In a previous post, I discussed how high-risk, high-reward careers can be a better deal for those who want to do good: if you strike it rich, buying a tenth car will add very little to your personal quality of life, but vaccinating a tenth child will help that child about as much as the first one. This matters in practice: most venture-backed startups fail, but the average (mean) financial gain to founders is measured in millions.
However, it would be a mistake to think of the returns to entrepreneurship as predictably stemming from just showing up and taking a spin at the wheel of startup roulette. Instead, entrepreneurship is more like poker: a game where even the best players cannot predictably win over a single night, but measurable differences predict that some will earn much more than others on average. By paying attention to predictors of entrepreneurial success (whether good news or bad), you can better tell whether you have a winning hand or should walk away for a different game. And even if the known predictors don’t bear on your own situation, knowing about these predictors can dispel the “lottery illusion”, and can let you know that success is not magic, and that it is worth investing in skill, hard work, strategy, and an understanding of the game.
Let’s take a look at some of those predictors…
Does your invention make business sense?
In Canada, the university of Waterloo is something of an “MIT of the north,” an engineering-intensive school where many students create innovative technologies and sometimes go on to build successful businesses out of them, most famously Mike Lazaridis the creator of the BlackBerry. To help would-be inventors from Waterloo and elsewhere, the government offers an Innovator’s Assistance Program (IAP) to evaluate innovations and their potential for commercialization (e.g. by patenting and licensing, creating a business to sell the technology, etc). Since a huge number of innovators have used the program and outcomes of participants are tracked, a number of academic studies have used this dataset, and found that the program did impressively well in predicting inventor success. This was highlighted in Daniel Kahneman’s recent book “Thinking, Fast and Slow.”
The program assigned numerical scores for a number of categories (market demand, existing competition, difficulty of manufacturing, etc) and combined them to form letter grades from A to E. In a sample of over a thousand inventions, the 2% of inventions with the highest grade were commercialized more than half the time, while the 15% with the lowest grade were never commercialized. Inventions with B and C grades were commercialized about four times as often as D grade inventions (a majority of the total), and 2.7 times as often considering only inventors who continued work after learning their grade.
Percent of all
Percentage that continue
A – recommended for
B – may go forward, but
need to collect more data
C – recommended to go
forward, returns likely
D – doubtful, further
E – strongly recommended
to stop further
Source: Åstebro, T. and Gerchak, Y. (2001) ‘Profitable Advice: The Value of Information
Provided by Canada’s Entrepreneur’s Assistance Program’, Innovation and New
Technology 10(1): 45-72.
Aside from the surprisingly high predictive power of the test (and the high success rate of “A” grade inventions), one thing that stands out is the overconfidence of those receiving the lowest grades: almost half of those with “E” grades persevere in attempting to commercialize their inventions, even though every single one fails. Kahnemann uses this dataset to highlight unrealistic entrepreneurial optimism, even as it displays the accuracy of the “Critical Factors Assessment” test, a simplified version of which is available for free online. If you can update on negative as well as positive information, you will be ahead of the game.
Past start-up success predicts future start-up success
One of the most damning facts about the investment management industry is that, for the vast majority of funds, past returns have almost no correlation with future returns. In other words, most of the skilled professionals in that industry are doing no better than chance for their investors, and worse than that after their fees are taken into account. How does the situation compare for startup entrepreneurs?
One attempt to tackle this question comes from a 2006 paper by Gompers, Kovner, Lerner, and Schwartzstein (2006). They use data on companies receiving venture capital funding between 1975 and 2000, and contrast entrepreneurs receiving venture capital funding for the first time, with serial entrepreneurs who had received venture investments in previous startups. They then measure “success” by whether the firm had made an initial public offering by 2003. Their raw data show that while 25.3% of first-time VC-backed entrepreneurs reach a successful IPO, 29.0% of the serial entrepreneurs do on their second try, and serial entrepreneurs who succeeded the first time are substantially more likely to succeed the second time than serial entrepreneurs who failed first.
When the authors go on to match firms, based on variables such as firm age, they compare the chance of IPO for a firm with typical characteristics on these axes, save for the past experience of the entrepreneur. The chance of IPO 30% with a previously successful (VC-backed) entrepreneur, 20% with one who has previously failed, and 18% with a first-time entrepreneur. The paper also finds that performance differences between experienced and inexperienced venture capitalists are greatest with respect to first-time entrepreneurs and first-time-failed entrepreneurs, but small with respect to entrepreneurs who have previously succeeded, i.e. that expert VCs have some skill in identifying “diamonds in the rough.” (The fact that it is possible to develop such skill indicates that there are identifiable differences between success-prone and failure-prone startups.)
If those numbers seem low for companies that have already received VC funding, you’re right: they don’t include companies that were acquired rather than conducting an IPO; see the Woodward and Hall paper I discussed in my last post on entrepreneurship for do-gooders, for more inclusive numbers (with first-time founding teams exiting with at least $1 million a third of the time). the authors claim that including acquisitions would give results “qualitatively similar” to the aggregate results, so readers would do well to assume the effect sizes are at least somewhat smaller (given researchers’ tendencies to present data in the most interesting light). 1
Tips from the right tail: how smart is Bill Gates?
Physicist and polymath Steve Hsu offers another angle on predictors of entrepreneurial success: look at the very most extreme examples of entrepreneurial success and note their deviations from the norm. In that post he works his way down the first three slots of the 2009 Forbes magazine list of the world’s richest people, finding Bill Gates, Warren Buffett, and Carlos Slim, and considers strong evidence that they are easily in the top 1%, and perhaps much higher:
Bill Gates scored 1580 on the pre-1995 SAT. His IQ is clearly >> 145 and possibly as high as 160 or so.
Warren Buffett graduated high school at 16 ranked in the top 5 percent of his class despite devoting substantial effort to entrepreneurial activities. Most people who know him well refer to him as brilliant, that folksy quote above notwithstanding. I would suggest the evidence is strong that his IQ is above 135, perhaps higher than 145.
Carlos Slim studied engineering and taught linear programming while still an undergraduate at UNAM, the top university in Mexico. He reportedly discovered the use of compound interest at age 10. I would suggest his IQ is also at least 135.
So it would appear that the three richest men in the world all have IQs that are higher than 90 percent or even 99 percent of the > 120 IQ population. (Relative to the general population they are all likely in the 99th or even 99.9th percentile.) The probability of this happening in the Igon Model (on which cognitive ability above the 90th percentile has little impact on entrepreneurial success) is less than 1 in 1000; (i.e. intelligence matters, even at the high end.)
Other demographic statistics from Forbes’ various “Richest X” lists can help shed light on the importance of education, parental success, and other indicators.
Eyes of the Incubators
Many venture capitalists invest relatively late in the lifecycle of a startup, after it has proven itself in a number of ways. But tech incubators like YCombinator and TechStars invest in numerous very early-stage software/web startups, enough to gain significant expertise, and one that tentatively appears to be confirmed in the results of their investments, which seem comparable to those of funds with later investment schedules. The choices of these organizations can give interesting information for prospective altruist entrepreneurs.
This article examines the Linkedin social network profiles of entrepreneurs backed by these incubators to determine the courses they studied and the universities they attended. About 50% studied Computer Science, 14% Engineering, 4% Physics, 3% math, with almost all the remainder taking social science or humanities degrees. Around 27% attended what the authors described as “top schools” in the U.S., meaning members of the Ivy League, Stanford, and the Massachusetts Institute of Technology. However, this is understates the representation of selective academic programs, since it excludes a number of universities known similarly high-quality student bodies or excellence in computer science, e.g. Caltech, Swarthmore and other top liberal arts colleges, and Carnegie Mellon.
One can also consult the public statements of the incubators, although obviously such statements are biased by the need to conceal business secrets of the selection process, and the desire to encourage entrepreneurs to apply to their programs. For instance, in this interview Ycombinator founder Paul Graham discusses (in addition to high intelligence and skill as a hacker) determination and aggressiveness. If one can arrange a private and frank discussion with such an investor, that will give feedback which is hard to beat for accuracy.
Value of Information
Someone considering entrepreneurship as a way to do good has reason to care about the expected value of their income, including the chances of big success. Taking into account factors like the output of the CFA test for inventors, the results of psychometric tests, past track records, and the (honest) estimates of skilled venture capitalists can multiply or divide the expected value of that course by several fold. That’s reason enough to go out of your way to gather such info, both about yourself and your prospective start-up, when weighing it as a career and as a way to do good.
We might also worry about the potential for various biases and errors to send regressions awry, and the possibility that a good chunk of these effects may reflect access to connections rather than skill, with top VCs being better at supplying assistance to their investments in such matters, while entrepreneurs with past success enjoy easy access to funding. ↩