In 2007, a 32-year-old software engineer named Marcus left a well-paying job at a financial services company to build a mobile application for restaurant reservations. He had savings, a co-founder, a working prototype, and the absolute certainty, as he described it later, that this was his moment. He was not wrong that the timing was reasonable. He was wrong about almost everything else: the feature set, the sales model, the target customer, and the unit economics. The business failed quietly fourteen months later. Marcus took a job at another startup, spent three years learning its sales and operations, and started again at 38 with a different idea, more realistic assumptions, and a network built during the interval. The second company was acquired in 2017 for a reported $40 million.
Marcus's story is not the one that gets told about entrepreneurship. The culture of entrepreneurship is saturated with narratives about young visionaries, garage breakthroughs, and the triumph of passion and audacity over conventional caution. These stories are real. They are also wildly unrepresentative of the actual distribution of outcomes. Scott Shane, an entrepreneurship researcher at Case Western Reserve University, spent years compiling what the data actually shows in his 2008 book The Illusions of Entrepreneurship. The picture is considerably more complicated than the mythology suggests, and understanding the actual evidence is the prerequisite for giving entrepreneurship serious thought.
This article examines what peer-reviewed research shows about why entrepreneurs succeed and fail, what traits consistently matter, what role timing and market conditions play, and what the gap between first-time and serial entrepreneurs actually reflects. It is not a discouragement from entrepreneurship. It is an attempt to replace inspiring falsehoods with useful truths.
"The biggest problem with entrepreneurship research is survivorship bias. We study the winners and then draw conclusions about what makes people win. But the losers had most of the same traits." -- Scott Shane, The Illusions of Entrepreneurship (2008)
Key Definitions
Survivorship bias: The logical error of drawing conclusions from a sample that includes only the survivors of a selection process, while ignoring those who failed. In entrepreneurship, studying successful founders without studying failed ones with similar initial characteristics produces systematically misleading conclusions about what causes success.
Effectuation: A decision-making logic identified by Saras Sarasvathy in her 2001 research, contrasted with causation. Causation starts with a goal and works backward to find the means. Effectuation starts with available means (who you are, what you know, who you know) and asks what possible goals those means might allow. Expert entrepreneurs tend to reason effectually rather than causally.
Locus of control: A psychological construct measuring the degree to which individuals believe they control their own outcomes versus being controlled by external forces. Internal locus of control (believing outcomes depend on one's own actions) is consistently associated with entrepreneurial activity and persistence.
Pivot: A structured course correction in which a startup changes a fundamental element of its strategy, often its product, target customer, or business model, while retaining lessons from prior work. Eric Ries popularized the concept in The Lean Startup (2011).
Product-market fit: The degree to which a product satisfies strong market demand. Marc Andreessen, who coined the term in a 2007 blog post, described it as being in a good market with a product that can satisfy that market. Most startup failures can be attributed to its absence.
| Factor | Mythology | What Research Shows |
|---|---|---|
| Age at founding | Young founders disrupt faster | Mean age of high-growth startup founders is 45 (Azoulay et al., 2018) |
| Passion and drive | Passion predicts success | Passion is widespread among failed founders too; weak predictor of outcomes |
| Serial vs. first-time | First-timers bring fresh thinking | Serial entrepreneurs have higher success rates due to accumulated learning |
| Timing vs. execution | "Idea is everything" | Market timing and conditions explain more variance than individual traits |
| Risk tolerance | High risk tolerance required | Expert entrepreneurs minimize risk through affordable loss principle (Sarasvathy) |
| Education | Dropouts succeed, credentials don't matter | Domain knowledge from experience matters; college completion also correlates with success |
The Base Rate Problem
Before examining what makes entrepreneurs succeed, it is necessary to establish honest base rates. The failure statistics cited in popular entrepreneurship culture are often imprecise, sometimes inflated, and routinely misinterpreted. What the serious research shows is this:
The US Bureau of Labor Statistics Business Employment Dynamics data, which tracks a much larger sample than most studies, shows that approximately 50 percent of new employer businesses fail within five years and roughly two-thirds fail within ten years. These are not all technology startups; they include all types of new businesses across industries. Failure rates vary substantially by industry: food service and retail fail at higher rates; professional services and healthcare at lower ones.
Among venture-backed technology startups, the numbers are somewhat different. Shikhar Ghosh at Harvard Business School analyzed 2,000 venture-backed US companies funded between 2004 and 2010 and found that approximately 75 percent failed to return investors' capital. Approximately 95 percent failed to achieve their projected returns. A small number, roughly 10 to 15 percent, produced returns that justify the risk-capital model.
These numbers are not arguments against entrepreneurship. They are arguments for going in with clear eyes. The founder who understands that most businesses fail is better positioned to make the structural changes that increase their specific probability of success than the founder who believes passion and determination are sufficient.
Survivorship Bias and the Knowledge Problem
The fundamental epistemological problem in entrepreneurship research and culture is survivorship bias. When we study Steve Jobs, Elon Musk, Sara Blakely, or Patrick Collison, we are studying individuals who succeeded conspicuously. We are not studying the thousands of founders with comparable intelligence, work ethic, and passion who failed. And we cannot easily learn from that larger population because failure rarely produces books, speaking engagements, or Forbes profiles.
Nassim Nicholas Taleb's work on silent evidence, particularly in The Black Swan (2007), is useful here. He describes a thought experiment: imagine assessing the value of astrology by interviewing successful astrologers. They will have compelling stories, apparent pattern-recognition, and apparent predictive accuracy. You would draw confident conclusions about astrology's efficacy. But you would have no information about the much larger population of practicing astrologers who were less accurate, or who got things wrong in ways that were never reported. The survivors define the sample.
In entrepreneurship, this problem produces confident claims about the causal role of traits that may simply be common to all founders, failed and successful alike. The passion, determination, and high risk tolerance that appear in post-hoc accounts of successful founders appear with similar frequency in the early-stage presentations of founders who never make it. Melissa Cardon at Pace University, who has spent years studying entrepreneurial passion empirically rather than anecdotally, finds that passion is widespread among entrepreneurs at founding and is a poor predictor of eventual success. What matters more is how founders respond when the initial passion meets market resistance.
Saras Sarasvathy and the Logic of Expert Entrepreneurs
In 2001, Saras Sarasvathy, then at the University of Virginia Darden School of Business, published one of the most genuinely original pieces of entrepreneurship research in decades. She gave twenty-seven expert entrepreneurs (each with at least fifteen years of experience and at least one successful exit) the same business case and asked them to think aloud as they worked through it. She then compared their reasoning patterns to those of expert managers and MBA students.
The difference was striking. Expert entrepreneurs did not start from a goal and reason backward to means, the causation logic that business schools teach and that most people assume is the natural approach. Instead, they started from what they had: their own knowledge, skills, and networks; the resources available; the partners who expressed interest. From these available means, they worked forward to ask what possible goals those means might allow, then selected among those goals based on what they found interesting and achievable. Sarasvathy called this effectuation.
The effectuation logic also fundamentally reconceptualizes risk. Causation asks: what is the expected return on this investment? Effectuation asks: what is the most I am willing to lose to learn whether this is viable? This affordable loss principle, as Sarasvathy named it, changes the decision from a prediction problem (which no one can solve reliably) to a commitment problem (which is personally answerable). Expert entrepreneurs do not predict the future; they negotiate it, building partnerships with stakeholders who each commit something, progressively reducing uncertainty through action rather than analysis.
This research has been replicated and extended substantially since 2001. It has practical implications: founders who wait until they have the perfect plan and complete information will wait indefinitely, while those who begin from available means, test quickly, and adjust based on real feedback are engaging in the reasoning pattern that research associates with expert entrepreneurship.
What Age Actually Predicts: The Azoulay et al. Finding
The image of the twentysomething founder is so embedded in technology culture that it functions almost as a prerequisite in some circles. The data does not support it.
Pierre Azoulay, Benjamin Jones, J. Daniel Kim, and Javier Miranda published research in 2018 in the American Economic Review: Insights analyzing the full population of US startups that received venture funding or reached high-growth status, drawing on IRS records and Census Bureau data for all new businesses between 2007 and 2014. Their finding: the mean age of founders of the fastest-growing new companies in their sample was 45. The highest-performing startups, those in the top 0.1 percent of job creation or revenue growth, had even older founders on average. Founders aged 40 to 49 created companies that were significantly more likely to achieve high growth than those founded by people in their twenties.
The mechanisms the researchers identified are plausible and consistent with what we know about expertise development. Prior industry experience provides pattern recognition that reduces certain categories of error. Professional networks built over decades provide access to capital, talent, and customers. Accumulated knowledge of industry economics and competitive dynamics allows more realistic planning. And psychological factors may matter: the research on cognitive aging suggests that certain forms of judgment, particularly those involving integration of complex, ambiguous information, improve through middle age rather than declining.
None of this means young founders cannot succeed spectacularly, as the genuine examples confirm. It means the modal successful founder looks different from the cultural archetype, and that experience is a genuine asset rather than a liability to be apologized for.
Serial Entrepreneurs and the Learning Advantage
Paul Gompers, Anna Kovner, Josh Lerner, and David Scharfstein published research in 2010 in the Journal of Financial Economics examining the performance differences between first-time and serial entrepreneurs in venture-backed companies. The results produced a nuanced picture.
Serial entrepreneurs who had previously founded a successful company had a success rate (defined as IPO or profitable acquisition) of approximately 30 percent on their next venture. First-time entrepreneurs had a success rate of approximately 21 percent. But the more interesting finding concerned failure: serial entrepreneurs who had previously failed had a success rate of approximately 22 percent, essentially indistinguishable from first-timers statistically, but dramatically higher than one might expect if failure simply repeated itself. The implication is that the learning is real, but takes more than one failure to fully transfer.
What do serial entrepreneurs learn? The research and practitioner accounts converge on several themes. They are more realistic about timelines: most first-time founders dramatically underestimate how long sales cycles take, how slowly large organizations make decisions, and how much runway runway actually requires. They are faster to recognize when a strategy is not working, having experienced the feeling of pushing against a market that simply does not want what you are selling. They are more deliberate about early hiring, having seen the cost of early hires whose values or work styles do not fit the culture being built. And they have better networks, which matters for both fundraising and recruiting.
Reid Hoffman's concept of blitzscaling, described in his 2018 book of the same name with Chris Yeh, addresses a different dimension of serial advantage: the understanding of when to prioritize speed over efficiency. Blitzscaling argues that in certain market structures, specifically winner-take-most markets enabled by network effects, the competitive advantage of growth speed justifies accepting high operational inefficiency and risk. This is a very specific strategic logic applicable in specific circumstances, not a general principle. Serial entrepreneurs are better at recognizing which circumstances warrant it.
Market Timing and the Bill Gross Analysis
Bill Gross, founder of Idealab, a startup studio that has launched hundreds of companies, analyzed the factors behind the success and failure of two hundred companies in his portfolio and presented the findings in a widely viewed 2015 TED talk. He assessed each company on five factors: idea, team, business model, funding, and timing.
Timing, defined as whether the market was ready for what the company was offering, was the single most important factor, explaining approximately 42 percent of the difference between success and failure. Team and execution came second at approximately 32 percent. The idea itself was third, at 28 percent. Business model and funding were less predictive than any of these.
Gross illustrated with examples. Airbnb succeeded partly because the 2008 financial crisis made people genuinely receptive to renting out a room for income. YouTube succeeded because broadband had just reached sufficient penetration to make video streaming practical. Pets.com failed because the internet had not yet developed the logistics infrastructure to make its model work economically. These were not failures of idea or team; they were failures of timing.
This finding is well-supported by academic research. A market that does not yet exist cannot be created through better execution alone. Timing is partly about luck, but partly about the disciplined analysis of whether the enabling conditions for a business model, technological, economic, regulatory, or cultural, are actually present.
The Pivot and the Lean Startup Evidence
Eric Ries published The Lean Startup in 2011, synthesizing ideas from Toyota's lean manufacturing, Steve Blank's customer development methodology, and his own experience as a founder. The book's central contribution was operationalizing the idea that startups are not small versions of large companies but rather organizations searching for a repeatable and scalable business model under conditions of extreme uncertainty.
The build-measure-learn feedback loop that Ries describes, deploying small experiments quickly and adjusting based on real customer behavior rather than assumptions, is now standard vocabulary in technology entrepreneurship. The pivot, a structured course correction in which a fundamental element of the strategy changes while learning from prior work is retained, became the mechanism for distinguishing adjustment from failure.
The empirical evidence for lean methodology is mixed but directionally positive. A 2014 study by Ernesto Dal Bo, Frederico Finan, Nicholas Li, and Laura Schechter found that lean-style iterative testing produced better outcomes in a randomized controlled experiment in the development economics context. Research by Thomas Eisenmann at Harvard Business School on startup failure found that premature scaling, building out infrastructure and hiring before achieving product-market fit, is one of the most reliably fatal errors, and that the lean methodology's emphasis on validation before scaling directly addresses this.
The core insight that survives scrutiny is simple: evidence beats assumptions. A startup that learns from real customer behavior before building the full product will make fewer expensive errors than one that executes a detailed plan based on assumptions about what customers want.
Gender, Funding, and Structural Disadvantage
Any honest account of entrepreneurship outcomes must address a structural factor that research has documented thoroughly: the gender gap in venture funding is large, persistent, and cannot be explained by quality differences in the underlying companies.
Paul Gompers and Sophie Wang published research in 2017 finding that women received only about 8 percent of venture capital invested in the United States, despite founding businesses that, on comparable metrics, performed as well or better. Dana Kanze and colleagues published a 2018 study in the Academy of Management Journal analyzing Q&A sessions at a TED-style pitch competition and found that investors asked men predominantly promotion-focused questions (about gains and opportunities) and women predominantly prevention-focused questions (about losses and risks). When founders of either gender responded in promotion-focused terms to all questions, funding outcomes improved. The asymmetry in questioning biased outcomes against women not through explicit discrimination but through the interaction between question framing and response.
The structural nature of this disadvantage means that strategies directed entirely at individual founder behavior and traits, while valuable, are incomplete. The entrepreneurship ecosystem's failure to direct capital proportionate to merit is both an equity problem and an inefficiency problem: it systematically underinvests in a segment of founders whose outcomes the data suggests would justify investment.
Practical Takeaways
Get specific about the market before the product. The most common fatal error documented in startup post-mortems is building something the market does not want badly enough. Talk to potential customers before building. Ask about their existing behaviors and the problems they are currently trying to solve, not whether they would use your hypothetical product.
Take the base rates seriously. Knowing that the majority of new businesses fail, and that the survival rate in your specific industry is what it is, is not demotivating. It is the information needed to make the structural choices (adequate runway, realistic timelines, early validation) that improve your specific probability.
Think effectually when resources are limited. Start from what you have and what you know. The founder who waits for the right conditions will wait forever; the founder who begins from available means and builds iteratively is engaging in the reasoning pattern that research associates with expert entrepreneurship.
Prior experience is an asset, not a delay. The Azoulay et al. data is clear: industry experience, networks, and accumulated pattern recognition are genuine predictors of success. If you lack them, building them through employment in your target industry before starting is a legitimate strategy.
Write down the affordable loss, not just the expected return. Before committing time and capital, define what you are willing to lose to find out whether the opportunity is real. This converts an unanswerable prediction problem into an answerable commitment question.
Study your failures as carefully as your successes. The serial entrepreneur advantage is largely a learning advantage. Post-mortems of what did not work, conducted honestly and in writing, are the mechanism through which experience transfers into improved future judgment.
References
- Shane, S. (2008). The Illusions of Entrepreneurship. Yale University Press.
- Sarasvathy, S. D. (2001). Causation and effectuation: Toward a theoretical shift from economic inevitability to entrepreneurial contingency. Academy of Management Review, 26(2), 243-263.
- Azoulay, P., Jones, B. F., Kim, J. D., & Miranda, J. (2018). Age and high-growth entrepreneurship. American Economic Review: Insights, 2(1), 65-82.
- Gompers, P., Kovner, A., Lerner, J., & Scharfstein, D. (2010). Performance persistence in entrepreneurship. Journal of Financial Economics, 96(1), 18-32.
- Gompers, P. A., & Wang, S. Q. (2017). Diversity in innovation. NBER Working Paper 23082.
- Kanze, D., Huang, L., Conley, M. A., & Higgins, E. T. (2018). We ask men to win and women not to lose. Academy of Management Journal, 61(2), 586-614.
- Ries, E. (2011). The Lean Startup. Crown Business.
- Cardon, M. S., Wincent, J., Singh, J., & Drnovsek, M. (2009). The nature and experience of entrepreneurial passion. Academy of Management Review, 34(3), 511-532.
- Ghosh, S. (2012). The Venture Capital Secret: 3 Out of 4 Start-Ups Fail. Harvard Business School Working Paper.
- Hoffman, R., & Yeh, C. (2018). Blitzscaling. Crown Business.
- Graham, P. (2009). Founder Mode. paulgraham.com.
- Taleb, N. N. (2007). The Black Swan. Random House.
Related reading: founder mythology examined, disruption rhetoric examined, how to make better decisions
Frequently Asked Questions
What percentage of startups actually succeed?
