time, and as of this writing, sits at around $800 billion. (Fun fact: In March 2000, Amazon’s market cap was a mere $30 billion).
So, back in 2000, everyone was on a collective sugar high to end all sugar highs. What could possibly go wrong? As it turns out, a lot.
The Nasdaq index began a precipitous decline from its March 2000 peak, falling all the way to its nadir of just above 1,300 in August 2002. While there is much Monday-morning quarterbacking about the impetus for the decline, many market analysts point to the Federal Reserve’s aggressive interest rate tightening in early 2000, which created a big debate as to the sustainability of heavy borrowing that many technology infrastructure companies had undertaken. Regardless of the ultimate cause, in about two and a half years, the index lost nearly 80 percent of its value, tech companies laid off record numbers of employees, VCs stopped investing in new companies, and the few companies that had sufficient cash to sustain themselves were focused purely on self-preservation at the expense of everything else.
That’s why you probably don’t remember most of the companies I mentioned before. Yet this was the environment in which I began my professional career.
Despite graduating from Stanford University in 1993 and Stanford Law School in 1996, sitting right at the epicenter of the tech boom the whole time, I was largely oblivious to what was happening around me. So, after graduating from law school, I left Silicon Valley to spend a year in my hometown of Houston, Texas, clerking for the Court of Appeals for the Fifth Circuit. This was an incredible learning experience and a fun way to spend a year, but, as it would turn out, it had zero relevance to my longer-term career.
I moved back to Silicon Valley to work for Lehman Brothers. Lehman, of course, was later a victim of the global financial crisis, suffering an ignominious bankruptcy in September 2008. My job at the time, in addition to being an all-around grunt, was to help life sciences companies raise capital, go public, and make acquisitions. Those were noble things to do, but for the fact that despite the raging bull market in technology in Silicon Valley, the investor appetite for life sciences was largely dormant.
Lucky for me, a friend had just taken a job at Credit Suisse First Boston, a scrappy investment bank that had brought on Frank Quattrone to build out their technology banking practice. Frank is a legend in the technology banking world, having started his career at Morgan Stanley, where he led IPOs for companies such as Apple and Cisco and advised on a huge range of important mergers and acquisitions. He is still a dominant figure in the technology space, having founded in March 2008 a leading mergers and acquisitions advisory firm named Qatalyst.
So I joined Credit Suisse First Boston and drank from the fire hose of the developing tech bubble. A few years into my job, on the eve of finishing an IPO for E.piphany, one of the marketing executives I had worked with to help them prepare for the IPO told me he was leaving to join a new startup called LoudCloud. Cofounded by Marc Andreessen, the already revered cofounder of Netscape, LoudCloud was trying to create a compute utility (much like Amazon Web Services has now created). Among the other cofounders was Ben Horowitz.
This was the fall of 1999, and the dot-com excitement was in full swing. I had finally opened my eyes to what was happening around me, and I wanted to be a part of it. When my friend at E.piphany offered me the chance to meet Marc Andreessen and Ben Horowitz and see what they were doing, it was too much to pass up. My wife, who was about five months pregnant at the time with our first child and who was busy closing on the first house we were buying together, didn’t see it quite the same way. She had a pretty good argument, to be honest. Why quit a great job with Credit Suisse First Boston where the business was going gangbusters, meaning the chance for both financial and professional success was palpable, in order to join a startup where I’d get paid next to nothing in salary for the promise of some equity appreciation in the future from stock options? She did, however, ultimately acquiesce, likely against her better judgment at the time.
I’ll never forget my interview with Marc. Although I had never met him before, like everyone in the tech industry I knew of his accomplishments and media fame. So when he asked me to meet him at a little Denny’s restaurant in Sunnyvale for my interview, I was a bit surprised.
But it didn’t take long to get excited about the LoudCloud market opportunity. Marc took a napkin from the table and began drawing some barely decipherable sketch of how LoudCloud was going to take over the computing world. Only now, with the benefit of more than eighteen years of working with Marc, have I come to learn that doodling in all its glory is among his many skills.
