Despite all the attention and investment that Silicon Valley’s recent start-ups have received, they have done little but lose money: Uber, Lyft, WeWork, Pinterest, and Snapchat have consistently failed to turn profits, with Uber’s cumulative losses exceeding $25 billion. Perhaps even more notorious are bankrupt and discredited start-ups such as Theranos, Luckin Coffee, and Wirecard, which were plagued with management failures, technical problems, or even outright fraud that auditors failed to notice.Below the fold, some reflections on Funk's insightful analysis of the "larger, more systemic problem".
What’s going on? There is no immediately obvious reason why this generation of start-ups should be so financially disastrous. After all, Amazon incurred losses for many years, but eventually grew to become one of the most profitable companies in the world, even as Enron and WorldCom were mired in accounting scandals. So why can’t today’s start-ups also succeed? Are they exceptions, or part of a larger, more systemic problem?
Funk introduces his argument thus:
In this article, I first discuss the abundant evidence for low returns on VC investments in the contemporary market. Second, I summarize the performance of start-ups founded twenty to fifty years ago, in an era when most start-ups quickly became profitable, and the most successful ones rapidly achieved top-100 market capitalization. Third, I contrast these earlier, more successful start-ups with Silicon Valley’s current set of “unicorns,” the most successful of today’s start-ups. Fourth, I discuss why today’s start-ups are doing worse than those of previous generations and explore the reasons why technological innovation has slowed in recent years. Fifth, I offer some brief proposals about what can be done to fix our broken start-up system. Systemic problems will require systemic solutions, and thus major changes are needed not just on the part of venture capitalists but also in our universities and business schools.
Is There A Problem?Funk's argument that there is a problem can be summarized thus:
- The returns on VC investments over the last two decades haven't matched the golden years of the proceeding two decades.
- In the golden years startups made profits.
- Now they don't.
VC Returns Are Sub-Par
a small percentage of investments does provide high returns, and these high returns for top-performing VC funds persist over subsequent quarters. Although this data does not demonstrate that select VCs consistently earn solid profits over decades, it does suggest that these VCs are achieving good returns.It was always true that VC quality varied greatly. I discussed the advantages of working with great VCs in Kai Li's FAST Keynote:
Work with the best VC funds. The difference between the best and the merely good in VCs is at least as big as the difference between the best and the merely good programmers. At nVIDIA we had two of the very best, Sutter Hill and Sequoia. The result is that, like Kai but unlike many entrepreneurs, we think VCs are enormously helpful.One thing that was striking about working with Sutter Hill was how many entrepreneurs did a series of companies with them, showing that both sides had positive experiences.
Startups Used To Make ProfitsBefore the dot-com boom, there used to be a rule that in order to IPO a company, it had to be making profits. This was a good rule, since it provided at least some basis for setting the stock price at the IPO. Funk writes:
There was a time when venture capital generated big returns for investors, employees, and customers alike, both because more start-ups were profitable at an earlier stage and because some start-ups achieved high market capitalization relatively quickly. Profits are an important indicator of economic and technological growth, because they signal that a company is providing more value to its customers than the costs it is incurring.Funk's Table 2 shows the years to profitability and years to top-100 market capitalization for companies founded between 1975 and 2004. I'm a bit skeptical of the details because, for example, the table says it took Sun Microsystems 6 years to turn a profit. I'm pretty sure Sun was profitable at its 1986 IPO, 4 years from its founding.
A number of start-ups founded in the late twentieth century have had an enormous impact on the global economy, quickly reaching both profitability and top-100 market capitalization. Among these are the so-called FAANMG (Facebook, Amazon, Apple, Microsoft, Netflix, and Google), which represented more than 25 percent of the S&P’s total market capitalization and more than 80 percent of the 2020 increase in the S&P’s total value at one point—in other words, the most valuable and fastest-growing companies in America in recent years.
Note Funk's stress on achieving profitability quickly. An important Silicon Valley philosophy used to be:
- Success is great!
- Failure is OK.
- Not doing either is a big problem.
