Investing in a ’unicorn’ in the early stages is exceedingly rare. The odds are not much better than seeing an actual unicorn in a forest.
For eevery company founded by aspiring entrepreneurs, only 1% receive external funding of any sort. From this universe, only 1% will then go on to reach that coveted billion-dollar plus valuation.
So how do professional venture capitalists, who serve as the stewards of their investors’ capital and are expected to pick unicorns as if they were apples on a tree, approach this well-trodden path of ‘failure’?
The power law of venture capital fund returns is extremely important. This theory implies that for a fund to be successful (i.e. return 30% IRR+), investing in businesses that only have the potential to become a 100 million dollar plus outcome will be inadequate once the failure rates of early-stage companies are factored in.
Basically, funds are better off shooting for unicorns every chance they get as most of their value will be created by a very small number of companies while the rest slowly ride off into the sunset. It’s called the ‘Babe Ruth effect’: swing big with everything you’ve got and you’ll either hit big or miss big. To use another baseball analogy, your ability to hit home runs is more important than your ability to get off base.
Venture capital funds will also employ a strategy that gives them the best chance of uncovering a disproportionately large number of unicorns relative to the rest of the industry. This may come from investing in a particular sector they believe they have an unfair advantage in (e.g. fintech). It may be based on a certain type of technology that they have experience working with (e.g. Internet of Things). It could also revolve around a business model (e.g., B2C v B2B).
Unfortunately, for all the time and effort VCs invest in their strategy, there is no secret formula to picking special companies that go on to change industries and become household names. Or is there?
I recently read a book called ‘Super Founders’ by Ali Tamaseb, who presented a novel approach to how data can be used to identify unicorns. Of course, there is no one formula to unlock startup value, but I would like to run through a few of his key data driven observations. To be clear, the ideas presented and discussed (much more eloquently and in much more detail) in the book belong to Ali Tamaseb. I have merely consumed and summarised them (adding some of my own commentary along the way).
For background, 30,000 data points that form the basis of the study were collected from companies between 2005 and 2018. During this period, there were over 200 billion-dollar companies founded that formed the billion-dollar group. An additional 200 randomly selected companies were used for the random group. As clarified in the book, correlation is not causation.
The categories that those data points were organised into were as follows: Founders, Product, Market and Capital.
OK, let’s begin.
Founders are where every investment decision starts. How good are they? But what is the ‘right’ age and where should the founders come from?
When one thinks of the prototype unicorn founder, the brilliant 25-year-old MIT drop-out springs to mind. One, maybe two mates with lofty dreams keen to roll the dice in the hope of making it big. Right? Wrong.
Apart from founders of billion-dollar companies in the health/biotech sectors (which are on average older than other sectors), the distribution of age is similar across the unicorn group and the random group (with a median of 34 years). Younger or older does not present any material advantage with the success of a company.
And to those investors who shy away from either single founder companies or those five person founding teams at the other end of the spectrum, you can relax. According to Tamaseb, there is no statistical correlation between the number of founders and the success of a company. The most common combination for a billion-dollar startup was two founders (representing 35%), followed by three, one, four and five or more. These numbers were very similar to those in the random group, suggesting no specific advantage or disadvantage to any scenario.
CEOs and CXOs (CTO, COO, CRO, etc.) of billion-dollar startups, however, were more likely to have been technical (someone who could build the original product, etc.) when compared with the random group. For CEOs in the billion-dollar group, the number of technical founders was 50%, with the random group showing 40%. And more often than not, that founder is supported by a technical CXO of which the number was 70% versus 60% in the random group. Not massive, but still meaningful.
Looking at the education of founders, it was clear that billion-dollar founders went to better-ranking schools compared to those in the random group, but what played a more important role than the ranking of the university itself was the location and the entrepreneurial culture of the university.
As most unicorns have been produced in America, it is no surprise to see Stanford, Harvard and MIT topping the list with almost as many billion-dollar start-ups as every other institution combined. One theory is that candidates from top schools have an easier time getting in front of investors and may have had more support/safety nets to fall back on should they not be successful, which facilitated the necessary level of risk.
The work experience of founders presented one of the biggest statistical variances across the two groups. Founders of billion-dollar startups were almost twice as likely to have worked at a ‘Tier 1’ company versus the random group. ‘Tier 1’ is obviously subjective, but think of the big management consulting firms, global technology companies, investment banks, etc. The other item to note was that founders within the billion-dollar group were more likely to have worked for themselves. Roughly 30% of these founding CEOs had not worked for anyone other than themselves before.