The idea of LoudCloud was elegant in its simplicity; it turned out that the execution of the business was anything but. In basic terms, Loudcloud sought to turn computing power into a utility. Just as when you plug your phone charger into the wall socket you don’t need to know (or care) about how the electricity got there, you just use it, LoudCloud’s mission was to do the same for computing capacity. As an engineer, you should be able to develop your custom application and then just “plug it in” to the compute utility that could run the application seamlessly for you. You shouldn’t have to worry about what kind of database, networking equipment, application servers, etc., underlie the utility; it should simply work. It was a great idea—one that Amazon Web Services has built into a multibillion-dollar business today.
LoudCloud was probably about ten years ahead of its time, an oft-repeated lesson, by the way, in the startup world. Though timing isn’t everything, timing is definitely something—it’s a big reason why we now see many ideas that failed in the dot-com bubble being reincarnated as successful businesses two decades later. As market conditions change—in the case of the dot-com businesses, the market size of available customers was simply too small relative to the cost of acquiring those customers—business models that previously failed can become viable. Marc likes to remind us that when he was building Netscape, the total size of the internet population was about 50 million people, nearly all of whom were accessing the internet on clunky dial-up connections. Thus, no matter how much utility the browser provided, the end-user market simply wasn’t that big. Contrast that to today, where we have about 2.5 billion smartphone users with ubiquitous connectivity to the internet and the potential for that number to double over the next ten years. All of a sudden, businesses that couldn’t work profitably at 50 million users take on a very different look when they can appeal to a mass-market audience.
After meeting with Marc, I also interviewed with a number of other members of the team, including cofounder Ben Horowitz. The setting for that interview was more normal, as we met on a Saturday at the company’s offices. But I remember being surprised by Ben’s attire—he was fully decked out in Oakland Raiders garb, including T-shirt, watch, and baseball hat. I now know, after many years of working side by side with Ben, that his attire was completely in character. In fact, to this day, Ben keeps a life-size dummy of a fully outfitted Oakland Raiders football player in his office. For the uninitiated, that can be quite a surprise!
I got the job as a business development manager at LoudCloud. This title was the euphemistic way of saying, “You were an investment banker in your previous job and might have some skills to add to the company, but we’re not quite sure yet exactly what those will be.” (Over my seven-year tenure at LoudCloud, I had the opportunity to take on a number of different roles, including running financial planning and investor relations, corporate development, some engineering teams, customer support, and field operations, which included support, professional services, and pre-sales engineering.)
I was in, I was thrilled (my wife was less so), and we at LoudCloud set out to build the first compute utility, flush with what we thought was plenty of cash. In its first few months, the company had raised nearly $60 million of debt and equity. But then again it was early 2000 and we were all living the dot-com dream. VC money was raining from the rafters.
We naturally decided to raise more money—$120 million, to be exact. In some respects the money was free (as the valuation at which we were able to raise was over $800 million—this for a less-than-one-year-old company!). But it was not in fact free, for with it came the expectations of growth for which the VCs had provided the money.
And grow we did. We topped six hundred employees before the company was even two years old. We decided to go public in March 2001, which was definitely not the greatest timing, right in the wake of the dot-com meltdown. In fact, LoudCloud was one of only a very small number of tech companies to go public that year (fewer than twenty tech IPOs happened in 2001, versus nearly five hundred in the prior year). The portfolio managers with whom we met during the IPO road show of back-to-back meetings could not have been more shell-shocked about the decimation they were seeing in their portfolios. They looked at us as if we had three heads when we dutifully gave the LoudCloud IPO pitch. Recall that Nasdaq was at about 2,000 at this time, down significantly from its roughly 5,000 peak a year prior, but still not at the bottom it would reach in August 2001.
But we went public because it was the only viable source of capital available to LoudCloud. We desperately needed the additional funding to continue to run the business. Despite having raised a lot of money to date, we were dangerously low on cash due to the post-2000 dot-com collapse. This was because we had originally targeted our service offerings to other startup companies; they seemed like a natural customer base given that they could benefit from being able to pay LoudCloud to worry about their computing infrastructure while they focused their activities on the internal development of their custom applications.