The statistics vary depending on definition and time horizon, but they are consistently sobering. Scott Shane, an entrepreneurship researcher at Case Western Reserve University, analyzed data from the US Bureau of Labor Statistics in his 2008 book The Illusions of Entrepreneurship and found that approximately 50 percent of new businesses fail within five years and roughly 66 percent within ten years. Among venture-backed startups, the failure rate is even higher: research by Shikhar Ghosh at Harvard Business School found approximately 75 percent of venture-backed companies fail to return investors' capital. These numbers are not widely advertised in entrepreneurship culture, which tends to select and amplify the most dramatic successes while rarely examining the distribution of outcomes systematically.
What traits do successful entrepreneurs have in common?
The research is more nuanced than popular accounts suggest. Genuine consistent findings include high tolerance for ambiguity and uncertainty, strong internal locus of control (the belief that outcomes are determined by one's own actions rather than external forces), and above-average propensity to take calculated rather than reckless risks. Melissa Cardon and colleagues' research on entrepreneurial passion found that passion matters but is neither sufficient nor the primary predictor of success. Meta-analyses of personality research find modest but consistent effects for conscientiousness, openness to experience, and emotional stability. Paul Graham's essays on founder traits, drawing from Y Combinator's experience funding thousands of startups, emphasize determination above intelligence and the ability to learn and adapt over having the right initial idea.
Is entrepreneurship more about talent or timing?
Both matter, but timing is systematically underweighted in popular accounts that focus on founder heroics. Bill Gross, founder of Idealab, analyzed the factors behind 200 startups and found that timing, defined as whether the market was ready for the product, was the single largest predictor of success, explaining approximately 42 percent of the difference between success and failure, more than team, idea, business model, or funding. However, timing interacts with execution: identifying the right moment is only valuable if the team can move quickly enough to capture it. Saras Sarasvathy's effectuation research suggests that expert entrepreneurs think differently about timing, working with what is available rather than predicting and planning for an ideal future state.
What does research say about the best age to start a company?
The popular image of the twentysomething founder is substantially at odds with the data. Pierre Azoulay, Benjamin Jones, J. Daniel Kim, and Javier Miranda published research in 2018 in the American Economic Review: Insights analyzing the complete population of US startups and finding that the average age of founders of the fastest-growing new companies was 45, not 25. Among the very top performers, those in the top 0.1 percent by growth, the average founder age was even higher. Prior industry experience, professional networks, and the accumulated pattern recognition that comes with years in a domain appear to be significant assets that younger founders typically lack. The Zuckerberg and Gates narrative is real but deeply unrepresentative of the actual distribution.
How important is passion versus market opportunity?
The research suggests market opportunity is the more reliable foundation. Melissa Cardon and colleagues' work on entrepreneurial passion shows that passion for the work increases persistence and energy, which can matter in early stages, but passion for a product in a market with insufficient demand does not overcome the structural problem. Scott Shane argues in The Illusions of Entrepreneurship that most small businesses fail not because founders lack passion but because they enter industries with unfavorable economics or underestimate competition. Paul Graham's view, synthesized from thousands of funded startups, is that founders should work on things they find genuinely interesting but that the market test, whether customers want the product badly enough to pay for it, is the central question no amount of passion can substitute for.
What do first-time versus serial entrepreneurs do differently?
Research by Paul Gompers, Anna Kovner, Josh Lerner, and David Scharfstein on venture-backed companies found that serial entrepreneurs who had previously founded a successful company had a 30 percent success rate on subsequent ventures, compared to 21 percent for first-time entrepreneurs. Part of the advantage is learning from failure: entrepreneurs who failed in a prior venture did nearly as well as those who had succeeded, both outperforming first-timers. Serial entrepreneurs tend to be more realistic about timelines, more selective about co-founders and early hires, faster at recognizing when a strategy is not working, and better networked with investors and potential customers. They also tend to apply Sarasvathy's effectuation logic more naturally, working iteratively from available resources rather than executing a fixed plan.
What are the most common reasons startups fail?
CB Insights has published regular analyses of startup post-mortems, and the most frequently cited causes cluster around a few themes. No market need, cited in approximately 35 to 42 percent of post-mortems, is consistently the leading cause: building something that not enough people want badly enough to pay for. Running out of cash is cited in roughly 29 percent of cases, but often reflects the underlying product-market problem rather than a purely financial failure. Wrong team is cited in approximately 23 percent. Scott Shane's analysis emphasizes that many failures are predictable from industry selection: businesses entering markets with high barriers to entry, thin margins, or well-established incumbents face structural disadvantages that founder quality cannot easily overcome.