Unicorns, Not So MuchWhat are these unicorns? Wikipedia tells us:
In business, a unicorn is a privately held startup company valued at over $1 billion. The term was coined in 2013 by venture capitalist Aileen Lee, choosing the mythical animal to represent the statistical rarity of such successful ventures.Back in 2013 unicorns were indeed rare, but as Wikipedia goes on to point out:
According to CB Insights, there are over 450 unicorns as of October 2020.Unicorns are breeding like rabbits, but the picture Funk paints is depressing:
In the contemporary start-up economy, “unicorns” are purportedly “disrupting” almost every industry from transportation to real estate, with new business software, mobile apps, consumer hardware, internet services, biotech, and AI products and services. But the actual performance of these unicorns both before and after the VC exit stage contrasts sharply with the financial successes of the previous generation of start-ups, and suggests that they are dramatically overvalued.Hey, they're startups, right? They just need time to become profitable. Funk debunks that idea too:
Figure 3 shows the profitability distribution of seventy-three unicorns and ex-unicorns that were founded after 2013 and have released net income and revenue figures for 2019 and/or 2020. In 2019, only six of the seventy-three unicorns included in figure 3 were profitable, while for 2020, seven of seventy were.
Furthermore, there seems to be little reason to believe that these unprofitable unicorn start-ups will ever be able to grow out of their losses, as can be seen in the ratio of losses to revenues in 2019 versus the founding year. Aside from a tiny number of statistical outliers ... there seems to be little relationship between the time since a start-up’s founding and its ratio of losses to revenues. In other words, age is not correlated with profits for this cohort.Funk goes on to note that startup profitability once public has declined dramatically, and appears inversely related to IPO valuation:
When compared with profitability data from decades past, recent start-ups look even worse than already noted. About 10 percent of the unicorn start-ups included in figure 3 were profitable, much lower than the 80 percent of start-ups founded in the 1980s that were profitable, according to Jay Ritter’s analysis, and also below the overall percentage for start-ups today (20 percent). Thus, not only has profitability dramatically dropped over the last forty years among those start-ups that went public, but today’s most valuable start-ups—those valued at $1 billion or more before IPO—are in fact less profitable than start-ups that did not reach such lofty pre-IPO valuations.Funk uses electric vehicles and biotech to illustrate startup over-valuation:
For instance, driven by easy money and the rapid rise of Tesla’s stock, a group of electric vehicle and battery suppliers—Canoo, Fisker Automotive, Hyliion, Lordstown Motors, Nikola, and QuantumScape—were valued, combined, at more than $100 billion at their listing. Likewise, dozens of biotech firms have also achieved billions of dollars in market capitalizations at their listings. In total, 2020 set a new record for the number of companies going public with little to no revenue, easily eclipsing the height of the dot-com boom of telecom companies in 2000.The Alphaville team have been maintaining a spreadsheet of the EV bubble. They determined that there was no way these companies' valuations could be justified given the size of the potential market. Jamie Powell's April 12th Revisiting the EV bubble spreadsheet celebrates their assessment:
At pixel time the losses from their respective peaks from all of the electric vehicle, battery and charging companies on our list total some $635bn of market capitalisation, or a fall of just under 38 per cent. Ouch.
What Is Causing The ProblemThis all looks like too much money chasing too few viable startups, and too many me-too startups chasing too few total available market dollars.
Funk starts his analysis of the causes of poor VC returns by pointing to the obvious one, one that applies to any successful investment strategy. Its returns will be eroded over time by the influx of too much money:
There are many reasons for both the lower profitability of start-ups and the lower returns for VC funds since the mid to late 1990s. The most straightforward of these is simply diminishing returns: as the amount of VC investment in the start-up market has increased, a larger proportion of this funding has necessarily gone to weaker opportunities, and thus the average profitability of these investments has declined.But the effect of too much money is even more corrosive. I'm a big believer in Bill Joy's Law of Startups — "success is inversely proportional to the amount of money you have". Too much money allows hard decisions to be put off. Taking hard decisions promptly is key to "fail fast".
Nvidia was an example of this. The company was founded in one of Silicon Valley's recurring downturns. We were the only hardware company funded in that quarter. We got to working silicon on a $2.5M A round. Think about it — each of our VCs invested $1.25M to start a company currently valued at $380,000M. Despite delivering ground-breaking performance, as I discussed in Hardware I/O Virtualization, that chip wasn't a success. But it did allow Jen-Hsun Huang to raise another $6.5M. He down-sized the company by 2/3 and got to working silicon of the highly successful second chip with, IIRC, six weeks' money left in the bank.
Funk then discusses a second major reason for poor performance:
A more plausible explanation for the relative lack of start-up successes in recent years is that new start-ups tend to be acquired by large incumbents such as the faamng companies before they have a chance to achieve top 100 market capitalization. For instance, YouTube was founded in 2004 and Instagram in 2010; some claim they would be valued at more than $150 billion each (pre-lockdown estimates) if they were independent companies, but instead they were acquired by Google and Facebook, respectively.18 In this sense, they are typical of the recent trend: many start-ups founded since 2000 were subsequently acquired by faamng, including new social media companies such as GitHub, LinkedIn, and WhatsApp. Likewise, a number of money-losing start-ups have been acquired in recent years, most notably DeepMind and Nest, which were bought by Google.But he fails to note the cause of the rash of acquisitions, which is clearly the total Lack Of Anti-Trust Enforcement in the US. As with too much money, the effects of this lack are more pernicious than at first appears. Again, Nvidia provides an example.
Just like the founders and VCs of Sun, when we started Nvidia we knew that the route to an IPO and major return on investment involved years and several generations of product. So, despite the limited funding and with the full support of our VCs, we took several critical months right at the start to design an architecture for a family of successive chip generations based on Hardware I/O Virtualization. By ensuring that the drivers in application software interacted only with virtual I/O resources, the architecture decoupled the hardware and software release cycles. The strong linkage between them at Sun had been a consistent source of schedule slip.
The architecture also structured the implementation of the chip as a set of modules communicating via an on-chip network. Each module was small enough that a three-person team could design, simulate and verify it. The restricted interface to the on-chip network meant that, if the modules verified correctly, it was highly likely that the assembled chip would verify correctly.
Laying the foundations for a long-term product line in this way paid massive dividends. After the second chip, Nvidia was able to deliver a new chip generation every 6 months like clockwork. 6 months after we started Nvidia, we knew over 30 other startups addressing the same market. Only one, ATI, survived the competition with Nvidia's 6-month product cycle.
VCs now would be hard to persuade that the return on the initial time and money to build a company that could IPO years later would be worth it when compared to lashing together a prototype and using it to sell the company to one of the FAANMGs. In many cases, simply recruiting a team that could credibly promise to build the prototype would be enough for an "aqui-hire", where a FAANMG buys a startup not for the product but for the people. Building the foundation for a company that can IPO and make it into the top-100 market cap list is no longer worth the candle.
But Funk argues that the major cause of lower returns is this:
Overall, the most significant problem for today’s start-ups is that there have been few if any new technologies to exploit. The internet, which was a breakthrough technology thirty years ago, has matured. As a result, many of today’s start-up unicorns are comparatively low-tech, even with the advent of the smartphone—perhaps the biggest technological breakthrough of the twenty-first century—fourteen years ago. Ridesharing and food delivery use the same vehicles, drivers, and roads as previous taxi and delivery services; the only major change is the replacement of dispatchers with smartphones. Online sales of juicers, furniture, mattresses, and exercise bikes may have been revolutionary twenty years ago, but they are sold in the same way that Amazon currently sells almost everything. New business software operates from the cloud rather than onsite computers, but pre-2000 start-ups such as Amazon, Google, and Oracle were already pursuing cloud computing before most of the unicorns were founded.Remember, Sun's slogan in the mid 80s was "The network is the computer"!
|Virtua Fighter on NV1|
- An experienced, high-quality team. Initial teams at startups are usually recruited from colleagues, so they are used to working together and know each other's strengths and weaknesses. Jen-Hsun Huang was well-known at Sun, having been the application engineer for LSI Logic on Sun's first SPARC implementation. The rest of the initial team at Nvidia had all worked together building graphics chips at Sun. As the company grows it can no longer recruit only colleagues, so usually experiences what at Sun was called the "bozo invasion".
- Freedom from backwards compatibility constraints. Radical design change is usually needed to take advantage of a technological discontinuity. Reconciling this with backwards compatibility takes time and forces compromise. Nvidia was able to ignore the legacy of program I/O from the ISA bus and fully exploit the Direct Memory Access capability of the PCI bus from the start.
- No cash cow to defend. The IBM-funded Andrew project at CMU was intended to deploy what became the IBM PC/RT, which used the ROMP, an IBM RISC CPU competing with Sun's SPARC. The ROMP was so fast that IBM's other product lines saw it as a threat, and insisted that it be priced not to under-cut their existing product's price/performance. So when it finally launched, its price/performance was much worse than Sun's SPARC-based products, and it failed.
