पाठशाला Pathshala · मन Man, The founder · Lesson 01 · Start

Why start a company: the honest question before the first one

Every founder can say why they want to start. Few have tested the reason against what the next ten years will ask of them. This is the test, and the reasons that fail it.

Pathshala, The Founder Library · 11 October 2026 · 10 min read

Nobody asks a founder why they started until the company is in trouble, and by then the answer is a story. The useful time to ask is before the first one, when the reason can still be tested against what it will cost. Most reasons fail the test. The ones that pass are quieter than the ones in the pitch.

This lesson prices the commitment in years rather than money, because years are what you will actually spend. It sorts the common reasons into the ones that predict quitting and the ones that hold. And it ends with a page to write and a date to reread it, because the question does not stay answered.

The cost, in years

Paul Graham’s oldest description of a startup is still the most honest one. In How to Make Wealth he calls it a way to compress your whole working life into a few years, and the sentence that follows is the one founders skip: to make a million dollars you have to endure a million dollars’ worth of pain. The compression is real. So is the pain.

The years are more than a few. Carta’s data desk looked at over thirty-one thousand primary fundraises on its platform and found that in 2023 the average company was 1.6 years from incorporation at seed, 3.3 at Series A, 4.9 at Series B and 6.9 at Series C. Those are American companies that succeeded in raising each round. An exit sits beyond Series C. A company that works is a decade; a company that does not is three or four years you do not get back, which is the figure Graham uses in Why to Not Not Start a Startup: if the startup succeeds it will consume at least three or four years. The failures are shorter, but not short.

The years are also not ordinary years. In What Startups Are Really Like Graham collected what founders said had surprised them most. The second item on the list: running a startup is not like having a job or being a student because it never stops. The third: things seem great one moment and hopeless the next, and there is a limit to how much of that a person can take. In Before the Startup he puts it in one line. If you start a startup it will take over your life to a degree you cannot imagine.

So the honest version of the question is not whether you want to start a company. It is whether you want this particular problem to be most of what you think about from the age you are now until the age you will be in seven to ten years, through a period in which your income falls, your weekends go and the outcome is more likely to be failure than not. Put the ages on paper. If you are thirty-one, write forty-one. Then ask the question again.

The reasons that predict quitting

Jessica Livingston, who interviewed founders for Founders at Work and then watched thousands of them pass through Y Combinator, gives the mechanism plainly in her nine lessons from founding YC: build on what genuinely interests you, because the work is hard enough that a lack of interest will make you quit. Everything in this section is a version of that. A reason predicts quitting when it is satisfied, or disproved, before the company is.

Escaping a boss. The boss goes away in month one. In month four the bank balance becomes a worse boss, because it does not take meetings and cannot be persuaded. Founders who left to be free of someone discover that a company is the least free arrangement they have ever been in, and the discovery arrives before any customer.

The title. Founder is a word that costs nothing to acquire and is awarded on day one. A reason fully paid out on day one has nothing left for month eighteen, when the product is late and a customer has churned. It is the reason most often dressed up as something else, and the easiest to test: would you do the same work for the same decade if nobody outside the company ever knew your role?

Money, quickly. Graham’s arithmetic above is the answer. The money comes, if it comes, late, lumpy and after a decade, and in the meantime the founder is usually the lowest-paid senior person in the building. Someone whose reason is money is better served by a well-paid job and a disciplined savings rate, and finds that out when the first funded competitor appears.

A technology rather than a problem. A tool you are excited about is a reason to build a prototype, not a company. The company needs a person with a problem, and when the tool turns out to be the wrong one for that person, a founder who loved the tool leaves and a founder who cared about the person changes the tool. Graham’s advice in Before the Startup is the same thought from the other side: the way to get startup ideas is not to try to think of startup ideas.

Everyone else is doing it. Friends raising rounds, a city full of announcements, a sense that the window is now. The window is real for some ideas and the pressure is real for all of them, but a reason that lives in other people’s timelines ends when those timelines change. A company started because the market was hot is orphaned when the market cools, and the market always cools.

The reasons that hold

The reasons that survive a decade share a shape. They are about the problem and the people rather than about the founder, they are not paid out by any single event, and they are things the founder would be doing some version of anyway.

A problem you cannot leave alone. The founder who has had the problem, has watched others have it, has tried the existing answers and found them wanting, and keeps returning to it in the shower. This is the reason behind most of the companies worth studying, and it is the only one that gets stronger when things go badly, because a bad month teaches you more about the problem.

Customers you already know. Someone who has spent six years selling into hospital procurement or running operations for a logistics company does not need to find the customer; they have the customer’s phone number and the customer’s complaint in their own handwriting. Graham’s counterintuitive list in Before the Startup puts expertise in your users above expertise in startups, and this is why. The founder with the customers in hand can be wrong about everything else and still get to a first sale.

The wish to live by your wits. The last of Graham’s sixteen reasons not to start a startup is that a job is the default, and his answer is that defaults are powerful precisely because they operate without any conscious choice. Some people find that they want the responsibility, the uncertainty and the direct relationship between what they do and what happens, and want it more than they want the structure a job provides. That is a legitimate reason and a durable one, as long as it is tested against the next section rather than assumed.

Curiosity about whether you can. Graham’s fifth counterintuitive point is that you cannot predict whether you will succeed, so you have to try to find out. For some founders that is enough, and it holds because it does not depend on the outcome. It does need to be the real reason rather than the respectable one, and the test is simple: if the answer turns out to be no, will you be glad you found out?

A reason that is paid out on day one has nothing left for month eighteen. Choose a reason that gets stronger when things go wrong.

Graham’s sixteen excuses, and the three to take seriously

In 2007 Graham listed sixteen reasons people give for not starting a startup and argued that most of them are not real: too young, too inexperienced, no idea, no co-founder, no room for more startups, parents want you to be a doctor. His case against each is worth reading in full and takes twenty minutes. But he is honest about three, and so should this lesson be.

A family to support. Graham does not advise anyone with a family to start a startup, not because it is a bad idea but because the downside is carried by people who did not choose it. His softer suggestions are a consulting business that can become a product company, or joining someone else’s startup first. In India this reason weighs more, not less. Households are more often joint in their finances, parents are more often dependants, and a single income supports more people. The lesson on [quitting your job](/library/quitting-your-job-timing-savings-family) treats it as the central question rather than a footnote, which is where it belongs.

Not ready for the commitment. Three or four years minimum, and in practice closer to the decade above. This is not a character flaw. It is information. Someone who knows they want to study further, move countries or have a child in the next two years is not wrong to wait, and a company started by someone who is half committed is worse for everyone in it than no company at all.

Fear of uncertainty. Graham’s answer is that if you start a startup it will probably fail, and that nobody will blame you so long as you made a serious effort. The first half is true everywhere. The second half was written about Silicon Valley in 2007, and a founder in India should check it against their own industry and their own family rather than take it on trust. In most of the Indian technology economy a well-run failure is now read as experience. In some professions and some families it is not yet, and pretending otherwise does not help.

The honest question, written down

Reasons stated aloud are performances. Reasons written down can be checked. Before any resignation letter write one page with four parts, and keep it where you will find it again.

The problem, in one paragraph, without the solution. Who has it, how often, what they do about it today, what it costs them. If the paragraph is about a technology or a market rather than a person, the reason is in the first list above. The lessons on [telling a problem from a complaint](/library/how-to-tell-a-problem-from-a-complaint) and on [problems worth a decade](/library/problems-worth-a-decade) are the tests to run on this paragraph.

The cost, in your own figures. Your age now and in ten years. Your current income and what you will draw in year one, which for most founders is nothing or a fraction. The savings the first eighteen months will consume. The things you will not do: the course, the move, the holiday, the promotion that would have arrived. Write them as losses, because that is what they are. A reason that cannot look at this list is not a reason.

Three conditions under which you would stop. A date, a number and a state. For example: thirty-six months without a repeatable sale; personal savings below a figure you set with your family; a health or family event of a kind you name now. Founders who write stopping conditions in advance stop with dignity and start again. Founders who do not write them stop when the money does, which is always later and worse.

The same page, with the outcome changed. Reread the problem paragraph and ask whether you would still want ten years of it if the company never raised a rupee, if it reached ₹5 crore of revenue and never grew further, and if someone else solved the problem first. A yes to all three is rare and is the strongest signal this lesson knows. A no to the first two means the reason is the shape of the company rather than the problem, and the lesson on [whether to raise at all](/library/should-you-raise-at-all) will tell you what that shape costs.

A ritual: the founding page, reread quarterly

Date the page. Show it to the two or three people whose lives it will change, and to one person who has built a company and has no stake in your answer. Then, on the first working day of every quarter, read it again and add only a note in the margin: what is still true, what has stopped being true, whether any stopping condition is close. Four readings a year is more thought than most founders give to the reason they are doing any of it, and it is the only way to tell persistence, which Graham calls more important than raw intelligence, from the sunk cost of a reason that ran out two years ago.

Nothing here is advice to start or not to start. It is a question and the cost of answering it honestly. The sources are below; Graham’s three essays take under an hour and are the best hour available on this subject.

Sources

1. Paul Graham, Why to Not Not Start a Startup, March 2007
2. Paul Graham, How to Make Wealth, May 2004
3. Paul Graham, What Startups Are Really Like, October 2009
4. Paul Graham, Before the Startup, October 2014
5. Jessica Livingston, 9 Things She Learned From Founding YC, Y Combinator blog, July 2018
6. Peter Walker, Carta Data Desk, Startups Take Longer to Reach Each Venture Stage, February 2024 — Average years from incorporation to each round, 2023, from 31,000+ primary fundraises on Carta.