Prevailing notions are important because they guide critical decisions made by founders, tech executives, and investors. For example, many founders locate (or relocate) their companies in Silicon Valley in the belief that this will help them attract engineering talent or venture capital. Some founders rush to raise venture capital early in their startup formation history, and others desperately seek to attach machine learning, blockchain, or computer vision to the end of their startup pitches in the hopes of attracting investors.
At 645 Ventures, we like to test preconceived notions, and so we decided to study the billion dollar tech companies of the past decade to understand what they looked like in their earliest days. We were inspired by a conversation we had with our friend Will Quist from Slow Ventures, where we discussed whether top early-stage venture capital firms systematically overlook certain types of exceptional companies.
We began with a Crunchbase data set comprising U.S. companies in the software and Internet categories that reached a valuation of $1 billion or greater over the past decade. We analyzed companies that reached this milestone in the private markets, via acquisition, or via initial public offering.
We chose not to include companies from capital-intensive industries, such as cleantech, biotech, or hardware companies, because their capital needs are generally higher than for software companies, and they are less likely to be bootstrapped. To further simplify the task, we focused only on U.S. companies.
There were 234 companies in our data set. The list included household names such as Lyft, Airbnb, Uber, Facebook, and Slack. But it also included much lesser-known businesses such as Ibotta, KeepTruckin, and Lytx.
The data was surprising in many ways and led us to question many assumptions prevalent in early-stage investing. As a result, we’ve structured our learnings into the five biggest myths about billion-dollar startups.
As an early-stage fund, we were very curious about the capital raising tendencies of this elite group of companies. We expected that these companies would be well-funded at the early-stage, due to high investor demand. The data showed the opposite. Only 56% of companies raised a seed round of any kind. This figure included rounds from angels and friends and family, so the actual percentage of institutional seed rounds was in fact even lower.
Additionally, for companies that did raise seed rounds, round sizes were relatively small. The median seed round size was $1.5m. The average seed round size was $3.7m, skewed upward from the median by very large seed rounds. Almost ⅔ of companies raised either no seed capital, or less than $1m of capital, at the seed. For Series A rounds, the median round size was $9 million, which is also a relatively conservative round size, especially in this age of supergiant venture funding rounds.
We interpret this data to suggest that capital efficiency is a virtue for early-stage startups, particularly at the seed stage. And as the table below shows, many of the largest outcomes in our data set raised small seed rounds or no seed capital at all. Given the factors that have driven down the cost of starting a company over the past decade, including cloud infrastructure and low-cost software distribution via the app store, capital efficiency may be even more of a virtue for startups over the next decade.
|Company||Founded||Seed Round||Latest Valuation|
|Dropbox||2007||$1.2M from VCs||$8B|
|2009||$250k from angels||$19B|
|Uber||2009||$1.25M from VCs and angels||$76B|
|StitchFix||2011||$750k from VCs||$2.6B|
|Snap||2011||$500k from VCs||$24B|
|Coinbase||2012||$600k from YC and angels||$8B|
WhatsApp’s first seed round was a $250k funding round in 2009, followed by an $8m Series A round in 2011. The company raised only $60m of capital throughout its history. The company was acquired by Facebook for $19 billion in 2014.
One of the prevailing notions about venture capital is that it’s a game of access, with only the best firms getting access to the best deals. This enables the rich to get richer. Andy Rachleff, Co-Founder of Benchmark Capital and Founder/CEO of Wealthfront, described this prevailing wisdom in his blogpost series, “Demystifying Venture Capital Economics”. “The top 20 firms… generate approximately 95% of the [venture] industry’s returns. These 20 firms don’t change much over time and are so oversubscribed that they are very hard for new limited partners to access.”
Even if the top VC firms do generate the majority of the industry’s returns, we wanted to understand if this translates into their monopolizing the billion-dollar outcomes, or whether there was a lot more inefficiency at the early-stage than we think. We compiled a list of 70 of the best-known VC firms with established track records, using our own internal investor analytics platform, as well as publicly-available data. This list included premiere seed firms such as First Round Capital, Floodgate, Baseline Ventures, and Founder Collective, as well as the best Series A/B firms including Sequoia, Union Square Ventures, Benchmark and Accel. We then analyzed the percentage of the billion-dollar companies that they invested in at Seed and Series A.
We found that these firms missed a much higher percentage of billion-dollar companies than we expected. At the seed stage, only 31% of the companies raised capital from one of these top 70 firms. Almost 70% did not have one of these firms in their cap table at the seed.
While the best-known firms showed up at the seed, they were only present in a fraction of the rounds. Leading the pack were a16z and Accel, who were both in 7 unicorns, equivalent to 3% each of all companies. They were followed by Sequoia and Y-Combinator, who were each in both in 6 unicorns.