For us to provide this service, however, we had to procure significant amounts of data center space and a ton of computer equipment. We paid for this infrastructure up front with the idea that we would amortize the payback of these costs as we grew our customer base. That worked for the first year or so until the cascading effects of the air being let out of the dot-com balloon caught up to us. As a result, our dot-com customers started going out of business and naturally had no VCs willing to fund their ongoing operations. We were stuck with a very high fixed cost base of capital infrastructure against a diminishing base of customers—a recipe for significant cash consumption.
And, as noted above, by this time, the VCs had essentially stopped writing checks, so the only other option for us was to raise money from more buyout-oriented investors. Buyout investors are different from VC firms in a few ways. Namely, they tend to invest in companies that are beyond pure startup stage, and they generally make what are called “control” investments. Control means that they often own a majority of the company and control a majority of the seats on the board of directors; this gives them the ability to be the major determiners of the company’s strategy. Buyout capital can often be more expensive than VC because the upside opportunity for these investors is more constrained given the later stage at which they invest. This was the case for us, meaning that the valuation at which they would fund the company was much lower, and thus the amount of ownership we would have to give up was much higher. In addition, the control aspects of the buyout alternatives we had were less palatable than our desire to preserve more degrees of freedom in running the business.
In an odd way, therefore, going public seemed to provide the lowest available cost of capital and the apparent path of least resistance. We originally intended to sell shares to the public at a range of ten to twelve dollars per share. (When companies file to go public, they put what’s known as an “initial filing range” out to the market to signal the price range at which they hope to sell shares to the public. IPOs that are in demand are often oversubscribed, meaning there is more institutional demand to purchase shares than there are shares to sell, and naturally in that case the company will increase the filing range accordingly.) But the stock market continued to deteriorate over the course of our IPO marketing period, and we ultimately sold stock to the public at six dollars per share. This is definitely not your typical IPO story. But, the IPO allowed us to raise sufficient capital to give ourselves a shot at success without having to give up day-to-day control of the business.
“Live to fight another day” is another great startup mantra to always keep front and center in your mind. Of course, as John Maynard Keynes reminded us, this applies to almost every financial endeavor: “The markets can remain irrational longer than you can remain solvent.” Cash is undoubtedly king in the startup world—and in the business world more generally.
But perhaps the most poignant phrasing of this lesson that I ever heard came from the late Bill Campbell. Bill is a Silicon Valley legend (Apple, Intuit, GO Corporation, Google, etc.) and in his later years was referred to as “Coach,” for he spent tireless hours coaching entrepreneurs as they were building their businesses. He was also once a “real” coach of the Columbia University football team, but suffice it to say that his coaching record there paled in comparison to his many business successes over a long career. We were privileged to have Bill on our board at LoudCloud, where he constantly reminded us in very simple terms of the critical role that cash plays in a startup’s life cycle: “It’s not about the money. It’s about the F-ing money.” Enough said.
In 2002, we ultimately sold most of the LoudCloud business to Electronic Data Systems (EDS) and essentially restarted as an enterprise software business named Opsware. In addition to being the new name of the company, Opsware was also the name of the software we had developed to use internally when we were running the LoudCloud business—the name was a contraction of “Operations Software.” Because as LoudCloud we had to manage a whole series of servers, network devices, storage devices, and applications, we developed the Opsware software to reduce the amount of manual labor needed by automating a variety of the technology management tasks. When EDS acquired the LoudCloud business, it licensed the Opsware software but allowed us to retain the core intellectual property. So we did what any enterprising startup would do and created a new business selling the Opsware software to other large-enterprise customers who could benefit from automating their own technology management processes.
And we did all this while still being a publicly listed company, albeit with a nascent business and a market cap that appropriately reflected that (im)maturity. The stock hit a low of thirty-four cents, but we stuck with it for another five years and ultimately built a nice software business at Opsware that Hewlett-Packard purchased in 2007 for $1.65 billion. My partner Ben has written extensively about the transformation of the business in his own book, The Hard Thing about Hard Things , which I highly recommend. (And that’s not just because he’s still my boss!)
