Here’s a game you can play in your own head. Think of the most successful person you know, or know of. Someone whose career or life trajectory you find genuinely impressive. Now ask yourself, honestly: how much of their success is because they are exceptional, and how much is because they got lucky?
Most people, when asked this about someone else, give a more honest answer than they would give about themselves. We can see other people’s luck with a clarity we rarely turn on our own lives. Yours feels earned; theirs looks like they caught a wave. And beneath this asymmetry — which is widely shared and, it turns out, predictable — sits one of the more uncomfortable findings in social science: luck plays a far larger role in who ends up where than our narratives of success generally admit.
The economist who measured luck
The most careful modern treatment of this comes from the economist Robert Frank at Cornell, whose 2016 book Success and Luck marshalled a surprising range of evidence. Frank’s argument isn’t the trivial claim that luck exists — everyone agrees with that. His argument is stronger and stranger. In modern winner-take-all markets, small initial advantages compound into enormous differences in outcome. Two people of roughly equal talent, working equally hard, can end up orders of magnitude apart in visible success, based on tiny differences in starting conditions that nobody at the time recognised as decisive.
Frank draws on a range of data. In hiring for senior positions, the difference in objective qualifications between the top several candidates is usually small — in many cases indistinguishable from measurement noise. But the candidate chosen gets the job, the salary, the reputation boost, and the platform from which to build further success. The next-best candidate, equally qualified, takes a slightly less prestigious role. Over twenty years, those two initially similar careers can diverge dramatically, not because one person was really better, but because a series of small advantages compounded.
The same pattern shows up, starkly, in the arts and in entrepreneurship. Many debut novels are genuinely good; a few become bestsellers; most don’t. The difference between the ones that succeed and the ones that don’t is usually not quality — it’s timing, reviewer attention, algorithmic amplification, or the accident of a small early signal that cascaded. This isn’t comforting for the novelists who didn’t succeed. But it’s worth saying clearly: being overlooked is usually not evidence of not deserving attention.
The experiment that made this visible
One of the most influential pieces of evidence for Frank’s argument comes from a study by the sociologist Duncan Watts at Columbia, who built a clever experiment he called the Music Lab.
Watts recruited about 14,000 teenagers and had them listen to and rate songs by unknown bands. Crucially, he divided the participants into two conditions. In one, participants saw the songs listed without any social information — just titles and artists. In the other, they saw how many other participants had already downloaded each song, meaning early popularity was visible.
Watts also did something sneaky. He ran the second condition as eight parallel worlds, each populated by a different random sample of listeners. Each world started with zero downloads; whatever happened in each world happened independently.
Two findings emerged. First, in every world where social information was visible, songs that did well early — because they happened to catch the first few listeners — kept doing well, snowballing into hits. Songs that did poorly early, even if they were objectively good, stayed obscure. Second, and more striking, the songs that became hits were mostly different across the eight worlds. One song would become the number-one hit in one world and rank 40th in another. Same song, same listeners drawn from the same pool, entirely different outcomes — driven entirely by the accident of which few listeners happened to click first.
The only world where there was any consistency about which songs did well was the one without social information. The “quality” of a song, stripped of social proof, produced somewhat similar rankings across trials. The moment social proof was introduced, small random fluctuations became everything.
Watts’s result was sobering. It suggested that in any market with strong network effects — music, books, films, viral videos, certain products, certain companies — the specific winners of the market are much less predictable, and much less meritocratic, than we tend to assume. Somebody was going to be the hit. That it was this specific one, rather than one of the many other plausible candidates, is partly accident.
The physicist who studies success
A different angle comes from the physicist and network scientist Albert-László Barabási, who has applied the tools of complex-systems analysis to the question of how success actually unfolds across a career.
Barabási and his collaborators have found that scientific careers — a domain where outputs can be measured and where success is relatively objective — follow patterns that are surprising. The most influential paper of a scientist’s career tends to come not from their most productive period, nor from the middle of their career when they have the most experience, but from a sequence of highly variable attempts whose individual success can’t be predicted in advance. The scientists who become famous aren’t the ones who wrote one brilliant paper early; they’re the ones who kept writing, many of whose papers were forgettable, until one of them happened to align with a field that was ready to receive it.
