Part 4: Prior startup experience may not help much
Over the last while, I've frequently heard that serial entrepreneurs have a higher probability of success. Mistakenly, though, I've taken this statement to mean that entrepreneurs improve by learning from past experiences. Serial entrepreneurs do have a higher rate of success, but their success may have more to do with them having the raw talent than learning while on the job. This is disappointing because, in a way, it could indicate that entrepreneurship is hard to learn. However, if startups are highly uncertain environments then it makes sense to me that learning in such an environment should be difficult (see Part 0).
Elon Musk, a serial entrepreneur known for X.com, Tesla, Solar City and SpaceX. Credit: Dublin, Wednesday 31th October 2013: Pictured at the The Web Summit 2013, RDS. Photo by Dan Taylor/Heisenberg Media
Truth be told, there are a limited number of studies on the success rate of serial entrepreneurs. The most relevant in my opinion, in the startup context, is a 2008 study from the Harvard Business School. That study considers thousands of entrepreneurs who received at least one round of venture capital funding. Defining success as an initial public offering (IPO), the study examines the success rates of one-off and serial entrepreneurs who raised venture capital.
A sub-set of the results may be summarised as follows:
First time entrepreneurs have a success rate of 20.9%.
Serial entrepreneurs who failed in their prior venture had a 22.1% rate of success.
Serial entrepreneurs who had a track record of prior success had a 30.6% rate of success in their next venture.
The first thing to say here is that, for sure, serial entrepreneurs have higher success rates (roughly 50% higher). The second thing to say is that it's very hard to know if these higher success rates are because serial entrepreneurs are naturally talented or because they have learned something from prior success. If talent didn't matter, and only learning mattered, then the success rate for serial entrepreneurs should be higher and it should be the same regardless of whether their previous venture failed or succeeded. Clearly, this is not the case, so we can say that raw talent, for sure, has an important role. By contrast, if only talent had a role and learning were unimportant, then the success rate should be higher for entrepreneurs with prior success then those with prior failure. This fits the data. Unfortunately, the case where both learning and talent are important could also lead to a scenario where the success rate is higher for entrepreneurs with prior success than those with prior failure. So, it therefore seems that talent is definitely important and learning is possibly important.
If I had to draw a conclusion, however, I would lean towards thinking that talent is dominant over learning in determining success. I say this because I would hypothesise that those entrepreneurs who fail and then go back out again (which are few and far between when it comes to raising venture capital again after a first failure) are perhaps amongst the most talented of the entrepreneurs in the initial batch of first time entrepreneurs. Therefore, if their success rate second time round (22.1%) is only slightly above that of all first time entrpreneurs (20.9%), I would question whether they really are learning much from prior experience.*
The question I am posing here is whether serial entrepreneurs have higher success rates because they have learned from prior experience or because they are fundamentally talented. My fear is that educators, mentors and advisors might wish to believe the former, while the latter may be more true. If talent is more important than learning in determining success I think we should consider the implications:
For entrepreneurs, I think this would mean reduced focus on past experience and more focus on heuristics (rules of thumb).
For investors, this means there is potential edge for those who can recognise talent at an early age, before there has been a public signal of a prior success.
All in all, I hope this gets you to think again about whether startups are a learning-by-doing experience. At least when it comes to venture backed startups, I think there is no evidence that entrepreneurs learn by doing but at the same time, there is no clear evidence against.
*On top of the issues I've raised here, the authors consider, in detail, a number of other issues with interpreting the raw data. To some extent, success breeds success in that top VC firms are more likely to back serial entrepreneurs and serial entrepreneurs may be sought out by better employees. These effects can serve to amplify the success rate of entrepreneurs in subsequent ventures and result in an overstatement of the extent to which entrepreneurs are learning from past ventures.