In short, today’s start-ups have targeted low-tech, highly regulated industries with a business strategy that is ultimately self-defeating: raising capital to subsidize rapid growth and securing a competitive position in the market by undercharging consumers. This strategy has locked start-ups into early designs and customer pools and prevented the experimentation that is vital to all start-ups, including today’s unicorns. Uber, Lyft, DoorDash, and GrubHub are just a few of the well-known start-ups that have pursued this strategy, one that is used by almost every start-up today, partly in response to the demands of VC investors. It is also highly likely that without the steady influx of capital that subsidizes below-market prices, demand for these start-ups’ services would plummet, and thus their chances of profitability would fall even further. In retrospect, it would have been better if start-ups had taken more time to find good, high-tech business opportunities, had worked with regulators to define appropriate behavior, and had experimented with various technologies, designs, and markets, making a profit along the way.But, if the key to startup success is exploiting a technological discontinuity, and there haven't been any to exploit, as Funk argues earlier, taking more time to "find good, high-tech business opportunities" wouldn't have helped. They weren't there to be found.
How To Fix The Problem?Funk quotes Charles Duhigg skewering the out-dated view of VCs:
For decades, venture capitalists have succeeded in defining themselves as judicious meritocrats who direct money to those who will use it best. But examples like WeWork make it harder to believe that V.C.s help balance greedy impulses with enlightened innovation. Rather, V.C.s seem to embody the cynical shape of modern capitalism, which too often rewards crafty middlemen and bombastic charlatans rather than hardworking employees and creative businesspeople.And:
Venture capitalists have shown themselves to be far less capable of commercializing breakthrough technologies than they once were. Instead, as recently outlined in the New Yorker, they often seem to be superficial trend-chasers, all going after the same ideas and often the same entrepreneurs. One managing partner at SoftBank summarized the problem faced by VC firms in a marketplace full of copycat start-ups: “Once Uber is founded, within a year you suddenly have three hundred copycats. The only way to protect your company is to get big fast by investing hundreds of millions.”VCs like these cannot create the technological discontinuities that are the key to adequate returns on investment in startups:
we need venture capitalists and start-ups to create new products and new businesses that have higher productivity than do existing firms; the increased revenue that follows will then enable these start-ups to pay higher wages. The large productivity advantages needed can only be achieved by developing breakthrough technologies, like the integrated circuits, lasers, magnetic storage, and fiber optics of previous eras. And different players—VCs, start-ups, incumbents, universities—will need to play different roles in each industry. Unfortunately, none of these players is currently doing the jobs required for our start-up economy to function properly.
Business SchoolsSuccess in exploiting a technological discontinuity requires understanding of, and experience with, the technology, its advantages and its limitations. But Funk points out that business schools, not being engineering schools, need to devalue this requirement. Instead, they focus on "entrepreneurship":
In recent decades, business schools have dramatically increased the number of entrepreneurship programs—from about sixteen in 1970 to more than two thousand in 2014—and have often marketed these programs with vacuous hype about “entrepreneurship” and “technology.” A recent Stanford research paper argues that such hype about entrepreneurship has encouraged students to become entrepreneurs for the wrong reasons and without proper preparation, with universities often presenting entrepreneurship as a fun and cool lifestyle that will enable them to meet new people and do interesting things, while ignoring the reality of hard and demanding work necessary for success.One of my abiding memories of Nvidia is Tench Coxe, our partner at Sutter Hill, perched on a stool in the lab playing the "Road Rash" video game about 2am one morning as we tried to figure out why our first silicon wasn't working. He was keeping an eye on his investment, and providing a much-needed calming influence.
Focus on entrepreneurship means focus on the startup's business model not on its technology:
A big mistake business schools make is their unwavering focus on business model over technology, thus deflecting any probing questions students and managers might have about what role technological breakthroughs play and why so few are being commercialized. For business schools, the heart of a business model is its ability to capture value, not the more important ability to create value. This prioritization of value capture is tied to an almost exclusive focus on revenue: whether revenues come from product sales, advertising, subscriptions, or referrals, and how to obtain these revenues from multiple customers on platforms. Value creation, however, is dependent on technological improvement, and the largest creation of value comes from breakthrough technologies such as the automobile, microprocessor, personal computer, and internet commerce.The key to "capturing value" is extracting value via monopoly rents. The way to get monopoly rents is to subsidize customer acquisition and buy up competitors, until the customers have no place to go. This doesn't create any value. In fact once the monopolist has burnt through the investor's money they find they need a return that can only be obtained by raising prices and holding the customer to ransom, destroying value for everyone.