Barabási calls this the random impact rule: impact is distributed randomly across a career, not concentrated at any particular stage. His research suggests that luck — in the form of timing, the field’s readiness, and the unpredictable cascades of attention — plays a larger role in which specific works become influential than the scientist’s overall talent does.
But here’s the subtle point. Barabási’s research also shows that producing many works matters enormously. Scientists who write only a few papers rarely have any become influential, regardless of quality. Those who produce steadily, across decades, have more shots — and the hit rate of any individual paper is so random that more shots genuinely matters. Talent, in this framework, is necessary but not sufficient; production volume multiplies your chances at the lottery.
The counter-view: talent and preparation still do work
Not every researcher has been persuaded that success is mostly luck. The writer Eric Barker and the psychologist Angela Duckworth have both argued, from different angles, that while luck matters, it mostly determines who wins among the already-talented and prepared. The general population doesn’t get lucky into careers as novelists or professors or chief executives. Those who get lucky are the ones who had already shown up many times, practised the craft, built the skills, and put themselves in positions where luck could find them. Luck, in their framing, selects among the prepared; it doesn’t select the unprepared.
This seems broadly right. The novelist who wins the literary prize after writing ten novels is not lucky in the sense that the prize fell from the sky; the prize fell on the head of someone who had produced ten novels, and could only have fallen on someone who had. The engineer who becomes a successful founder had built skills and relationships for a decade before the company took off. The person who “got lucky” with a job offer had spent years developing the capabilities the offer required.
So the honest position is probably this: luck plays a large role in which of the prepared people succeed, but preparation plays a large role in whether you’re in the candidate pool at all. The two explanations aren’t competing — they’re layered. Preparation is the price of admission. Luck is the lottery drawn from the admitted.
What this should change about how you think
The research on luck has some uncomfortable implications, and some useful ones. It’s worth sitting with both.
The uncomfortable: many of the people you admire, envy or measure yourself against did not earn their success as purely or completely as the retrospective narrative implies. Their preparation and talent were usually real, but so was their luck. The biographies tend to smooth the luck away, retrospectively making everything look inevitable. It wasn’t inevitable. If you ran history again, different people would occupy those positions. The talented population is always larger than the successful population.
Also uncomfortable: your own absence of certain visible successes is probably not a reliable signal of lacking talent or having worked less hard. You may have done everything right and caught the wrong wave. Many people in that position spend years diagnosing themselves for deficiencies that aren’t really there, adjusting their careers in response to signals that were mostly statistical noise.
The useful: if luck plays a large role, the rational response is to increase the number of lottery tickets you hold. This means producing more — more writing, more applications, more attempts, more reaches, more projects shipped. Barabási’s research is clear that prolific people, of ordinary talent, outperform sparse people of high talent. Volume isn’t vulgar; it’s the honest response to an unpredictable market.
Also useful: generosity toward other people’s setbacks becomes more appropriate when you understand that their failures are often the same statistical noise that your successes are. The people who didn’t get the thing you got weren’t usually less talented than you. They were, statistically, you in a world where the coin came up differently.
The question that remains
The deepest thing the research on luck teaches is a kind of humility that is hard to hold while staying motivated. If success is partly luck, what does it mean to take credit for what you have? If failure is partly luck, what does it mean to accept blame for what you don’t?
The honest answer is probably that both credit and blame should be softer than our culture encourages them to be. Your preparation was real; your luck was also real; the combination is what produced you. Other people’s combinations came out differently. Judging anyone’s life — including your own — on its visible outcomes is a kind of confusion about what those outcomes actually measure.
The question to hold is one that, once you’ve considered it, you may not want to fully answer:
If you ran your life again from the start with the same temperament, intelligence and effort, what do you think would have happened — and how much of what you’re proud of would still be there?
Key research referenced: Robert Frank, Success and Luck (2016); Duncan Watts’s Music Lab experiment (Salganik, Dodds and Watts, Science, 2006); Albert-László Barabási, The Formula (2018); Angela Duckworth and Eric Barker’s counter-arguments on the role of preparation.