The market became more efficient at Series A. 71% of companies raised Series A from a top firm. However, there were many big winners in the remaining 29% of companies that didn’t raise from a top firm. We provide a few below. These overlooked companies were frequently in markets that were not perceived to be large or attractive at the time, such as Peloton or Pandora. This data suggests that there is opportunity for new firms to generate alpha by focusing on markets that are perceived not to be ready for prime time.
Peloton, which went public in 2019 at an $8B market cap, was passed on by hundreds of VC firms throughout its history. In the words of Founder/CEO Jon Foley, "Every round, for six rounds. Andreessen, Bessemer, Sequoia ... they passed again and again." This was largely due to the non-conventional nature of the company’s business model and market focus.
To better understand the unique characteristics of billion-dollar companies, we looked closely at the user behavior, business model, and technology innovation of each company in its early days. Our goal was to understand whether companies fit the traditional perception of their taking great risk in these areas.
The characteristics we evaluated were:
1) User Behavior: How widespread was the user behavior the startup was aiming to promote? For example, for Airbnb, we assessed how widespread the practice of staying in a stranger’s home was at the time of the startup’s founding. We categorized user behavior into three types: non-existent, existing, and widespread. Widespread behaviors were the least risky in terms of adoption. Existing behaviors required behavioral change on behalf of the target user or customer, while non-existent behaviors required the startup to evangelize the new behavior
2) Business Model: Did the company establish a new monetization model, or was it leveraging a business model that was already proven? Proven business models connote less risk in monetization. For example, the cryptocurrency platform Coinbase had an unproven business model because individuals had not purchased cryptocurrency in an established model.
3) Product Improvement: Was the product improvement a gimmick, incremental, radical, or disruptive innovation? We previously defined each of these in our Venture Investment Triangle article.
Conventional wisdom about venture investors is that they invest in startup moonshots, represented by companies with non-existent user behavior, non-proven business models, and/or disruptive innovation. However, our data found none of these assumptions to be generally true for the data set. 84% of companies were operating in a category that had existing or widespread behavior at the time of the company’s founding, which significantly reduces the risk of user adoption.
However, while there were fewer companies that took on behavioral risk, those who assumed this risk appear to have generated larger exits. The average last valuation for those companies was $5.1 billion, versus $3.8 billion for the overall dataset. For example, Uber and Lyft are examples of companies that were facing nonexistent user behavior, the biggest uphill climb, at the time of their founding. Individuals hadn’t hailed cars with their mobile devices before these companies arrived, and they catalyzed a completely new behavior.
Airbnb, the home and vacation rental marketplace, struggled to raise seed capital in its early days, due to investors’ skepticism regarding whether the behavior of renting another’s home online would be widely adopted. As a result, the company was repeatedly passed on by many investors as it attempted to raise its seed round. However, the behavior that Airbnb (founded in 2008) sought to capitalize on was existing, due to the presence of companies such as HomeAway (founded in 2004) and VRBO (founded in 1995). Leveraging a proven business model and incremental technology innovation, and expanding on the innovations of companies that came before, Airbnb scaled rapidly to billions of dollars of revenues after its initial struggle.
83% of companies leveraged a proven business model, and this percentage increases to 92% when analyzing the companies seeded by top investors. This reflects the fact that the majority of billion-dollar companies were not taking as much business model risk as we might assume.
Finally, 92% of companies were developing a product that was offering either incremental or radical innovation compared to incumbents, with incremental innovation comprising 58% of companies. This percentage increased to 96% if we analyze companies seeded by top investors. Truly disruptive companies, such as the virtual reality company Oculus or the Robotic Process Automation Company UiPath, were quite rare.
These results are quite surprising, as we often hear a narrative of billion-dollar startups catalyzing new user behavior or pioneering disruptive technologies. In fact, billion-dollar startups are much more likely to capitalize on existing user/customer behavior trends, or build on existing technologies that are already ready for prime time.
Our study revealed that although the Bay Area is by far the most frequent hub of billion-dollar companies, it doesn’t have a monopoly on billion-dollar tech company formation. In actuality, 46% of all billion-dollar companies in our data set, nearly half of them, were formed outside of Silicon Valley.
Outside of the Bay Area, the next top 5 regions comprised 29% of total billion-dollar companies, and the next 12 regions comprised 38%. These next 5 regions, ranked in order, were New York, Los Angeles, Boston, Washington (Seattle) and Chicago. New York comprised 14% of all companies, with L.A. following next at almost 5% of all companies, followed by Boston, Washington and Chicago in the 3-4% range.
Interestingly, several billion-dollar companies were created in regions not known to be technology hubs. Florida produced 4 billion-dollar companies in the group, while Atlanta and Colorado produced each produced 3. There also was a long tail of regions and cities that produced one to two billion-dollar companies in the group.
