Since the Great Recession, the global private tech ecosystem has witnessed a remarkable paradigm shift. The advent, growth and proliferation of “unicorns” have led to a fundamental change in the way all stakeholders in the tech ecosystem inter-operate. While this unicorn phenomenon has been in the making for the past 6-7 years, ongoing private investing and fundraising trends seem to imply that it could be another decade for the unicorn gold dust to settle. And, another five years for all the stakeholders to fully realize lessons learned from this fundamental paradigm shift.
Our game plan at SharesPost Research is to continuously analyze the drivers and market conditions leading up to the creation & proliferation of unicorns, to debate for and against the “Unicorn Bubble”, and to review fundamental investment trends and issues at well-known unicorns. While doing so, we plan to analyze the current state of tech ecosystem and understand the motivations and biases of all stakeholders, including CEOs/Founders, employees, VCs, and non-traditional private investors. While we have no interest in joining the long list of crystal ball gazers, it is clear to us that an objective, analytical, and data-driven framework is likely to yield valuable insights and lead to spirited debates.
With this backdrop, in our first white paper, we provide a big picture snapshot of unicorns today. We analyzed trends in VC Investments & unicorn creation since 2009. Key takeaways from our analysis are as follows:
Globally 170 Paper Unicorns & Counting
Today, Fortune magazine tracks 174, CB Insights’ identifies 169, and TechCrunch has a list of 168 companies. Approximately 52% or 90 out of the 170-ish Paper Unicorns (i.e., companies closing financings implying a value of $1B or more) are U.S.-based tech companies. This compares to roughly 70 U.S.-based Real Unicorns “minted” since 2009. We use the term "Real Unicorn" to describe a VC-backed tech company with a liquidity event with at least a #1B valuation.
Zero To One Billion in 66 Months
The median duration for a VC-backed company to reach $1B valuation since incorporation is roughly 66 months. U.S.-based tech unicorns tend to reach this status a tad later at 76 months. Non-U.S. tech companies that have become unicorns do so at a median pace of 57 months after incorporation, with typical Chinese Paper Unicorns reaching this milestone in 53 months, or 4.5 years.
VC Batting Average Increased For Unicorns Founded 2005-2009
We estimate roughly 22,000 U.S. tech companies were founded & received VC funding since 1995, resulting in roughly 40 Public, 30 Acquired, and 90 Paper Unicorns. Excluding outlier years with very high (1996 & 2005) and very low (1999 & 2000) batting averages, we estimate a normalized VC batting average ranges between 80-100 at-bats (VC-backed startups) for each unicorn “home run” (or approximately 1.5%). And, recent Paper Unicorn proliferation implies that VC batting average increased 25% year over year for companies founded between 2005 and 2009.
Roughly 30% Of Paper Unicorns Likely To Lose Status
Our analysis involved applying a normalized VC batting average and adjusting for growing investment velocity to the current crop of Paper Unicorns. As a result, we’d expect roughly 30% of the current crop of the roughly 90 U.S. tech Paper Unicorns (and a likely greater proportion of international Paper Unicorns) to have sub $1B liquidity outcome. We’d, however, expect a still significant inflow into the asset class from 2010-2015 cohort of VC-backed companies, given that more than 13,000 U.S. tech companies received VC funding since 2010 versus fewer than 8,000 combined from 1995 through 2009.
Roughly 170 Paper Unicorns & Counting
Renowned tech investor, Aileen Lee coined the term “Unicorns” in November 2013, and estimated 39 such unicorns existed back then. The original unicorn definition was as follows: U.S.-based software or internet companies started since 2003 and valued at over $1 billion by public or private market investors. Since then, the unicorn definition has evolved into the following: VC-backed private tech companies with a private valuation of $1B or higher. Ms. Lee’s original list included 14 private companies (as of November 2013). And, today, about 32 months later, Fortune magazine tracks 174 unicorns globally, CB-Insights has 169 unicorns updated in real-time, and TechCrunch tracks 168 private tech companies with a reported valuation exceeding $1B. TechCrunch also maintains an Emerging Unicorn leaderboard featuring 46 companies with a private valuation between $500MM and $1B. Today’s Paper Unicorns have an aggregate market capitalization of roughly $600B versus roughly $100B in combined market cap of private unicorns as of November 2013. This translates to a 500% increase in market capitalization of late-stage private tech companies in just two and a half years.
