Skip to content

Harvard Can Count Its Unicorn Founders, But Can It Prove It Made Them?

Entrepreneurship has become one of the defining promises of modern business education. Business schools have built entrepreneurship centres, incubators, accelerators, venture competitions and entire programmes around the idea of helping students become founders. Successful entrepreneurs are prominently featured in alumni communications, and unicorn founders have become particularly powerful symbols of institutional success.

There is nothing surprising about this. At a time when fewer students see a traditional corporate career as the only definition of success, entrepreneurship is an increasingly important part of what business schools are expected to offer.

But there is an interesting problem with the way we measure it.

Running the startup analysis backwards

Most of the evidence starts with the institution. A school looks at its alumni population, identifies successful entrepreneurs and points to them as evidence of its entrepreneurial impact. We decided to reverse the direction of the analysis.

Instead of starting with business schools, we started with unicorns.

At Deans Insight, we analysed 8,595 founder records associated with 4,113 highly valued companies from Boston to Beijing to Bangalore, and then traced their academic backgrounds. For 5,499 founders, around two-thirds, we were able to identify educational information. The question was deliberately simple: where did the people who built these companies actually study?

Business schools are certainly there, and in good number. Harvard Business School appears 178 times in the data, Stanford Graduate School of Business 139 times, Wharton 93, INSEAD 66 and MIT Sloan 47. Berkeley Haas, Columbia Business School, Northwestern Kellogg, HEC Paris and London Business School also appear in the educational histories of unicorn founders.

Those numbers are impressive. But they become more interesting when placed alongside the rest of the educational landscape.

Stanford University appears 339 times, MIT 268, Harvard University 203 and UC Berkeley 176. Across the dataset, universities with deep connections to technology, engineering, science and research appear repeatedly.

The point is not that STEM universities create entrepreneurs while business schools do not. The data cannot support such a simplistic conclusion. What it suggests is that business education is only one pathway among several through which the founders of extremely high-growth companies emerge. 

And that immediately makes the word “produce” much more complicated.

No business school stands alone

Consider Stanford GSB. It is one of the world’s leading business schools and it sits prominently in our data, which invites an obvious interpretation: that the school is exceptionally good at producing founders.

But Stanford GSB does not exist in isolation. It sits inside Stanford University and, more importantly, within Silicon Valley. An MBA student there spends two years surrounded by one of the densest concentrations of engineers, computer scientists, researchers, venture capitalists, technology companies and serial entrepreneurs anywhere in the world.

If that student subsequently builds a unicorn, what exactly produced it?

Was it what happened in the classroom? The business school network? Stanford’s engineering and computer science capabilities? Access to investors down the road? Silicon Valley itself? Or the fact that Stanford was able to attract an exceptionally talented and ambitious individual who might have become an entrepreneur anyway?

The answer is almost certainly some combination of these factors. 

The same pattern appears elsewhere. MIT is embedded in the Boston-Cambridge innovation ecosystem. Berkeley sits within the gravitational pull of the Bay Area. London brings together universities, finance, technology and one of Europe’s deepest pools of venture capital within a few miles of each other. Paris increasingly connects business schools, engineering schools, founders, investors and technology companies within the same ecosystem.

A billion dollars is easier in some places

Geography, it turns out, runs through the whole dataset. More than 5,000 of the 8,595 founder records are associated with US companies. China and the UK follow at a considerable distance, followed by Germany, India and France.

Are people in Americans really five or ten times more entrepreneurial than elsewhere? What the numbers reflect is the extraordinary concentration of capital, technology, networks and company valuations in particular entrepreneurial ecosystems.

A founder can build a highly profitable company employing thousands of people without ever creating a unicorn. Another can operate in an ecosystem awash with venture capital and reach a billion-dollar valuation relatively early in the company’s development.

So when we count unicorn founders by academic institution, we may believe that we are measuring education when we are partly measuring geography, access to capital and ecosystem density. 

The missing denominator

So is Wharton more entrepreneurial than MIT Sloan or LBS? There is another statistical problem hiding behind the impressive numbers: the denominator.

A school with 100 unicorn founders among its alumni sounds more entrepreneurial than one with 50. But if the first institution has 100,000 alumni and the second has 20,000, the picture changes completely.

Absolute numbers measure size, but tell us nothing about likelihood. A much more revealing metric might therefore be the number of unicorn founders per 10,000 alumni. Such a measure could produce a very different hierarchy from the one generated by simply counting famous founders.

Yet even that would leave the most difficult question unanswered.