In addition, we found that common wisdom regarding cities producing unicorns in specific industries does not generally hold true. For example, one would assume that New York is more likely to produce billion-dollar fintech startups vs. infrastructure software companies, due to the fact that it is a financial services hub. However, New York in fact produced 3 infrastructure software unicorns and only 2 fintech unicorns. Similarly, while Boston is perceived as a hub for biotech but having lost its edge in software and Internet, it in fact produced unicorns across categories including infrastructure software, fintech, ecommerce, and security.
We believe that the trend toward increasing geographic dispersion of billion-dollar companies will continue. Due to the increasing ability of companies to scale distributed teams, combined with the dispersion of talent and capital, we believe that Silicon Valley will continue to lead but won’t be as dominant. Below are examples of billion-dollar companies built in regions not perceived to be tech hubs.
|Epic Games||Cary, NC||$15B|
|Chewy||Dania Beach, FL||$8.8B|
|Magic Leap||Plantation, FL||$6.3B|
MongoDB, the cross-platform document-oriented database software company, was founded in New York and raised a $1.5m Series A led by Union Square Ventures, based in New York as well. The company eventually went public in October 2019, and has a $7 billion market cap today.
Lastly, our study demonstrated that founders come from unexpected places in terms of their career paths, academic accomplishments, and previous entrepreneurial experiences. Beginning with entrepreneurial experience, some founders and investors believe that previous successful entrepreneurial experience is a prerequisite for building a billion-dollar company. However, our data instead showed that ~40% of the founders in the dataset were first-time founders, a much higher figure than we expected. Furthermore, an additional 17% had a previous startup that failed. In total, almost 60% of billion-dollar founders did not have a successful entrepreneurial past.
In specific technology industries, there is an even greater percentage of first-time founders. For example, ~54% and ~48% of infrastructure software and fintech unicorns, respectively, were founded by first-time founding teams.
Along with considering previous entrepreneurial experience, VCs also tend to rate certain previous employers higher than others. In our study, the companies below showed up most frequently in terms of producing billion-dollar founders.
While Google and Facebook are widely considered as top employers, we were intrigued by companies such as Oracle and McAfee being included on the list as well as the connection each of these companies had to a specific industry. Amongst all the previous employers we analyzed, Oracle produced 18% of infrastructure software founders, while McAfee produced 36% of security founders.
We also were intrigued by Amazon and Microsoft. In the venture industry, a common wisdom has been that these two companies don’t cultivate entrepreneurial talent well, and therefore they wouldn’t rank highly in terms of producing unicorns. The opposite is actually true. Those companies rank #2 and #4 in our list in terms of mafia formation, producing 21 billion-dollar companies between them.
Similarly, we’re also seeing newer companies started in the last decade that have already begun to produce unicorns. Warby Parker, for example, founded less than ten years ago, has already produced 3 unicorn companies.
|Employer||Number of companies|
Root Insurance, the auto insurance insurtech platform, was founded by all first-time founders. However, both co-founders had impressive backgrounds. Alex Timm (CEO) was an actuarial science major who had previously spent almost 4 years at Nationwide Insurance. His co-founder & CTO, Dan Manges, was previously the CTO of Braintree, which PayPal acquired for $800 million. Furthermore, Root raised its Series A from Drive Capital, which was started by former Sequoia investors and is today one of the largest Midwestern-based venture capital firms. According to the company, Root Insurance wrote more than $187 million in insurance premiums during the first six months of 2019, growing over 800% from the same period in 2019.
We found ourselves inspired by the findings of this study. Our firm mission is “bringing the invisible to life”, and one of our core tenets is empowering founders to do what might seem to be impossible. Furthermore, we’re inspired by the underdog.
Many of our findings are consistent with that of improbable founders building billion-dollar businesses. The finding that a majority of founders had not had a previous successful startup, and 40% were first-time founders, suggests a more even playing field than we expected. The fact that the majority of founders don’t raise large VC rounds in their early days is also consistent with the resourcefulness and scrappiness that we have observed in exceptional founders. That many of them are missed by the top VC firms supports the notion that a founding team doesn’t need to be anointed by the best firms in its early days to succeed.
Although the majority of unicorns are located in the Valley, locating a startup there doesn’t necessarily improve its probability of success, given the level of competition there. Further, our findings show that billion-dollar companies can be built even in small cities without a history of technology entrepreneurship.
Most perplexing, and perhaps a valuable subject of future analysis, is the data relating to the types of business models, user behaviors, and technologies that reach billion-dollar scale. Founders and investors in billion-dollar startups are embracing less risk than we expected. Is this because the failure rate associated with building truly transformative companies is higher, or because founders embark upon them less frequently?
Our overall takeaway from this study is that founders and companies do not fall into preconceived buckets. Whether it’s the founder background, the geography in which the company is located, or the capital needs of the business, great companies often follow uncharted paths. We hope this study encourages both investors and founders to consider new possibilities as they fund and build new billion-dollar companies.