In order to learn more about market conditions that led to the creation and proliferation of such unicorns, we analyzed data available on PitchBook, Preqin, & National Venture Capital Association as well as publicly available information from news articles, company websites, CrunchBase, LinkedIn, Wikipedia and public market data on Yahoo! Finance.
Since 2009, CB Insights has tracked 169 private tech companies with valuations exceeding $1B. From 2010 through 2013, there were about eight unicorns minted per year. However, as illustrated below, this trend clearly stepped up in 2014 and 2015. On a 2016 YTD basis, the rate at which unicorns are minted has ticked lower, with roughly 20 net new unicorns minted so far in 2016. Nonetheless, this YTD rate of unicorn creation appears to be roughly in-line with 2014 levels, and clearly a step above the levels observed in 2011 through 2013.
Approx. 90 U.S. Tech Paper Unicorns Minted Since 2009
There has been a gradual shift towards non-U.S.-based Paper Unicorns in the past 24-30 months. At the end of 2013, more than 70% of global Paper Unicorns were companies headquartered in the U.S. During 2014-2016, 44% of Paper Unicorns minted were international companies. Out of the 170-ish Paper Unicorns today, 98 companies or 58% of all Paper Unicorns are U.S.-based and the remaining 72 are international companies. For the purposes of this white paper, we have excluded 11 U.S. Healthcare/Biotech unicorns due to our primary research focus on traditionally defined tech companies. Out of the 72 non-U.S.-based Paper Unicorns, 33 are Chinese companies, followed by 12 in Western Europe (incl. U.K., France, Germany, Sweden), and seven in India.
Median Paper Unicorn Age at Birth: 66 Months
As a next step, we rank-ordered all Paper Unicorns by age. Our objective was to look for signs of a “bubble” (i.e., to determine whether there were any pockets of unusual emergence of Paper Unicorns over the past couple of years). As illustrated in Exhibit 1, we witnessed one Paper Unicorn minted every 9 days in 2014, and one unicorn minted every five days during 2015. Below we have rank ordered global Paper Unicorns by founding year. The oldest Paper Unicorn in the current crop of 170-ish companies was founded in 1995. There are three Paper Unicorns less than two years old (i.e., founded in 2015). During the early days of Paper Unicorn phenomenon, VCs were investing in approximately four to six companies per year that would later become Paper Unicorns. For instance, SurveyMonkey was founded in 1999, and became a Paper Unicorn in 2011 (or 153 months later). Palantir was founded in 2004, and also became a unicorn in 2011 (or about 89 months later).
When we categorized Paper Unicorns by geography, interesting themes began to emerge. We illustrate the general breakdown with respect to U.S.-based and international Paper Unicorns ordered by “founding year” in the chart below. Digging deeper in the data, we came across the following interesting observations:
- During 1995 through 2005, relatively few international tech companies were founded and funded per year that would become unicorns later in their respective lifetimes;
- From a U.S. unicorn perspective, 2006, 2007, and 2009 were probably the best years to incorporate a “unicorn-to-be” as such companies had the highest likelihood to become a unicorn. During each of these three years, more than 10 Paper Unicorns were founded.
- The median age of Paper Unicorns from the cohort before 2014, (i.e., from 2010 to 2013) is roughly 72 months or six years. Median age of the 2014 cohort is 58 months whereas 2015 cohort companies became Paper Unicorns roughly 71 months after incorporation.
- Currently, there are 11 Paper Unicorns on the list that reached $1B valuation level in less than 24 months since incorporation with five being U.S.-based companies.
- During the Paper Unicorn proliferation era of 2014-2015, 115 Paper Unicorns were minted. Roughly 70 out of 115 (or 60%) were U.S.-based companies and 30% were either Chinese or Indian companies.
