If you’ve been following my recent posts, you may know that I spent the last four years conducting one of the largest data-driven studies into startups, trying to understand what was different between the startups that became billion-dollar successes versus those that did not. I published the results in a book called “Super Founders: What Data Reveals about Billion-Dollar Startups”.
This is one of the key questions I keep getting asked: What does your data reveal about founders as CEOs versus professional/hired CEOs. After all, many famous founders were replaced at some point including Steve Jobs, Jack Dorsey (Twitter), Andrew Mason (Groupon), and Jerry Yang (Yahoo).
So, I had to dig up the dataset to find the answer. Here’s what it revealed:
The Majority of Unicorns Are Founder Led
Among all the unicorns founded in the past 15 years, 65% still have the original founder as the CEO. Of those which were acquired or had an IPO that valued them over $1 billion, 73% were founder-led at the time of the acquisition or IPO.
These numbers are encouraging and seem to be much higher than what they would’ve been in the 90s when companies like Yahoo were founded and a sign that the majority of founders were able to hold on to the CEO seat and drive their companies to a billion-dollar acquisition.
As a reference point, I measured the same metric among failed unicorns (those that had achieved a billion-dollar valuation, but later failed, or their valuation fell below $1Bn): 47% were founder-led and 53% had a professional CEO. It might be easy to make the conclusion here that “hired” CEOs led to failure, but that’s probably not an accurate take. It’s often the case that when a company is troubled and things are not working, the board of directors gets influenced to bring a professional CEO, hence why the failed unicorns were more likely not to have the founder as the CEO anymore.
Those in Tech Remained CEO for 2X Longer Compared with Biotech
The average tenure of the founding CEOs that were replaced by a professional CEO was 5.47 years among tech unicorns and 2.75 years among biotech unicorns.
This again shouldn’t come as a surprise, since it’s more customary in biotech for the founders to develop the science and bring on a professional CEO at the commercialization point. There are many examples of exceptional biotech companies with their founders remaining as commercial CEO (Zymergen and Gingko Bio as two recent examples) and that trend is getting stronger as well over the past years.
Founder-Led Startups Created Larger Value
The average valuation of founder-led unicorns was 10.8% higher than those with a hired CEO and among those that had an IPO or were acquired for over $1B, the average valuation of founder-led ones was 18.5% higher than those with a professional CEO.
It’s important to note that among all the charts discussed in this article, there was no control group, as in we don’t know what would’ve happened to a specific company had it not replaced the founding CEO. Hence it’s not a straightforward conclusion that founding CEOs are always better than hired CEOs. There are always nuances and special cases for each company.
I personally believe founders make for great CEOs and that it’s to the benefit of startups, in the long run, to have founders remain as CEO (of course, as long as they’re acting in good faith, not doing anything illegal or alienating their employees). Even if the learning curve costs the company some delays and mishaps in the initial years, exceptional founders learn fast, make decisions with a 10-year view rather than a 3-year view and can attract great talent by selling their vision and passion. It’s encouraging to see the data support that as well. Other research has also shown that founder-led public companies were more likely to invest heavily in R&D, take less time, and less investment to achieve larger outcomes.
I discuss what data reveals about competition, defensibility, team, and fundraising patterns among unicorns vs. non-unicorns in my book Super Founders. If you love data-driven insights as much as I do, check out: