Average Age of Successful Entrepreneurs: What 2.7 Million Census Records Show
By Lucas Palma · Last Updated: October 26, 2026 · Reading time: 11 minutes
The average age of successful entrepreneurs is 45. That number comes from administrative tax and employment records covering roughly 2.7 million American founders, not from a survey or a magazine profile, and it has been repeated in enough headlines that most people who look this up find it within seconds. It is also the least useful thing in the study it came from. Age is not a lever. Nobody reads that figure and becomes 45. What sits one table further into the same paper is a variable you can actually act on — and it quietly contradicts the most common piece of advice given to people starting out online.
Where the Number Actually Comes From
The study is Age and High-Growth Entrepreneurship by Pierre Azoulay, Benjamin Jones, J. Daniel Kim and Javier Miranda, published in American Economic Review: Insights and circulated earlier as NBER Working Paper 24489. The full methodology is in the Census Bureau’s own working paper version, which is the version worth reading if you want to see how the figures were built.
It is not survey data. The authors matched three Census-held administrative datasets: the Longitudinal Business Database, which tracks U.S. establishments; employment histories from the Longitudinal Employer-Household Dynamics programme, which records where people worked before; and IRS filings that identify business owners. The result is a population, not a sample — around 2.7 million founders of U.S. businesses that hired at least one employee, founded between 2007 and 2014.
Two definitions matter before any number means anything:
“Successful” in this paper mostly means upper-tail growth — landing in the fastest-growing 1 in 1,000 new ventures by employment growth over five years. That group contains roughly 1,700 founders out of 2.7 million. The paper also reports a separate outcome: acquisition or IPO, about 4,000 founders.
“Founder age” is age at the point the business was registered, taken from tax records rather than self-reported.
So this is not a study about whether a business works. It is a study about who ends up in the extreme right tail.
The Average Age of Successful Entrepreneurs Rises With the Bar You Set
The headline figure is one point on a gradient. The paper’s own numbers:
| Group | Mean founder age |
|---|---|
| All new employer businesses | 41.9 |
| Fastest-growing 1 in 100 | 43.7 |
| Fastest-growing 1 in 1,000 | 45.0 |
| Acquired or taken public | 46.7 |
The pattern is the finding. The harder the outcome, the older the founder — and the authors state it without hedging: “Our primary finding is that successful entrepreneurs are middle-aged, not young.”
Supporting results from the same paper:
- A 50-year-old founder is about 1.8 times more likely to reach upper-tail growth than a 30-year-old, and about twiceas likely to have a successful exit.
- The relationship between age and successful exit is “monotonically increasing up until about age 60 and declining slightly thereafter.”
- Performance is roughly flat between 25 and 35, jumps after 35, and surges again around 46.
- Founders under 25 “appear to do badly (or rather, do well extremely rarely).”
All of which is interesting and none of which you can do anything about. Which is why the next section is the one that matters.
The Finding That Does the Work
The authors also measured whether each founder had previously been employed in the same industry as the business they started. They could do this because the LEHD data links about 2.5 million founders to roughly 20 million previous jobs, each with its own industry code.
Here is what they found for the chance of reaching the fastest-growing 1 in 1,000:
| Prior employment in the start-up’s industry | Share reaching the top 1 in 1,000 |
|---|---|
| None at the broad (2-digit) level | 0.11% |
| 3+ years, broad match — 24 industries | 0.22% |
| 3+ years, narrow match — 312 industries | 0.24% |
| 3+ years, very narrow match — 1,065 industries | 0.26% |
Three years inside the broad sector doubles the rate. And the paper is explicit about the direction: “the closer the industry match, the greater the success rate,” with success rates, in the authors’ own summary, “rising up to 125%.”
Put next to the age effect, this is the larger and the more actionable of the two. Age you inherit. Industry time you choose.
“The Same Industry” Is Narrower Than Almost Anyone Assumes
The three rows above are not three ways of saying the same thing. They are three different levels of the North American Industry Classification System, and the paper spells out the scale: the 2012 version contains 24 two-digit industries, 312four-digit industries and 1,065 six-digit industries.
Two digits is a continent. “Retail trade” is two digits. So is “information”. Almost any job you have held sits inside one of 24 buckets, which is why that row is the easy one to satisfy.
Six digits is a street. At that level, selling software to hospitals and selling software to restaurants are different industries. Running a bakery and running a coffee shop are different industries. And it is the six-digit row that carries the highest success rate in the table.
This is the part that gets lost when the study is summarised as “experience helps.” Everyone agrees experience helps. What the data says is more specific and much less comfortable: the value is concentrated in experience inside the particular line of business, and it is measured in years, not in familiarity.
Why This Makes Changing Niches Expensive
Now hold that against how online business is usually taught.
The standard advice when something is not working is to change the niche. Try a different market. The product was wrong, the audience was wrong, pick a new one. It sounds like decisiveness, and in a dataset of 2.7 million founders it is the one move that resets the only founder-side variable with a measured effect back to zero.
A distinction worth making, because nobody makes it:
Changing the offer inside the same industry keeps your accumulated years. You learn what the customers actually pay for, which is the entire point of iterating.
Changing the industry does not. If you spent three years in a six-digit industry and then move to an unrelated one, you are back in the 0.11% row with respect to the new one, no matter how much general business skill you carry across.
The data cannot tell you which to do in your specific case. What it can tell you is that the second move has a measurable cost, and that the cost is invisible in the moment, because the thing you lose — industry-specific knowledge of customers, pricing norms, suppliers, regulation, and what tends to break — does not feel like an asset until you need it and no longer have it. If you have changed direction more than once already, the research on how long a skill actually takes to become sellable is the other half of this picture.
Most People Do Not Have Three Years Anywhere
The three-year threshold sounds modest until you check it against how long people stay in jobs.
The Bureau of Labor Statistics measures this. In January 2024, median tenure with a current employer was 3.9 years — down from 4.1 years in 2022 and the lowest reading since 2002. Broken down by age, the gap is large: 2.7 years for workers aged 25 to 34, against 9.6 years for those aged 55 to 64.
So the typical worker in their late twenties has not yet crossed the three-year line at their current employer, let alone accumulated three years in the six-digit industry they would like to start a business in. This is not a character flaw and it is not fixable by reading harder. It is a scheduling problem, and the honest version of it is that the experience usually has to be built deliberately — while employed, on contract, or on the side. If that is your situation, how many hours a full-time job actually leaves is the constraint to measure before planning anything.
There is also a second asset that comes bundled with industry time, and the study does not measure it: the people. Years inside an industry produce a set of former colleagues, clients and suppliers who have seen your work. That matters because customers who arrive through a referral behave differently on margin and retention than customers who arrive cold — which is an advantage someone entering a new industry does not have and cannot buy quickly.
Keep the Base Rates in View
It would be easy to read the experience table as encouraging. It should not be read that way, and the arithmetic is worth doing out loud.
A 0.26% success rate is roughly 1 in 385. The 0.11% rate is roughly 1 in 909. Doubling your odds of a 1-in-909 outcome gets you to a 1-in-385 outcome. Both are rare. Neither is a plan.
And the outcome being measured is not the one most readers care about. Most people starting something want it to survive and pay them. That is a different question, and BLS measures it separately. Of U.S. private-sector establishments born in March 2013, 79.6% were still operating after one year, 50.6% after five years and 34.7% after ten.
Those are two separate questions with two separate answers: about half of new establishments are still trading after five years, while between 0.11% and 0.26% of founders reach the growth outcome this study measures. The age-45 finding and the experience finding both belong to the second question. Anyone using them to reason about the first is using the wrong study.
For scale, Census Business Formation Statistics counts applications likely to become businesses with payroll. That series — published monthly and tracked by the Federal Reserve — recorded about 1.77 million such applications in the twelve months to September 2026, and it is revised every month, which is worth checking whenever you see the figure quoted without a date.
What Transfers and What Does Not
The study splits experience in one way only: same industry or not. In practice there are two kinds of experience, and conflating them is where people go wrong when they judge their own position.
Industry-specific. Who the customers are, what they will pay, which suppliers are reliable, what regulation applies, what the seasonal pattern looks like, which failures are normal and which are fatal. This is what the Census data measures, and it is the part that resets when you change industries.
Portable. Writing, selling, negotiating, managing your own time, reading a contract. These travel with you. They are not in the study’s experience variable at all, which means the study neither credits nor discredits them — a limitation, not a verdict. Some of them have their own evidence base: whether making the first offer in a negotiation helps or hurts you, for instance, has been tested directly, and the answer is less one-sided than the standard advice.
The practical reading: portable skills decide how well you execute. Industry time appears to decide which table you are sitting at.
How to Check Your Own Position Honestly
This takes about fifteen minutes and produces an uncomfortable number, which is the point.
- Find the industry code for the business you want to start. The Census Bureau maintains the NAICS classification and a public lookup; you want the six-digit code, not the broad category.
- Find the codes for your last three employers. Same system. Most medium and large employers can be identified from their main activity.
- Count matching years at each level. How many years do you have where the first two digits match? The first four? All six?
- Write down the honest total. Not adjacent experience. Not “I read a lot about it.” Years employed or contracted inside that code.
If the two-digit total is zero, you are in the 0.11% row for that business. That is not a reason to abandon the idea — the base rates are tiny in every row — but it is a reason to consider the two options the data points at: choose a business inside an industry you have actually worked in, or acquire the time inside the target industry first, including as an employee or contractor, before committing everything to it.
What This Study Cannot Tell You
This matters enough to state plainly rather than bury.
It covers employer firms only. Every founder in it started a business that hired at least one employee. The solo freelance or one-person online business most readers are building is not in the dataset. There is no reason to assume the figures transfer, and the authors do not claim they do.