Finally, we illustrate the overall summary of this analysis in the chart below. The headline is that a typical Paper Unicorn was minted after being in business for roughly 66 months (median duration). U.S.-based unicorns are a tad older, reaching a $1B paper valuation after being in business for about 76 months. On the other hand, international Paper Unicorns have typically reached the $1B valuation level in less than five years. Chinese tech unicorns are among the fastest at 53 months on average, probably hinting at a greater likelihood of a “bubble” in international unicorns, particularly the ones that reached this status during 2015. Finally, in order to provide additional context, the average duration after which a company has become a unicorn is roughly 79 months versus 66 months median duration. This discrepancy in median vs. mean can largely be attributed to few “really old” unicorns (i.e. more than 150 months old). We prefer to rely on the median in the sample set as the distribution around the median is relatively uniform, as more than 70% of unicorns occur within one standard deviation of the median age of 66 months.
Approx. 70 Real U.S. Tech Unicorn Exits Since 2009
As a first step, we decided to compare/contrast this ongoing Paper Unicorn creation with Real Unicorns (i.e. tech companies with $1B+ liquidity outcomes). We gathered data on VC-backed exits with value exceeding $1B within a variety of tech sectors. Because data available for such transactions prior to 2008-09 is quite limited, inaccurate or required us to make a larger number of assumptions than our comfort level permitted, we decided to do this analysis with data starting 2009.
First, we counted global billion-dollar IPOs since January 2009. We estimate there were roughly 100 IPOs globally with company values exceeding $1B at IPO since 2009. About 70% of these unicorn IPOs were U.S.-based companies, or U.S.-domiciled IPOs. 51 out of 100-ish Global IPOs were U.S. tech companies. 39 out of 51 U.S. tech IPOs can be regarded as IPOs of VC-backed companies. For instance, our count of 39 IPOs included companies such as Facebook, Twitter, Tableau, Zynga, Groupon, etc. However, we excluded IPOs such as Alibaba, Nielsen, Gogo, and Level 3 Communications since we believe that they either did not have VC investors or their public offerings were primarily for existing investors to sell shares.
Along similar lines, we counted global M&A transactions with consideration exceeding $1B since 2009. This exercise turned out to be a significantly arduous one. There have been roughly 500 billion-dollar plus M&A transactions since 2009. Approximately 326 of the 500-ish unicorn M&A deals have been with U.S.-based companies. Roughly 120 of these can be regarded as tech sector acquisitions. We believe fewer than 30 of these 120 tech acquisitions can be regarded as “exits of VC-backed companies”. We highlight that we have NOT included acquisitions of public companies that had VC investments prior to going public. Our rationale in doing so is that we wanted to avoid double counting of liquidity events or VC exits. For instance, we have not included Priceline’s acquisition of Kayak since this unicorn liquidity event would be counted under IPO (i.e. Kayak IPO event is a unicorn liquidity outcome event). All in, we haven’t included 23 total acquisitions of such VC-backed but public companies such as LinkedIn, Marketo, HomeAway, Trulia, OpenTable, Active Network, etc. Rest assured, we have included all these companies in our count of IPOs of VC-backed companies.
Below, we illustrate the overall trends in VC-backed U.S.-based tech IPOs and M&A trends since 2009. Key takeaways: 1) There have been six to eight IPOs of VC-backed tech companies every year over the past several years. This annual run rate has largely stayed within this range since the Great Recession; 2) There have been four to five M&As of VC-backed tech companies every year, except 2014.
Finally, we rank ordered all Real Unicorns by their respective founding dates. The overall trend is somewhat similar to Paper Unicorns, as highlighted in Exhibit 4. In particular, the 2005-09 cohort of VCfunded tech companies appear to have had a relatively higher likelihood of a $1B outcome versus prior two cohorts- 1996-2000 or 2001-2005. In addition, we have witnessed just four Real Unicorn outcomes from VC-backed companies founded since 2010. In a normalized scenario, we’d expect at least 5 unicorn exits per year. Or, we would have expected roughly 20 unicorn IPOs or M&A transactions of VC-backed companies founded since 2010. In fact, we have witnessed only four unicorn exits so far. Given the roughly 10 unicorn liquidity outcome events per year of VC-backed companies, we wouldn’t be surprised if another 15 VC-backed companies founded since 2010 have real $1B liquidity outcome events over the next couple of years.
VC Batting Average Increased for Unicorns Founded 2005-2009
According to Major League Baseball’s official definition, batting average means number of hits divided by at-bats. In the context of this research, we were trying to determine an answer to this question: how many startups do VC investors have to back on average to create a $1B+ outcome?
Back in 2009, Union Square Ventures’ Fred Wilson noted that VC’s try to optimize a target batting average around “1/3, 1/3, 1/3”. In other words, 1/3rd of VC investments tend to be money losers, 1/3rd of investments tend to return invested capital (1x return), and bulk of the entire fund’s return comes from remaining 1/3rd of invested companies. With the emergence of unicorns, the size of potential return has clearly been amplified. Effectively, we’d guess that VCs now tend to feel comfortable if a larger proportion of their investments end up being money-losers or return invested capital on a pro rata basis, so long as there is a unicorn outcome among the rest of their investments. Our guess is that VCs try to optimize around a target batting average as “33%, 33%, 31%, and 2%”. In other words, 1/3rd investments are money losers, 1/3rd investments tend to return invested capital (1x), and <1/3rd investments return about 1.5x to 2x capital. But the bulk of the fund’s returns come from 1-2% of VC investments, so called “home runs”.
In order to determine VC batting average over the past 15-20 years, we need two inputs: 1) number of VC-backed companies started in U.S. within tech sector every year (denominator); and 2) number of those companies that became unicorns from those investments (numerator).
Per PitchBook, there are roughly 22,000 to 24,000 such companies (i.e. VC-backed U.S.-based tech companies founded since 1995). For context, CrunchBase lists roughly 32,000 U.S.-based companies founded since 1995 and have at least one funding round from VCs, but this list includes peripheral sectors, and likely a greater margin of error given that CrunchBase is a crowd-sourced database.
Next, we simply added up the number of unicorn IPOs, unicorn M&A transactions, and Paper Unicorns over the past five years. We added up the annual totals illustrated in charts in Exhibit 4 and Exhibit 9. We illustrate this trend in the chart below, and this would become the numerator in our calculation of VC batting average.
Finally, we divided the number of unicorns created in any given year by the number of VC-backed companies in the corresponding year. The headline here is that the likelihood of a VC investment turning into a unicorn has generally stayed within 1.0% and 1.5% on an annual basis over the past 10-15 years. There have been certain periods of troughs and peaks over the past couple of decades, as one would expect.
Key observations include:
Cohort 1: 1996 to 1998
Prior to the onset of the dot-com bust from 1996-1998, VC batting average was consistently above 2.0%. We believe this was the “golden era” of early stage VC investing. Clearly, the pre-dot-com euphoria likely fueled the early stage valuations and directly impacted VC batting average. Additionally, we’d guess there were significantly fewer players in the game 20 years ago. We’d highlight that the number of companies that received VC funding was less than 20% of today’s annual run rate. For instance, from 1996 through 2000, VCs invested in fewer than 1,600 companies combined. Compare that to a range of 1,900 to 3,000 VCbacked companies every year for the past 6 years.
Cohort 2: 1999 to 2001
We noticed that VC batting average of companies founded in 2001 was the lowest observed in the past couple of decades. And, so far, there hasn’t been a single Real Unicorn – either public or acquired – from the 2001 cohort of VC-backed companies.
Cohort 3: 2002 to 2006
During the 5-year period following the dot-com bust, we estimate VC batting averages increased consistently on a year over year basis. During this period, roughly 74 unicorns were founded and funded out of 2,800 total VC-backed companies. This translated into a healthy 1.9% batting average. This includes 18 Public, 10 Acquired, and remaining 32 Paper Unicorns.
Cohort 4: 2007 to 2009 (& beyond)
Around the onset of the Great Recession, the proportion of VC-backed companies that remain Paper Unicorns has gradually increased. For instance, 19 Public or Acquired Unicorns were funded/founded during this 3-year period versus 27 Paper Unicorns. While the annual run-rate of unicorns increased during this period, the overall VC batting average modestly declined to 1.6%, apparently due to the uptick in number of VC-backed companies founded/funded per year.
Roughly 30% Paper Unicorns Likely to Lose Status
Finally, we tried to answer this question – what proportion of Paper Unicorns is likely to have a sub-$1B outcome? In order to estimate this number, we made a simplifying assumption by applying a somewhat conservative batting average (ranging from 1.30% to 1.50%) to the number of VC-backed companies founded/ funded from 2002 to 2009. As a result, we estimate that roughly 20-30 of today’s Paper Unicorns are likely to have a sub-$1B outcome. We illustrate the range of unicorn outcomes for a range of batting averages in the chart below.