It is true a startup that combines innovation in technology with innovation in business has an advantage. Once more, Nvidia provides an example. Before starting Nvidia, Jen-Hsun Huang had run a division of LSI Logic that traded access to LSI Logic's fab for equity in the chips it made. Based on this experience on the supplier side of the fabless semiconductor business, one of his goals for Nvidia was to re-structure the relationship between the fabless company and the fab to be more of a win-win. Nvidia ended up as one of the most successful fabless companies of all time. But note that the innovation didn't affect Nvidia's basic business model — contract with fabs to build GPUs, and sell them to PC and graphics board companies. A business innovation combined with technological innovation stands a chance of creating a big company; a business innovation with no technology counterpart is unlikely to.
ResearchFunk assigns much blame for the lack of breakthrough technologies to Universities:
University engineering and science programs are also failing us, because they are not creating the breakthrough technologies that America and its start-ups need. Although some breakthrough technologies are assembled from existing components and thus are more the responsibility of private companies—for instance, the iPhone—universities must take responsibility for science-based technologies that depend on basic research, technologies that were once more common than they are now.Note that Funk accepts as a fait accompli the demise of corporate research labs, which certainly used to do the basic research that led not just to Funk's examples of "semiconductors, lasers, LEDs, glass fiber, and fiber optics", but also, for example, to packet switching, and operating systems such as Unix. As I did three years ago in Falling Research Productivity, he points out that increased government and corporate funding of University research has resulted in decreased output of breakthrough technologies:
Many scientists point to the nature of the contemporary university research system, which began to emerge over half a century ago, as the problem. They argue that the major breakthroughs of the early and mid-twentieth century, such as the discovery of the DNA double helix, are no longer possible in today’s bureaucratic, grant-writing, administration-burdened university. ... Scientific merit is measured by citation counts and not by ideas or by the products and services that come from those ideas. Thus, labs must push papers through their research factories to secure funding, and issues of scientific curiosity, downstream products and services, and beneficial contributions to society are lost.Funk's analysis of the problem is insightful, but I see his ideas for fixing University research as simplistic and impractical:
A first step toward fixing our sclerotic university research system is to change the way we do basic and applied research in order to place more emphasis on projects that may be riskier but also have the potential for greater breakthroughs. We can change the way proposals are reviewed and evaluated. We can provide incentives to universities that will encourage them to found more companies or to do more work with companies.Funk clearly doesn't understand how much University research is already funded by companies, and how long attempts to change the reward system in Universities have been crashing into the rock comprised of senior faculty who achieved their position through the existing system.
He is more enthusiastic but equally misled about how basic research in corporate labs could be revived:
One option is to recreate the system that existed prior to the 1970s, when most basic research was done by companies rather than universities. This was the system that gave us transistors, lasers, LEDs, magnetic storage, nuclear power, radar, jet engines, and polymers during the 1940s and 1950s. ... Unlike their predecessors at Bell Labs, IBM, GE, Motorola, DuPont, and Monsanto seventy years ago, top university scientists are more administrators than scientists now—one of the greatest misuses of talent the world has ever seen. Corporate labs have smaller administrative workloads because funding and promotion depend on informal discussions among scientists and not extensive paperwork.Not understanding the underlying causes of the demise of corporate research labs, Funk reaches for the time-worm nostrums of right-wing economists, "tax credits and matching grants":
We can return basic research to corporate labs by providing much stronger incentives for companies—or cooperative alliances of companies—to do basic research. A scheme of substantial tax credits and matching grants, for instance, would incentivize corporations to do more research and would bypass the bureaucracy-laden federal grant process. This would push the management of detailed technological choices onto scientists and engineers, and promote the kind of informal discussions that used to drive decisions about technological research in the heyday of the early twentieth century. The challenge will be to ensure these matching funds and tax credits are in fact used for basic research and not for product development. Requiring multiple companies to share research facilities might be one way to avoid this danger, but more research on this issue is needed.In last year's The Death Of Corporate Research Labs I discussed a really important paper from a year earlier by Arora et al, The changing structure of American innovation: Some cautionary remarks for economic growth, which Funk does not cite. I wrote:
Arora et al point out that the rise and fall of the labs coincided with the rise and fall of anti-trust enforcement:It is pretty clear that "tax credits and matching grants" aren't the fix for the fundamental anti-trust problem. Not to mention that the idea of "Requiring multiple companies to share research facilities" in and of itself raises serious ant-trust concerns. After such a good analysis, it is disappointing that Funk's recommendations are so feeble.