It is observational, not experimental. Industry experience predicts higher success rates. The study does not and cannot show that experience causes them. Plausible alternative explanations exist, including that people with better information about an industry are also better at choosing when to enter it.
The employment histories exclude Massachusetts, which the authors note as a coverage gap in the LEHD data.
The founding years are 2007 to 2014. That window predates most of the current platform economy. Whether the same gradient holds for businesses built on top of marketplaces and app stores is an open question, not a settled one.
It is United States data. Nothing here establishes the same pattern in Europe, Brazil, India or anywhere else.
And the general caution that applies to everything on this subject: none of this is a forecast about you. It is a description of what happened to 2.7 million American founders in a specific period, measuring an outcome almost none of them reached.
Frequently Asked Questions
What is the average age of successful entrepreneurs? In the Census study, the mean age at founding was 45.0 among the fastest-growing 1 in 1,000 new ventures, 43.7 among the fastest-growing 1 in 100, and 46.7 among founders whose businesses were acquired or went public. The mean across all new employer businesses was 41.9.
Does this mean young founders never succeed? No. It means they are under-represented at the extreme top relative to how much attention they receive. The paper’s own phrasing is that founders below 25 “do well extremely rarely” — rarely is not never, and the dataset contains successful young founders. It argues against youth being a predictor, not against it being possible.
How much industry experience is enough? The study’s own threshold is three years or more, because that is the band it measured. It did not test four years against eight, so there is no evidence here for an optimal amount. What it did test is closeness of match, and closer was consistently better.
Does freelance or contract work count? The dataset is built from payroll employment records, so formally it counts jobs that appear in those records. Whether unpaid, informal or self-employed time in an industry confers the same benefit is outside what this study measured. Treat it as untested rather than as counted.
Is it too late to start a business at 50? On this data, 50 is a better age than 30 for the outcomes measured — a 50-year-old founder was about 1.8 times more likely to reach upper-tail growth. The relationship kept improving to around age 60 before declining slightly. That is a statement about averages across 2.7 million people, not a prediction about any individual.
Does the study prove experience causes success? No, and this is worth being strict about. It shows a strong, consistent association that survives controlling for age. Causation would require an experiment nobody has run.
I have no experience in the industry I want to enter. What should I do first? The two paths the data suggests are picking a business inside an industry you have already worked in, or getting inside the target industry — employed, contracted, or serving its customers in some capacity — before making it your only source of income. Neither is advice about your situation; both are what the gradient in the table points at.
Why does changing niches matter if I am still learning? Because the thing being measured is time inside a specific industry, and an unrelated change sets that to zero. Changing what you sell to the same industry does not. The distinction between the two is the practical content of this study.
Conclusion
The average age of successful entrepreneurs is a real finding from unusually good data, and it is the wrong thing to take from the research. It describes a correlation you cannot act on while sitting one table away from a correlation you can.
What the Census records show is that prior time inside the specific industry roughly doubles the rate of extreme success, that the benefit grows as the industry match gets narrower, and that the measured threshold is three years — slightly longer than the median American job. Against that, the default advice to change niches when progress is slow looks less like decisiveness and more like discarding the only input with evidence behind it.
The base rates remain brutal in every row of the table. Nobody should read 0.26% as a promise. But between two people with the same idea and the same work ethic, the data points at the one who already knows the industry — and that is a position you can choose to build rather than wait to inherit.
I am building a course right here, in public, from exactly this kind of research. Nothing is for sale yet. When there is, it will be built on results that exist rather than promises that do not.
A Note from Lucas
I went into this paper expecting to disagree with it, because “you need experience first” is the sort of thing people say to discourage others. What changed my mind was the six-digit column. I have spent two years circling online business as a category — content, tools, monetisation — and by the study’s own definition that is not three years in any single industry, it is three years in twenty-four of them. That stung, and it is the reason I stopped treating “pick a new niche” as a neutral decision. If you take one thing from this: the cheap pivot and the expensive pivot look identical from the inside.
About the Author
Lucas Palma writes Info Is The Key, where the method is to read the primary document — a government dataset, a platform’s own policy, a peer-reviewed study — and publish what it actually says, including the inconvenient parts. He is not a native English speaker and is building this business in public, which is the editorial angle rather than a credential. He holds no qualification in economics; every figure above is sourced so you can check it.
Sources
- Azoulay, P., Jones, B., Kim, J.D. & Miranda, J., Age and High-Growth Entrepreneurship — NBER Working Paper 24489
- U.S. Census Bureau, CARRA Working Paper 2018-03 — full methodology and tables
- U.S. Bureau of Labor Statistics — Median tenure with current employer was 3.9 years in January 2024
- U.S. Bureau of Labor Statistics, Business Employment Dynamics — 34.7% of establishments born in 2013 were still operating in 2023
- U.S. Census Bureau Business Formation Statistics, via Federal Reserve Economic Data — High-Propensity Business Applications, Total for All NAICS