Again, there are a lot of caveats to this analysis. For instance, this analysis assumes that only Paper Unicorns are likely to have a sub-$1B outcome from here on out. It is quite likely that some Public Unicorns (i.e., VC-backed companies that had a $1B+ IPO valuation) today will lose value, and end up becoming nonunicorns. However, this is at least partially offset by those VC-backed companies that went public with valuations below $1B, but ended up becoming Public Unicorns in the future. Also, arguably, VC batting average could have deteriorated as the number of investments increased. On the flip side, VC batting average could have improved with scale, largely driven by larger end-market opportunities or lower cost to scale tech businesses. For instance, the latest cohort of companies will likely benefit to a greater degree from the emergence of cloud computing and mobile internet, and all the new resulting business opportunities for early movers. In addition, arguably, VC batting average for overseas investments could be higher than U.S. investments, largely due to lack of incumbents or early mover advantages. Regardless, we believe that roughly 30% of today’s Paper Unicorns will have a sub-billion-dollar outcome in the future. Put in a more positive light, 70% of today’s Paper Unicorns are likely to become Real Unicorns.
Quick Note About Data Sources, Assumptions and Caveats
While analyzing macro-level data for the venture capital industry with specific focus on technology companies, we relied on data available on a variety of sources. This included PitchBook, Preqin as well as publicly available datasets prepared by CrunchBase, NVCA/Thomson Reuters, Dow Jones Venture Source, and Jay Ritter’s blog on IPO trends. In order to analyze and draw conclusions from data, we made several simplifying assumptions and tradeoffs. Below we have attempted to provide a list of caveats and assumptions in our analysis. While we have fairly high level of confidence in the big picture conclusions in this report, we’d highlight that relying on specific period or category data may not lead to accurate conclusions.
In order to calculate trends in VC batting average, we relied on the number of “information technology” companies with headquarters in U.S., founded since 1995, and have at least one venture capital or angel investments.
We gathered data for Software, Media & Entertainment, Telco, IT Services, etc. According to NVCA, VC investors have done approximately 2,000 to 3,000 new investments in these sectors every year in the U.S. Noteworthy outliers have been circa 2000 when there were more than 5,000 net new investments, and circa 2009 with fewer than 1,500 tech or related investments. Over the past couple of years, the run rate has clearly ticked higher to about 3,000 investments per year. We illustrate this trend in Exhibit 16 in the Appendix.
Per Preqin, approximately 8,000 to 10,000 VC deals are completed globally across all sectors every year. Roughly 60-70% of these deals are categorized as Internet, Software, and other related IT companies. Roughly 30-40% of VC investments are completed in the U.S. with the rest in Europe and APAC. This proportion has been slowly declining over the past five to seven years
We noticed that several recent late-stage private tech deals were not categorized as “venture capital” investments by public databases. But, we included such deals into our analysis.
We noticed that data for IPOs and VC investments prior to 2008 has been fairly inconsistent across public industry sources, particularly around industry category and pre-IPO funding. We relied on CrunchBase data for IPO & M&A trends, and cross-checked it with PitchBook data.
In our analysis, we have assumed certain thresholds of funding or fundraises to indicate a certain stage of company development or investment focus of a VC fund. We acknowledge that lines blur around these thresholds, and hence we haven’t relied on precise numbers within a range or basket to draw conclusions.
We noticed that VC investment data for international markets has not been captured to similar levels of historical accuracy as has been catalogued by NVCA/Thomson Reuters for the U.S. We have used U.S. VC investment activity as a proxy for international unicorns, and then developed sensitivity analysis on top of these assumptions.
We observed that there is a fundamental difference in the way industry data providers define or count “deals” or “investments” in tech sectors within venture capital. For instance, NVCA tracks roughly 4,400 investments per year in the U.S. across all sectors, Preqin reported approximately 10,000 deals globally, whereas PitchBook tracked approximately 8,000. We ended up making simplifying assumptions to gross up the number of VC investments globally from Thomson Reuters data (NVCA), by assuming a general shift toward international markets over the past 6-7 years.