There’s a quiet distortion built into nearly every success story you’ve ever been told. You hear about the founders who made it, the investors who called it right, the artists who finally broke through. What you don’t hear about is everyone else who tried the exact same thing and quietly disappeared. That gap between what we see and what actually happened has a name: survivorship bias. It shapes how we think about risk, talent, and what it takes to succeed, often in ways we don’t notice at all.
What Survivorship Bias Actually Means
Survivorship bias is a form of sampling bias that can lead to overly optimistic beliefs because multiple failures are overlooked, such as when companies that no longer exist are excluded from analyses of financial performance. The concept sounds simple, but its effects run deep. It occurs when a successful subgroup is mistaken as the entire group, due to the invisibility of the failure subgroup.
Survivorship bias occurs when researchers focus on individuals, groups, or cases that have passed some sort of selection process while ignoring those who did not. The failures don’t fill conference stages or get written up in magazines. They just vanish from the record. In his book The Black Swan, financial writer Nassim Taleb called the data obscured by survivorship bias “silent evidence.”
The World War II Bomber: Where the Concept Was Born
During World War II, statistician Abraham Wald described methods of estimating the vulnerability of various parts of an aircraft based on damage to surviving planes. The Statistical Research Group at Columbia University, of which Wald was a member, issued eight memoranda on methods of analyzing data obtained from damaged combat aircraft. Given that there was only data from surviving aircraft, this work directly addresses the issue of survivorship bias.
An engineer realized that the worst-hit planes never came back. So the spots showing no damage on surviving planes were actually the worst parts to be hit. Planes hit there never came back, but because they only focused on data from the returning planes, researchers were initially misled by survivorship bias. It’s a powerful illustration of how invisible data can be more important than the data you have in front of you.
The Startup World’s Most Uncomfortable Truth
While you may have heard that 90% of startups fail, recent data from the U.S. Bureau of Labor Statistics reveals a more nuanced picture: approximately 21.5% of private sector businesses fail within their first year, 48.4% within five years, and 65.1% within ten years. The overall picture, though, is still bleak. Global startup failure rate statistics show that 90% of businesses fail at some point in their lifecycle.
While startup failure rates are generally consistent across sectors, some industries show unique patterns: technology has a 63% failure rate within five years, the highest of any industry. In financial technology, the situation is worse. Three out of four fintech startups supported by investors don’t succeed, facing a 75% failure rate. This highlights the challenges these businesses encounter navigating complex regulatory requirements, establishing trust in the financial sector, and competing with established financial institutions. Yet the ones that do survive receive nearly all the media coverage.
Although about 100% of headlines on startup funding cover venture capital, only 0.05% of startups raise venture capital. The rest of the businesses raise the money for their startups through banks and loans, friends and family, credit cards, angel investors, and personal savings. That alone tells you how distorted the public picture of entrepreneurship really is.
What Mutual Fund Data Has Been Hiding
A study that exhibits survivorship bias can skew the returns positively as it only considers mutual funds that are currently in existence. If we only considered the funds that are still active, the average return calculated would be 9%. By contrast, if the study included all possible observations that met its criteria, the calculated average returns would be only 3%, two-thirds less than the return calculated under survivorship bias.
For actively managed US equity mutual funds over the period from 1991 to 2020, survivorship bias overstates the median fund alpha by 0.60% per year: the median fund alpha is negative 0.84% per year among surviving funds compared to negative 1.44% per year among both surviving and non-surviving funds. That might sound like a small number, but compounded over decades of investing, it adds up significantly. Ignoring the impact of mutual funds that do not survive means that a seemingly stellar track record may be an incomplete and biased reflection of reality.
The “Oldest People” Problem in Health Science
Dr. Bradley Elliott, Senior Lecturer in Physiology at the University of Westminster, explains that asking centenarians for their longevity secrets is not reliable due to survivorship bias. Survivorship bias occurs when conclusions are drawn from a group that has survived, while ignoring those who didn’t make it, giving skewed or incomplete insights.
It’s a cliché of reporting on people who reach 100 years of age to ask some variation of the question “What did you do to live this long?” Inevitably, some interesting and unexpected answer is highlighted: fish and chips every Friday, drinking a glass of strong liquor every day, bacon for breakfast every morning, wine and chocolate. While a popular news story, this is a relatively meaningless question that doesn’t help us understand why certain people have lived so long.
A more reliable approach would be to study people in their 60s over decades to identify factors that contribute to longevity. Asking only those who survived misses everyone who followed the same habits and didn’t make it to 100.
When Medical Research Gets It Wrong
Survivorship bias is one of the research issues raised in the provocative 2005 paper “Why Most Published Research Findings Are False”, which shows that many published medical research papers contain results that cannot be replicated. That paper, by John Ioannidis, sent ripples through the scientific community and sparked years of debate about research methodology.
While researchers often attempt to correct for survivorship bias by applying statistical techniques and improved design and analysis, after a study is completed, the process of publishing the study’s findings may suffer from a kind of survivorship bias termed publication bias. There is increasing concern in the sciences that many published research findings are false due in part to the effects of publication bias, the tendency of research journals to publish interesting results. Studies that find nothing tend to go unpublished, which means the scientific record is skewed toward positive findings. Those studies that show no evidence of a relationship between phenomena are more often left out of the journal.
Mental Health Research and the Dropout Problem
Research findings reveal significant survivorship bias among longitudinal survey respondents, indicating that restricting analytic samples to only respondents who provide repeated assessments in longitudinal survey studies could lead to overly optimistic interpretations of mental health trends over time. This matters more than it might seem. Survivorship bias can be problematic if individuals who make it past a selection process are different than those who do not. In the context of longitudinal mental health surveys, bias introduced by non-random differences in baseline mental health or mental health trajectories could result from restricting an analytic sample to respondents who consistently participated in surveys, ignoring individuals who dropped out.
Analysing mental health trends among only individuals who consistently respond to longitudinal mental health surveys can lead to overly optimistic interpretations of mental health trends by excluding individuals who less frequently respond to follow-up survey invitations. In other words, the people who feel worst may be the least likely to keep showing up, making overall mental health look better than it is.
The Dropout Founder Myth
Students in business school can recall how “unicorn start-ups” are commonly applauded within the classroom, serving as an example of what students should strive for, an archetypal symbol of success. The famous college dropout narrative follows the same logic. A handful of extremely visible tech founders left university and built billion-dollar companies. The much larger number who dropped out and failed simply aren’t talked about.
Only 1% of startups emerge as unicorn startups like Uber, Slack, Airbnb, and Stripe. Yet these are the stories that dominate business culture and inspire countless imitations. If a startup fund has a portfolio of 100 companies, most of its returns would come from one investment, ideally a unicorn, followed by the nine successful-but-not-huge companies. The 10 successful startups more than compensate for the 90 failures. The implication is that startup investors are searching for the home run and are willing to lose money on most of their investments to find that company. Investors understand this math. Most aspiring founders don’t.
Survivorship Bias in AI and Emerging Technology
Roughly 95% of generative AI pilot projects in enterprises fail to deliver any measurable ROI, with only about 5% yielding a positive return. The failure rate for AI startups reaches 90%, significantly higher than the roughly 70% seen among traditional tech firms. Given the enormous volume of AI investment coverage in recent years, those failure numbers rarely make headlines. Based on startup failure rate statistics, 42% of AI businesses fail due to insufficient market demand, the largest share of any category.
Startup failure statistics in 2025 showed that the cohort of AI startups launched in 2022 burned through $100 million in three years, a cash-burn speed that doubles that of earlier generations. According to private-market investment advisors, 85% of AI startups are expected to be out of business within three years. The survivors will be celebrated loudly. The rest will join the silent majority we never discuss.
How to Think More Clearly About Success Stories
When a study is affected by survivorship bias, we only pay attention to part of the data. This can have a number of consequences, such as overly optimistic conclusions that do not represent reality, leading us to think that circumstances are easier or more likely to work in our favor than they actually are. The antidote isn’t pessimism. It’s a deliberate effort to ask what you’re not seeing.
People may also take unwise risks because they are influenced by the success of others who survived such risks and were not exposed to the hardship caused by those risks; efforts to emulate these people can hamper one’s ability to learn from those who did not succeed. Seeking out the failure stories, the quiet exits, the ventures nobody writes about, is genuinely useful work. In the case of science careers advice, the bias arises because those who manage to stick to their chosen career path are there to advise the next generation of researchers on how to stay in their field. For objective careers advice, talk to those who left science as well as those who stayed.
The big wins are visible by design. They rise, they attract attention, and they get retold. The losses are quiet, numerous, and instructive. Paying attention to both sides of the ledger doesn’t dampen ambition, it just makes it more honest.