Immediately following the sale of Opsware to Hewlett-Packard, many of us had the opportunity to stay on as part of the HP Software business. At the time, HP Software was a roughly $4 billion division within the broader HP mother ship (HP sold everything from printers and ink cartridges to desktops, servers, networking equipment, and storage devices) that had been built on the foundations of HP OpenView, a set of software products that, like Opsware, helped companies manage their IT assets.
Over the years, HP Software had acquired a number of other software businesses in the broader IT management space, and thus the product line, employees, and customer base were very diverse and geographically dispersed. I had the opportunity to manage the integration of the Opsware team into HP Software and then to run the roughly $1 billion global software support business. With 1,500 employees scattered across every major global market, I logged more airline miles in that job than I have ever done in my professional lifetime to date. But it was a fun and exciting opportunity to manage a team at scale, as jobs and learning experiences of that kind can be hard to come by in the earlier-stage startup world.
Following the 2007 sale of Opsware to HP, Marc and Ben began investing in earnest as angel investors. Angels are traditionally individuals who invest in very-early-stage startups (generally known as “seed-stage companies”). In Silicon Valley in 2007, the angel community was pretty small, and there were not many institutional seed funds, meaning professional investors who raised money from traditional institutional investors to invest in seed-stage companies. Rather, angel investing was dominated largely by a loose collection of individuals who were writing checks out of their personal accounts. Interestingly, Marc and Ben made their angel investments through an entity known as HA Angel Fund (Horowitz Andreessen Angel Fund), a reversal of the now-well-known brand name for their venture fund.
Marc and Ben started investing at an exciting time when change was afoot in Silicon Valley. To understand this change, you have to understand a bit of the history of the VC industry.
As we’ll dive into deeper in subsequent chapters, the Silicon Valley VC business started in earnest in the 1970s and was characterized for most of the next thirty-odd years by a relatively small number of very successful firms that controlled access to startup capital. In simple terms, capital was the scarce resource, and that resource was “owned” by the then-existing VC firms, many of which are still very successful and active players in the current VC marketplace. Thus those who wanted access to that capital—the entrepreneurs—needed to effectively compete for that capital. The balance of power, therefore, as between the VC firms and entrepreneurs, was squarely in favor of the former.
Beginning in the early 2000s, though, there were a few significant transformations in the startup ecosystem that would change things in the entrepreneurs’ favor.
First, the amount of capital required to start a company began to decline; this continues in earnest even today. Not only did the absolute cost of servers, networking, storage, data center space, and applications begin to fall, but the procurement method evolved from up-front purchasing to much cheaper “renting” with the advent of what is known as cloud computing. As a startup, these changes are very significant, as they mean that the amount of money you need to raise from VCs to get started is much less than in the past.
The second material transformation in the startup ecosystem was the advent of an incubator known as Y Combinator (or YC for short). Started in 2005 by Paul Graham and Jessica Livingston, YC basically created startup school. Cohorts of entrepreneurs joined a “YC batch,” working in an open office space together and going through a series of tutorials and mentorship sessions over a three-month period to see what might come out the other end. Over the past thirteen years, YC has turned out nearly 1,600 promising startups, including some very well-known success stories such as Airbnb, Coinbase, Instacart, Dropbox, and Stripe.
But that’s not the most significant impact that YC has had on the VC ecosystem. Rather, the import of YC, I believe, is that it has educated a whole range of entrepreneurs on the process of starting a company, of which raising capital from VCs is an integral part. That is, YC cracked open the “black box” that was the VC industry, illuminating to entrepreneurs the process of startup company formation and capital raising.
In addition, YC created true communities of entrepreneurs among which they could share their knowledge and views both on company building and on their experiences working with VC firms. Prior to this time, the entrepreneurial community was more dispersed, and therefore knowledge sharing between members of the community was decidedly limited. But with knowledge comes power, thus the second material driver of the changing balance of power between entrepreneurs and VCs.
