Secrets of Sand Hill Road

by Scott Kupor

  How Can We Measure Success for a Venture Capital Firm?

  What are the implications of all this for investing in venture capital?

  First, diversification is a bad strategy for investing in VC firms. If you are an institutional investor who is lucky enough to have built a roster of successful firms whose returns are not the median but in the high-return section of the power-law curve, you don’t want to diversify. Returns in the top end of VC funds can often be as much as 3,000 basis points higher than at the bottom end; dispersion of returns is huge when you have power-law distributions. In general, diversification is likely to push you toward the median/low-return section of the power-law curve and thus be dilutive to overall returns. Thus, many institutional investors seek to have a concentrated venture portfolio—which, by the way, probably further exacerbates that power-law distribution of returns.

  And that brings us to the second implication—it’s very hard for new firms to break into the industry and be successful. Admittedly, that has changed a bit over the last decade—in part due to the changing nature of the financing environment, which we’ll cover later—but it’s still pretty tough. To become a top VC firm that institutional investors want to invest in, you have to get yourself into that good part of the power-law curve, but if you don’t have the brand to create the positive signaling that attracts the best entrepreneurs, it’s hard to generate the returns—you get the picture. It’s a classic chicken-and-egg problem.

  Forget the Batting Average

  But here’s the rub. VCs with the highest batting averages do not usually make the best VCs.

  There’s nothing like baseball to help us understand the venture business. For the baseball averse, let’s start with defining “batting average.” A player’s batting average is the quotient obtained by dividing the number of hits a player gets by the number of times at bat the player has (I know that walks don’t count in the denominator, but it doesn’t matter for my analogy). So a player who has a .300 batting average—which over the course of a career will get him into the Hall of Fame—gets a hit three out of every ten times that he comes to the plate.

  Good VCs get a hit about five times out of ten at bats (a .500 batting average). A VC “hit” means that the investment returns more than the original amount the VC invested in a company. At first blush, that may sound good. But it’s not—and, more importantly, it doesn’t really matter in determining success.

  For most VCs, the distribution of at bats looks something like this:

  50 percent of the investments are “impaired,” which is a very polite way of saying they lose some or all of their investment. Think about that for a second—VCs are completely wrong about half the time and lose most or all of the money that their investors entrusted to them as a result. In probably any other professional endeavor (baseball maybe being the exception that proves the rule), if you scored 50 percent, you’d likely be looking for a new job. But, hey, we VCs celebrate failure—sort of.

  20–30 percent of the investments are—to continue with the baseball analogy—“singles” or “doubles.” You didn’t lose all the money (congratulations on that), but instead you made a return of a few times your investment. That $5 million you invested in Cryptocurrency.com returned you $10–$20 million—not bad. However, if you include the 50 percent of the “impaired” investments, the VC is still in trouble—70–80 percent of its invested dollars have yielded a total return of about seventy-five to ninety cents on the dollar. That doesn’t sound like a recipe for success.

  Luckily, we still have 10–20 percent of our investments left—and these are our home runs. These are the investments where the VC is expecting to return ten to one hundred times her money.

  If you’re paying attention, this distribution of returns should remind you of the power-law curve discussion from the last section. It turns out that not only does the performance of VC firms follow the power-law curve, but so does the distribution of deals within a given fund.

  Over time, funds that generate two and a half to three times net returns to their investors will be in the good portion of the power-law curve distribution and continue to have access to institutional capital. We’ll talk about fees later, but to achieve two and a half to three times net returns (after all fees), VCs probably need to generate three to four times gross returns. That means if a VC has a $100 million fund, she needs to realize a total of $300–$400 million in proceeds from her investments in order to give $250–$300 million in net returns to the institutional investors.

  What this tells us is that a batting average isn’t the right metric for measuring success for a venture capital firm. In fact, the data show that firms with a better batting average often don’t outperform those with a lower one—how can that be?

  Because what matters most is your “at bats per home run.” In baseball, at bats per home run is the quotient obtained from dividing the number of times a player comes to bat by the total number of home runs achieved. Mark McGwire has the top stat here, with a lifetime at bats per home run of 10.61. That means that McGwire hit a home run roughly one out of every ten times that he came up to the plate.

  In VC, all we really care about is the at bats per home run. That is, the frequency with which the VC gets a return of more than ten times her investment—which we consider a home run. If you do the math, you’ll see that VCs can get a lot of things wrong. Their overall batting average can be even less than 50 percent, as long as their at bats per home run are 10–20 percent, better than the all-time best baseball players.

  And, as noted above, that’s in fact what we see in the industry. The difference between a top-performing venture fund and a poor-performing one is not the batting average, but rather the at bats per home run. In many cases, in fact, firms that do the best have a worse batting average than the firms that underperform: they’re like a baseball slugger who either strikes out or hits a home run every time he comes to the plate. It turns out that you can’t de-risk your way to a winning venture fund. If you want to be in this business, you have to have either a stomach made of steel or a lifetime supply of Maalox.

  Accel Partners is famous for, among other things, investing in a very early round of Facebook. At the time, Facebook was valued at roughly $100 million. Assume that Accel held those shares until Facebook went public, which it did at a market capitalization of around $100 billion. (We’re keeping the math here simple for illustrative purposes, so forget for now about whether Accel’s initial investment was diluted by subsequent rounds of financing, and forget about those first few days or weeks of Facebook’s trading in the public markets.)

  Rough math says that Accel made one thousand times its money on that investment—that certainly puts it in the home-run category. What do you think Accel made on the other investments it made as part of that fund? It’s a trick question. The answer is, “It doesn’t matter!” If you make one thousand times your money on one investment, you could be wrong on everything else and still have a top-performing fund, which Accel has done. It turns out that Accel did in fact make other great investments in that same fund, including AdMob, XenSource, and Trulia, among others—a curtain call of epic proportions. But all of that was financial gravy after the power-law return of the Facebook investment.

  The Venture Capital Industry Is Tiny but Punches Well Above Its Weight

  If you live in California, Massachusetts, or New York and are part of the VC or technology startup ecosystem, you can’t open your Twitter feed or even your local newspaper without being overwhelmed by news of the industry’s happenings. And it might make you think that VC is a really big industry, or at least that the earth revolves around it.

  In fact, venture capital is a really small business, particularly when you compare it with other financial asset classes. The year 2017 was big: investments in companies by VC firms topped $84 billion. That’s the largest amount in a while, and the business bottomed
out (in recent years) at just below $30 billion in 2009. If you look across the prior five years or so, US VC investments in portfolio companies tended to be around $60–$70 billion per year. Interestingly, the number of discrete investments has declined in recent years, as more of the dollars are being concentrated into the companies valued at greater than $1 billion; in 2017, $19 billion (nearly 25 percent of the total capital invested across all companies) went to a very small number of these greater than $1 billion valuation companies. There’s the power law again at work.

  The other size metric for the industry is the annual amount of dollars raised by VC firms from institutional investors. In 2017, US firms raised about $33 billion from investors. At the peak of the dot-com bubble in 2000, US VC firms raised about $100 billion from investors, so we are well off of the peak.

  To give you some perspective on these numbers, the global buyout industry raised about $450 billion in 2017. The hedge fund industry manages north of $3 trillion. The US GDP is about $17 trillion. So, by any measure, the venture capital industry represents a tiny amount of capital at work in the broader financial system.

  But the impact of venture-capital-funded businesses punches well above its weight.

  As we’ve already talked about, the five largest US market capitalization companies are all venture backed—Apple, Facebook, Microsoft, Amazon, and Google.

  Stanford University published a study in 2015 highlighting the concentration of venture-backed companies in the US public markets since 1974. Stanford picked this year because the VC industry dramatically expanded starting in 1979 with the passage of the “prudent man rule.” Prior to 1979, investing in VC was not considered “prudent” for most institutional investors. Thus, the industry largely attracted money from family offices, university endowments, and philanthropic foundations. With the introduction of this rule, pension funds were now permitted to invest in the VC asset class, and thus assets under management grew significantly. Despite the introduction of the rule in 1979, Stanford went back to 1974 to capture one or two significant companies—e.g., Apple—that would have otherwise been missed.

  Using the 1974 data cutoff, 42 percent of public companies are venture backed, representing 63 percent of total market capitalization. These companies account for 35 percent of total employment and 85 percent of total research and development spend.

  That’s pretty good for an industry that invests about 0.4 percent of the US GDP!

  CHAPTER 3

  How Do Early-Stage VCs Decide Where to Invest?

  Let’s take a look at how VCs decide which companies they should invest in and why. In an investing world glamorized by Shark Tank, Silicon Valley, and “unicorns,” there is not a whole lot of simple, straightforward communication about what drives VC decisions.

  As I mentioned, at the early stage of venture investing, raw data is very hard to come by. Obviously! At that point, the company usually hasn’t gone to market yet in any real way. So at the time when many VCs are evaluating a startup for possible investment, qualitative evaluations dwarf quantitative ones.

  As we’ll see later in the book, there are a lot of quantitative ways to model the potential future returns of an investment. These are great spreadsheet exercises when and if you have enough data to inform the assumptions in the spreadsheet.

  But the old adage “garbage in, garbage out” is particularly apt for early-stage venture investing. There simply aren’t enough financial metrics to meaningfully model future potential returns for a business that just doesn’t exist beyond the PowerPoint slides the entrepreneur has put together (sometimes just hours in advance of the pitch meeting with a venture firm).

  So what do you do? Well, it turns out that there are qualitative and high-level quantitative heuristics that VCs use to evaluate the prospects for an investment. And they generally fall into three categories: people, product, and market.

  1. People and Team

  Let’s start with people, as this is by far the most qualitative and, for early-stage investing, likely the most important evaluation criterion. When the “business” is nothing more than a very small collection of individuals—in some cases only one or two founders—with an idea, much of the VCs’ evaluation will focus on the team.

  In particular, many VCs delve deeply into the backgrounds of the founders for clues about their effectiveness in executing this particular idea. The fundamental assumption here is that ideas are not proprietary. In fact, VCs assume the opposite—if an idea turns out to be a good one, assume there will be many other founders and companies that are created to pursue this idea.

  So what matters most is, why do I as a VC want to back this particular team versus any of the x-number of other teams that might show up to execute this idea? The way to think about this is that the opportunity cost of investing in this particular team going after this particular idea is infinite; a decision to invest means that the VC cannot invest in a different team that may come along and ultimately be better equipped to pursue the opportunity.

  John Doerr, a VC from the firm Kleiner Perkins, is famous for purportedly saying that a cardinal rule of venture capital is “No conflict, no interest,” but the reality of modern VC is that conflict is king. Venture capitalists are de facto unable to invest in businesses pursuing the same opportunity, though of course the definition of conflict is always in the eye of the beholder.

  Why is that? Because a VC’s decision to invest in a company is effectively an endorsement of the company as the de facto winner in the space. After all, why would I invest in Facebook instead of Friendster if I felt as though Friendster were the likely company to dominate the social networking market? Recall the earlier discussion about the positive signaling effect of a VC’s brand on the portfolio company’s brand; they are deeply entangled as a result of the investment, for better or for worse. Thus, every investment decision has infinite opportunity cost in that it likely prevents you as a VC from investing in a direct competitor in that space; you have picked your horse to ride.

  In light of this, among the cardinal sins of venture capital is getting the category right (meaning that you correctly anticipated that a big company could be built in a particular space) but getting the company wrong (meaning that you picked the wrong horse to back). For example, you might have discerned in the early 2000s that social networking was going to be big, but then decided to invest in Friendster over Facebook. Or you might have recognized that search was going to be a big business in the late 1990s but elected to invest in AltaReturn over Google.

  So how do you evaluate a founding team? Different VCs of course do things differently, but there are a few common areas of investigation.

  First, what is the unique skill set, background, or experience that led this founding team to pursue this idea? My partners use the concept of a “product-first company” versus a “company-first company.”

  In the product-first company, the founder identified or experienced some particular problem that led her to develop a product to solve that problem, which ultimately compelled her to build a company as the vehicle by which to bring that product to the market. A company-first company is one in which the founder first decides that she wants to start a company and then brainstorms products that might be interesting around which to build one. Successful businesses can of course ultimately be created from either mold, but the product-first company really speaks to the organic nature of company formation. A real-world problem experienced by the founder becomes the inspiration to build a product (and ultimately a company); this organic pull is often very attractive to VCs.

  Many people are undoubtedly familiar with the concept of product-market fit. Popularized by Steve Blank and Eric Ries, product-market fit speaks to a product being so attractive to customers in the marketplace that they recognize the problem it was intended to solve and feel compelled to purchase the product. Consumer “delight” and repeat purchasing are the classic hallmarks of
product-market fit. Airbnb has this, as do Instacart, Pinterest, Lyft, Facebook, and Instagram, among others. As consumers, we almost can’t imagine what we did before these products existed. Again, it is an organic pull on customers, resulting from the breakthrough nature of the product and its fitness to the market problem at which it is directed.

  The equivalent in founder evaluation for VCs is founder-market fit. As a corollary to the product-first company, founder-market fit speaks to the unique characteristics of this founding team to pursue the instant opportunity. Perhaps the founder has a unique educational background best suited to the opportunity.

  We at a16z saw this with Martin Casado and his decision to found Nicira, a company that created software-defined networking (SDN). Martin not only worked on early iterations of SDN for the intelligence community, but he also earned his PhD at Stanford in the area. His entire professional career effectively led him to the development of Nicira, which by the way was ultimately acquired by VMware for $1.25 billion.

  Perhaps the founder had a unique experience that exposed her to the market problem in a way that provided unique insights into the solution for the problem. The founders of Airbnb fit this bill. They were struggling to make ends meet living in San Francisco and noticed that all the hotels were sold out locally whenever there was a major convention in town. What if, they thought, we could rent out a sleeping spot in our apartment to conference attendees to help them save money on accommodations and help us meet our rent obligations? And thus was born Airbnb.

  Perhaps the founder has simply dedicated his life to the particular problem at hand. Orion Hindawi and his father, David, founded a company called BigFix in the late 1990s. BigFix was a security software company that focused on endpoint management—the process by which companies provided virtual security for their PCs, laptops, etc. After selling the company to IBM, Orion and David decided to found Tanium, essentially BigFix 2.0. Incorporating all the lessons learned from BigFix and, as important, the changes in technology infrastructure that occurred over the intervening ten-plus years, Tanium is today a world-class, modern endpoint security company. Tanium represents the culmination of a lifetime of living and breathing enterprise security challenges.