If someone did their MBA at Harvard Business School and subsequently created a unicorn, what exactly can Harvard legitimately claim? Did they produce that startup success?

At the most basic level, we know that a unicorn founder attended Harvard Business School. That is a fact. Saying that Harvard Business School produced that founder is a much bigger claim.

Two companies, two different claims

And the difference is not merely theoretical. Consider Grab.

Anthony Tan and Tan Hooi Ling were classmates at Harvard Business School when they developed the idea that became MyTeksi and later Grab. They entered the HBS New Venture Competition, came runner-up, and the prize money helped fund the company’s launch.

That creates a relatively clear line between institution and outcome. The future founders met at the school. The idea was developed there. The school’s entrepreneurship ecosystem helped test it, and its venture competition provided early funding.

The founders brought their own ideas, experiences and ambitions, and what happened over the following years depended on more than their time at Harvard. But it is reasonable to say that the school played a meaningful role in the company’s creation.

And then there is Hugging Face, which Nvidia agreed to acquire for $12.93 billion. 

One of its co-founders, Clément Delangue, studied at ESCP Business School. That makes Hugging Face a perfectly legitimate company to associate with an ESCP alumnus. But he built the company with Julien Chaumond and Thomas Wolf, co-founders with very different technical and professional backgrounds, and the company developed within a much broader technology and open-source ecosystem.

Did ESCP play a role in Delangue’s entrepreneurial journey? It may well have done. But from the educational affiliation alone, we cannot know what that role was.

Yet in a dataset of unicorn founders, these examples can ultimately look remarkably similar: a successful founder associated with a business school.

It would be difficult to argue that the business school had the same impact in each case.

Past performance as the predictor

This distinction becomes even more important when we consider the selection mechanisms of elite institutions. Elite schools do not admit at random. Harvard, Stanford, INSEAD, Wharton and their peers select people who are already unusually ambitious, accomplished and often well connected. Some show up having already worked in startups before they ever show up on campus. Others arrive with well-articulated entrepreneurial ambitions. And there are those who have founded businesses before they ever enter the classroom.

When those people go on to succeed, separating the selection effect from the education effect is close to impossible with the data anyone currently holds.

None of this is unique to entrepreneurship. The same issue arises every time a university points to CEOs, political leaders, billionaires or other highly successful alumni as evidence of its impact. There is a real difference between identifying talented people, educating talented people and causing the outcomes those people later achieve.

Higher education has historically been much better at demonstrating the first two than proving the third. 

Moving beyond a tally of unicorns

Which is why the most interesting questions in this unicorn dataset have nothing to do with who comes top.

When did founders attend business school relative to the creation of their companies? Had they already founded businesses before enrolling? Did they meet their co-founder in the building? Did the company come out of university research? Did the first cheque arrive through an alumni network? How long after graduation did they incorporate, and how far from campus? And does any school still stand out once you adjust for alumni numbers, geography and proximity to capital?

Answers to those questions would tell us far more about entrepreneurial impact than another league table of famous alumni.

They might also change how business schools think about entrepreneurship itself.

Perhaps the most successful entrepreneurial institutions are not necessarily those with the largest entrepreneurship centres or the longest list of startup electives. Maybe what matters is the ability to connect: management with engineering, ideas with capital, researchers with operators, people who want to start something with people who already have.

Geography and network density may matter more than curriculum. Admissions selectivity might be more impactful than teaching. Or perhaps particular educational experiences really do dramatically increase the probability that someone becomes a successful founder.

Grab suggests that sometimes they do. Hugging Face reminds us that an alumni connection alone cannot tell us the whole story.

At the moment, we too often jump from correlation to causation without knowing which of these explanations made the difference. 

Business schools should carry on celebrating their founders. Alumni success matters, and institutions are entitled to be proud of who has passed through them.

But celebrating an alumnus and claiming an outcome are not quite the same thing, and for the first time the data to tell them apart is within reach.

The growing availability of public information about founders, companies, funding rounds and professional trajectories, combined with our ability at Deans Insight to analyse that information at scale, means that we can increasingly move beyond counting successful alumni. We can begin examining the journeys behind those outcomes and asking what role institutions may actually have played.

That is a much more demanding standard of impact, and fascinating more interesting one.

For years, business schools have been good at counting the unicorns among their alumni. The next challenge is demonstrating the role they played in creating them.

Benjamin Stevenin is the co-Founder of Deans Insight, helping higher education institutions move from claiming impact to demonstrating it.

Interested in this topic? You might also like…

Leave a Reply