And that takes us to the founding of Andreessen Horowitz, started in 2009 by Marc Andreessen and Ben Horowitz. What Marc and Ben saw was this fundamental shift in the landscape that would no longer make access to capital alone a sufficient differentiator for VC firms. Rather, in their view, VCs would need to provide something more than simply capital, for that was becoming a commodity, and instead, in this post-2005 era of VC, firms would need to compete for the right to fund entrepreneurs by providing something more.
What that “something more” would be was informed by their thinking around the nature of technology startup ventures. That is, tech startups are basically innovative product or service companies. In most cases, tech startups represent an amalgamation of engineers who identify some innovative way to solve an existing problem or create a new market by introducing a product or service that consumers didn’t even know could exist. This affinity between the identification of the problem to be solved and the development of the product or service that in fact solves the problem is a key component of successful tech startups. No doubt that effective sales and marketing, capital deployment, and team building, among others, are also crucial ingredients to success, but fundamentally tech startups need to “fit” a market problem to a compelling market solution to have a shot at success.
Thus, to increase the odds of ultimately building a widely successful and valuable company, Marc and Ben had a thesis that founders should ultimately be product/engineering types and that there should be a tight coupling between the product visionary and the individual responsible for driving the company’s strategy and resource allocation decisions. Those latter responsibilities are typically the province of the CEO. Therefore, Marc and Ben had a predilection for backing CEOs who were also the source of the company’s product vision.
But, while technical founding CEOs might be great at product development, they might often lack the rest of the skills and relationships required to be all-around great CEOs—technical recruiting, executive recruiting, PR and marketing, sales and business development, corporate development, and regulatory affairs, among others.
As a result, the “something more” that Marc and Ben decided to build Andreessen Horowitz around was a network of people and institutions that could improve the prospects for founding product CEOs to become world-class CEOs. And I was lucky enough to become employee number one as we launched the firm in June 2009.
Over the past ten years, we’ve gone from $300 million in funds under management and a three-person team to managing more than $7 billion in funds and roughly 150 employees. Most of our employees focus on that “something more,” spending their days building relationships with people and institutions that can help improve the likelihood of our founder CEOs building enduring and valuable companies.
We’ve been fortunate enough to invest in many great companies, some of which are household names today—Airbnb, Pinterest, Instacart, Oculus, Slack, GitHub—and many that we hope will become household names in the future. And we’ve learned a lot, sometimes by making the right decisions but also by making mistakes as we’ve built the business. We believe in being innovative and experimenting in our own business. In fact, we consistently tell our team to “make new mistakes,” which we hope translates into taking informed risks, iterating on product and service offerings, and learning from previous mistakes to avoid treading down the same dead-end path. Throughout the course of this book, we’ll spend more time on many of the lessons learned.
Most importantly, we believe deeply in the sanctity of the entrepreneurial process and work hard every day to respect the very difficult journey that aspiring entrepreneurs walk on their hopeful path to success. We know that the odds of success for most entrepreneurial endeavors are small and that the ones that make it do so due to a unique combination of vision, inspiration, grit, and a healthy dose of luck.
It is their stories, and those of LoudCloud, Opsware, and Andreessen Horowitz, that are in many ways the story of this book.
Startups thrive (or die) based on the availability of capital from VCs, particularly at the formative stages of their lives when the business itself is in growth mode and can’t support itself through operating cash flow. And VC, like all types of capital, is a great form of financing where the needs and desires of the entrepreneur and the VC are aligned; there is a mutual pact the two organizations enter into with an agreed-upon set of objectives they hope to accomplish together. Money from public, institutional investors can also be an important part of the financing equation as a startup gets to a later point of maturity and can then satisfy the demands for predictable earnings growth that such investors require.
In a similar vein, when interests diverge between entrepreneurs and VCs, the world is not a very fun place to be.
As I’ve already mentioned, the best way to set up a successful marriage between entrepreneurs and VCs is to level the playing field and make sure everyone understands how VC works. So now it’s time to roll up our sleeves and dig in.