This one is really interesting. Among billion-dollar start-ups, fewer than 50% of founding CEOs and fewer than 30% of founding CXOs had much, if any, work experience that directly related to their startups. To start a proptech company, you do not need to have worked in the property markets (nod to you @Rhys Rogers @BYB). Likewise, do you need to have worked in insurance to start an insurtech business.
What is obviously important, are the soft skills like hiring talented people, forming partnerships, being a good salesperson and solving problems. Bar the health/biotech sectors, there was no significant difference between the billion-dollar and random groups when it came to this domain expertise.
When it comes to the question of being a first time versus a repeat founder, and how that affected the probability of starting a billion-dollar company, a compelling 60% of founders had been there and done it before. It mattered less whether the previous venture was a big success or a dismal failure, and more that the founder had been in the hot seat on at least one previous occasion.
When that previous venture resulted in an exit at anything north of a $10 million valuation, the number in the billion-dollar group goes up to 70% versus 24% in the random group.
Ali Tameseb felt the data was so compelling that he named the book after these so-called ‘Super Founders’. There are plenty of well-known billion-dollar founders out there but what is less well-known are their fairly meagre efforts before hitting big.
Kalanick started Scour (filed for bankruptcy) and Red Swoosh ($19 million exit) before founding Uber. The Collision brothers founded Auctomatic (exited for $5 million, not quite $10 million but a result all the same) before founding Stripe. The same trend can be seen across the Pacific. Tech veteran, Michael Starkey founded iSelect in 1999 before he went on to found Athena Home Loans in 2017. Maybe, more incredible was Nick Molnar, who founded Iceonline when he was just 20 years old and went on to found the $42-billion behemoth Afterpay. Who said the past was not a reliable predictor of the future?
We strongly believe in the famous Marc Andreesen quote at Aura Ventures, “Software is eating the world”. It was made in 2011 and still rings as true as ever. Ali’s data supports this notion with 54% of billion-dollar companies being pure software plays versus 40% in the random group. The balance for reference was made up of consumer product companies, healthcare/biotech/pharma, physical business products, energy/minerals and finance. Within that software designation, the most common subsectors, which account for around 40% (of the 54%) are business/productivity software, social/consumer software and application software. There is no suggestion that these trends will continue as new industries are unlocked by additional technological capabilities.
Classification as a painkiller or a vitamin pill, as expected, had a big bearing on a company’s statistical chance of ultimate success. Where a painkiller takes away a well-defined pain point, a vitamin pill seeks to improve the way something is done either through value, efficiency or entertainment. Although just over 30% of billion-dollar companies were vitamin pills, which suggests there is merit to building them, the rest (70%) were painkillers in what is clearly a significant difference in the data.
When it comes to the company’s approach to the problems they are solving, the two needs that the billion-dollar start-ups are addressing are productivity (40%) and saving money (20%). The random groups are only 30% productivity and 10% saving money, with far more focus on convenience and health.
The level of engineering complexity of a company not only shapes its fabric but everything from its capital requirements to the types of employees that are required. Surprisingly, engineering light companies that focus on systems integrations, when compared with highly technical and deep tech companies, do not significantly underperform when it comes to achieving billion-dollar valuations. Overall, billion-dollar companies are slightly less likely to be system integration style companies and slightly more likely to be technical and deep tech-based.
Looking at whether a company had engineered a product that was highly differentiated versus incrementally different — over two thirds of those in the billion-dollar group had managed to achieve highly differentiated status (against 40% in the random group).
Not wanting to give all the good stuff in the book away and stop you from reading it yourself, there is an excellent quote that I wanted to include from Andy Rachleff, founder of Benchmark Capital and Wealthfront, “When a great team meets a lousy market, market wins. When a lousy team, meets a great market, market wins. When a great team meets a great market, something special happens.”
There is no denying that the market is an extremely important part of assessing any new early-stage investment and it was perhaps not surprising to read that the data supported the presence of a ‘large’ market in 60% of the billion-dollar group (versus 45% in the random group).
What was more interesting was what the data said in relation to creating a market versus competing for market share. Contrary to popular belief, only 32% of billion-dollar companies were creating a new market, with the rest competing for an existing market. These numbers were very similar across both groups, suggesting that there is actually no real statistical advantage of being first or being second or third, etc. While first mover advantage may be a real thing, what is more crucial is executing well, being differentiated, and addressing the right needs.
On the core sales segmentation of a company and whether it sells to consumers or to enterprises, both models were well represented suggesting similar rates of success for each.
Being early in the market is generally considered as being wrong. Unless you have an unlimited amount of capital to keep the lights on whilst waiting for the world to catch up with you, your star is going to fade out before the magic happens. This could not be illustrated better through some of the well-known success stories that were built on recycled ideas.
Think of all the search engines that existed before Google and the social networks before Facebook's time. Just as creating the market was not any more statistically significant than competing for market share, billion-dollar companies were no more likely to be the first to try an idea than be the tenth to do so. Those outside of the first five to try an idea were almost as likely to be a unicorn than those inside the first five. What has a greater bearing on success are things like: technological capabilities, regulation changes and behaviour shifts. Always ask ‘why now?’ as part of the investment process.
As a continuation of the market’s status at the time of launch and the order in the line of businesses to address its needs, is the concept of competition at the time of founding. Popular theory maintains that startups entering crowded markets will struggle to gain traction against well-funded and all-powerful incumbents. However, the data does not agree.
Over half of billion-dollar startups faced and overcame this struggle against one or more incumbents. This number was significantly lower for companies in the random group. Companies facing no or fragmented competition were pegged at the same rate across both groups. So, perhaps it is less a case of incumbents being a factor to shy away from and more of them validating a market worth pursuing?
If the success of a billion-dollar company is not highly correlated to the level of competition in a market, it is probably because they have managed to craft a strong level of product defensibility. The data showed that start-ups with network effects, engineering defensibility and branding were significantly more likely to become billion-dollar companies. Technical complexity, as mentioned above, was not enough to create a moat on its own and is commonly coupled with other moat types.
To make all of the good stuff happen, a startup obviously needs capital. How much capital exactly is addressed below, but it is generally thought that capital intensive businesses make poorer investments than their capital light cousins. This is because capital-intensive businesses at some stage will become dilutive to the founders and early investors (unless they keep funding the business, which is not always practical).
The billion-dollar cohort, however, had close to 60% of companies in the medium or high capital requirement designation. This was compared with about 35% in the random group, which were better represented in the light capital requirement section. I found this surprising as it may speak to the defensibility of a business with high capital needs.
There is also the question of capital efficiency, which is related to capital required but is not always correlated, i.e. you may have a high requirement for capital and be either capital efficient or capital inefficient. It highlights the need to assess the unit economics of a business during due diligence.
Another element of a start-up’s capital journey is how much it raises in its first and second rounds. Statistically, billion-dollar start-ups raised double the amount of the random group in their first raise (median of $5 million versus a median of $2.5 million) and more than triple in their second raise (median of $15 million versus a median of $5 million). This data in no way takes into account a company’s need for capital – and in that, highlights that the market in a lot of cases was able to identify the quality of a company early which led to bigger rounds.
And then finally, not only the quantum of the money you have raised can make a statistical difference, but so can the source. Startups that raised capital from ‘Tier 1’ investors were a massive three times more likely to reach that billion-dollar group. The inverse of this is also true of ‘Tier 3’ investors, i.e. there are three times as many from the random group to the billion-dollar group in that category. What constitutes a ‘Tier 1’ investor versus a ‘Tier 3’ investor is obviously subjective but that is a story for another time.
There it is. Not exactly a secret formula but some data points to keep in mind next time you are at the table discussing a new opportunity.
To summarise, you are looking for a startup with any number of founders of any age, ideally the CEO and CXO are technical, come from a ‘Tier 1’ university and have worked for a ‘Tier 1’ employer and/or themselves. But they do not need to be domain experts. If you can find a founder who has exited a company for more than $10 million that will help.
Next, you want them to be working on a software project that is a painkiller and addressing defined problems in productivity or a P&L. It should be highly differentiated but does not need to be deep tech. You want the founders to be going after a ‘large’ market but they do not need to be the first to market, and in fact being later to the scene may be an advantage.
Do not be scared of incumbents (but only because they are building a defensible moat). The company should raise about $5 million in its first round and $15 million in its second round (because it should be medium/high capital intensive), from the best VCs you can get in front of.
Hopefully this has debunked some myths for you around what qualities a potential unicorn may possess. It certainly did for me.
Again, all of the data in this piece has been sourced from the book ‘Super Founders’ by Ali Tamaseb. If you found this interesting, I highly recommend a full read of the book.