पाठशाला Pathshala · मन Man, The founder · Lesson 02 · Start

The founder’s first ninety days

The first quarter has one job: a paying customer, or honest proof that there is none to be had. Here is the quarter planned week by week, from leaving the job to the first review.

Pathshala, The Founder Library · 11 October 2026 · 8 min read

The first ninety days of a company are usually spent on things that feel like founding: a name, a logo, a company registration, a deck, a website with a waiting list. None of those is the job. The job is to find out, as cheaply and as quickly as the calendar allows, whether anyone will pay for the thing, and the quarter is exactly long enough to find out if nothing else is allowed in.

This lesson lays the quarter out by week. The weeks are a default pace rather than a rule; a founder with customers already in hand will move faster through the middle and one who is still employed will move slower through the whole. What does not change is the order, because each stage exists to make the next one cheaper, and skipping one is how a founder arrives at week thirteen with a product and no idea who it is for.

What the quarter is for

Jessica Livingston’s list of how not to fail begins where every list from Y Combinator begins: nothing else you do will matter if you are not making something people want. The first quarter is the cheapest moment in the company’s life to find that out, because there is no team to protect, no investor to update and no sunk cost beyond the founder’s own time. Every week after this one is more expensive.

So the quarter has one number, and the founder chooses it in week one. For most companies it is paying customers, and the target is one. For a consumer product with no price yet, it is users who returned in their second week without being asked, and the target is ten. Everything in the thirteen weeks either moves that number or is a distraction, and Livingston names the distractions in the same talk: fundraising coffees, networking, press, conferences. Founders who do those things in the first quarter are usually avoiding the one thing that can be measured.

Weeks one and two: the house in order

Two weeks on arrangements feels like a waste when the idea is burning. It is the opposite. A founder who has not computed personal runway makes every decision in the quarter against an imaginary number, and a founder who has not designed the working week spends the quarter answering messages. The lesson on [quitting your job](/library/quitting-your-job-timing-savings-family) has the runway arithmetic and the family conversation; the one on [the founder’s week](/library/time-the-founders-week-designed) has the week. Do both in the first fortnight and write the two results, a month and a timetable, where you will see them.

The co-founder conversation also belongs here, whether it ends in a partnership or a decision to go alone for now. Graham’s first surprise in What Startups Are Really Like is that founders underestimate how severely a co-founder’s character is tested, and the test begins in week one with a question about hours and equity that is easier to ask before either person has given anything up. The lesson on [vesting and the co-founder cliff](/library/vesting-and-the-cofounder-cliff) has the structure.

One more thing for the still-employed. Read the employment contract for the clauses on intellectual property and competing activity. Many Indian technology and services employers assign to themselves anything made on their time or equipment, and some reach further. A founder who learns this in month nine, from a lawyer acting for the former employer, learns it expensively. Learn it in week one and arrange the work accordingly.

Weeks three to six: the problem

Four weeks on the problem and none on the product. This is the part most founders cut, and it is the part that decides the quarter. Graham’s rule in How to Get Startup Ideas is to look for problems, preferably ones you have yourself, and his second rule is about shape: choose something a small number of people want a large amount over something a large number of people want a little. The well, not the crater. The four weeks exist to find the well.

The method is twenty conversations, held properly. Write the problem in a paragraph with no solution in it. List fifty people who might have it, starting with the ones who already know you, and reach twenty. In each conversation ask about the past rather than the future: when did this last happen, what did you do, what did it cost, what have you tried. Never ask whether they would use what you are imagining, because they will say yes to be kind and the yes is worthless. The lesson on [the customer interview](/library/the-customer-interview-done-properly) has the protocol, and YC’s talk on how to talk to users is the forty-minute version.

The output of the four weeks is a count. Of the twenty, how many have already paid money, hired a person or built something themselves to deal with the problem? That number is the first real number the company owns. If it is under five, what you have is a complaint rather than a problem, and the lesson on [telling the two apart](/library/how-to-tell-a-problem-from-a-complaint) will say what to do next, which is usually to change the person rather than the idea. If it is ten or more, name the narrowest group those ten belong to in a single line, and build for them alone.

Weeks seven to ten: the smallest thing that can be sold

Michael Seibel’s talk on building an MVP is the standard forty minutes on this stage. The pace this lesson sets is a first version built in weeks rather than months, for the narrow group found above, solving one case well enough to charge for. Hold the problem tightly and the solution loosely. If the first version needs more than three weeks it is a second version wearing a disguise, and the thing to cut is scope, not the deadline.

Before it exists, ask for money. Go back to five of the twenty whose count said yes and ask for a deposit, a pilot fee or a signed order for the thing you have described. A yes with money attached is the only yes that counts in this quarter. A no with a reason is next week’s work. A no without a reason, from someone who three weeks ago described the problem in detail, is the most useful answer of all, because it says the problem is real and the solution is wrong.

Then deliver by hand. Graham’s Do Things that Don’t Scale describes the most common unscalable thing founders do at the start, which is to recruit users one at a time, and names the technique Stripe’s founders used: when someone agreed to try it, set them up there and then rather than sending a link. Install it on their laptop. Enter their data yourself. Run the first week of the service as a human being with a spreadsheet if the software is not ready. The point is not efficiency. It is to be in the room when the product meets the problem, because that is where the next version comes from.

Weeks three to six decide the quarter. A founder who skips them arrives at week thirteen with a product and no idea who it is for.

Weeks eleven to thirteen: the first customer and the review

The last three weeks close. Either the first customer pays, at whatever price was written down in week nine, or the five who were asked have said no and the reasons are recorded word for word. Both are outcomes. Only the absence of either is a failure of the quarter, because it means the question was never put.

Alongside the payment, measure return. Of the users set up by hand in weeks nine and ten, how many came back in their second week without being prompted? A handful of users returning on their own is the earliest sign a product does something, and its absence is the earliest warning. Neither number is statistically anything. Both are more information than most companies have at this age.

Then write the review, one page, on a date fixed in week one. What was learned about the problem. The number, against the target. What the quarter cost in cash, against the runway computed in week one, because Livingston’s answer to why startups run out of money is that they spend too much, and the habit of checking starts now. And a decision for the next quarter from exactly three options: continue as is, change the customer or the product, or stop. The lesson on [decisions](/library/decisions-reversible-irreversible-how-fast) is about how to make that one well; the point here is that it is made in writing on a known day.

What the ninety days must not contain

Incorporation, until someone needs it. A private limited company brings a filing calendar, a bank account that takes weeks to open and a compliance cost from the first month. Incorporate when a customer needs an invoice from a company or a co-founder needs equity; before that a proprietorship can invoice and a shared document can hold the equity agreement in principle. The lesson on [incorporating in India](/library/incorporating-in-india-pvt-ltd-llp-or-opc) has the order and the portals.

Hiring. Nobody in the first quarter. A contractor for a specific piece of work is fine; a salary is a commitment made against a product that does not yet have a customer.

Fundraising. Investors fund evidence and the quarter is for producing it. A founder who spends weeks three to six in investor meetings instead of customer conversations arrives at week thirteen with a deck and no number to put on it. The lesson on [whether to raise at all](/library/should-you-raise-at-all) is for the quarter after this one.

Names, logos, websites, offices. A name is a domain bought in ten minutes. A logo is a word in a plain typeface. A website is one page that says what the thing does and has a phone number. An office is a table. All four can be improved in the second quarter if there is a second quarter, and none of them changes the number.

A weekly ritual, on Friday afternoon

Every Friday, in thirty minutes, write three lines: the number this week against last week, what was done that moved it, what was done that did not. Tick the checklist above and look at what is still open in the current fortnight. On the thirteenth Friday the thirteen notes are the review, already written. The founder who keeps this up for ninety days has a habit that will outlast the idea, and most first ideas are outlasted.

Nothing here is legal or financial advice; the clauses in an employment contract and the choice of entity deserve a professional’s eye. The sources are below. Graham’s two essays and the two YC talks together take under two hours, which is less than most founders spend on the logo.