Historically, many large labs were set up partly because antitrust pressures constrained large firms’ ability to grow through mergers and acquisitions. In the 1930s, if a leading firm wanted to grow, it needed to develop new markets. With growth through mergers and acquisitions constrained by anti-trust pressures, and with little on offer from universities and independent inventors, it often had no choice but to invest in internal R&D. The more relaxed antitrust environment in the 1980s, however, changed this status quo. Growth through acquisitions became a more viable alternative to internal research, and hence the need to invest in internal research was reduced.Lack of anti-trust enforcement, pervasive short-termism, driven by Wall Street's focus on quarterly results, and management's focus on manipulating the stock price to maximize the value of their options killed the labs:
Large corporate labs, however, are unlikely to regain the importance they once enjoyed. Research in corporations is difficult to manage profitably. Research projects have long horizons and few intermediate milestones that are meaningful to non-experts. As a result, research inside companies can only survive if insulated from the short-term performance requirements of business divisions. However, insulating research from business also has perils. Managers, haunted by the spectre of Xerox PARC and DuPont’s “Purity Hall”, fear creating research organizations disconnected from the main business of the company. Walking this tightrope has been extremely difficult. Greater product market competition, shorter technology life cycles, and more demanding investors have added to this challenge. Companies have increasingly concluded that they can do better by sourcing knowledge from outside, rather than betting on making game-changing discoveries in-house.
We have to add inadequate VC returns and a lack of startups capable of building top-100 companies to the long list of problems that only a major overhaul of anti-trust enforcement can fix. Lina Khan's nomination to the FTC is a hopeful sign that the Biden adminstration understands the urgency of changing direction, but Biden's hesitation about nominating the DOJ's anti-trust chief is not.
Update: Michael Cembalest's Food Fight: An update on private equity performance vs public equity markets has a lot of fascinating information about private equity in general and venture capital in particular. His graphs comparing MOIC (Multiple Of Invested Capital) and IRR (Internal Rate of Return) across vintage years support his argument that:
We have performance data for venture capital starting in the mid-1990s, but the period is so distorted by the late 1990’s boom and bust that we start our VC performance discussion in 20045. In my view, the massive gains earned by VC managers in the mid-1990s are not relevant to a discussion of VC investing today. As with buyout managers, VC manager MOIC and IRR also tracked each other until 2012 after which a combination of subscription lines and faster distributions led to rising IRRs despite falling MOICs. There’s a larger gap between average and median manager results than in buyout, indicating that there are a few VC managers with much higher returns and/or larger funds that pull up the average relative to the median.
The gap is pretty big:
VC managers have consistently outperformed public equity markets when looking at the “average” manager. But to reiterate, the gap between average and median results are substantial and indicate outsized returns posted by a small number of VC managers. For vintage years 2004 to 2008, the median VC manager actually underperformed the S&P 500 pretty substantially.Another of Cembalest's fascinating graphs addresses this question:
One of the other “food fight” debates relates to pricing of venture-backed companies that go public. In other words, do venture investors reap the majority of the benefits, leaving public market equity investors “holding the bag”? Actually, the reverse has been true over the last decade when measured in terms of total dollars of value creation accruing to pre- and post-IPO investors: post-IPO investor gains have often been substantial.
To show this:
We analyzed all US tech, internet retailing and interactive media IPOs from 2010 to 2019. We computed the total value created since each company’s founding, from original paid-in capital by VCs to its latest market capitalization. We then examined how total value creation has accrued to pre- and post-IPO investors6. Sometimes both investor types share the gains, and sometimes one type accrues the vast majority of the gains. Pre-IPO investors earn the majority of the pie when IPOs collapse or flat-line after being issued, and post-IPO investors reap the majority of the pie when IPOs appreciate substantially after being issued.
There are three general regions in the chart. As you can see, the vast majority of the 165 IPOs analyzed resulted in a large share of the total value creation accruing to public market equity investors; nevertheless, there were some painful exceptions (see lower left region on the chart).