पाठशाला Pathshala · विचार Vichār, The idea · Lesson 02 · Start

Problems worth a decade of your life

A company takes ten years. Before you give an idea that long, score the problem on three things: how long it will last, how often it hurts, and how much it hurts each time.

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

The median company that works takes about ten years to become what it was going to be. That is the real price of an idea: not the ₹20 lakh of savings or the first round, but a decade of the founder’s life. Most ideas are chosen in an afternoon and judged on whether they sound good. This lesson gives a way to judge them on whether they are worth the ten years.

The previous lesson, [how to tell a problem from a complaint](/library/how-to-tell-a-problem-from-a-complaint), was about whether a problem is real. This one assumes it is real and asks whether it is big enough and long enough. Three questions do the work. How long will the problem exist? How often does one person meet it? How much does it hurt each time? Each has a number. The three together produce a verdict that a founder can defend to themselves and later to an investor.

Durability: will it still hurt in 2036?

Peter Thiel’s seven questions for a business, set out in Zero to One, include two that most founders never ask of their idea: the timing question, which gets [its own lesson](/library/why-now-timing-argument), and the durability question, which he phrases as whether the market position will be defensible ten and twenty years into the future. The question applies to the problem before it applies to the position. A company built on a problem that disappears in year four has nothing to defend in year ten.

Problems disappear in three ways and each is checkable. A regulation removes it. Compliance pain is real and frequent and intense, which makes it attractive, and a single circular can abolish it; the GST return-filing market changed shape every time GSTN simplified a form. A platform absorbs it. If the problem is a missing feature in something large and well-run, assume the owner of the platform will add it; the more obvious the gap the sooner it closes. A habit ends it. Problems caused by a behaviour that is itself fading, such as anything built around cash handling after UPI, shrink with the behaviour. Ask of a problem: in 2036 does a reasonable person in India still have this? If the honest answer is probably not, the problem might make a good product and a poor company.

The opposite test is also useful. Problems rooted in arithmetic, biology, geography or human nature do not expire. Working capital is scarce for small traders in every decade. Parents want their children to do well in every decade. Goods have to move from where they are made to where they are bought in every decade. These problems change form constantly and never go away, which is exactly what you want under a ten-year company.

Frequency: how often does it hurt?

Kevin Hale, then a Y Combinator partner, listed the properties of a good problem in his 2019 Startup School lecture: popular, growing, urgent, expensive, mandatory and frequent. Of the six he called frequency super important, because every occurrence is another chance to convert the person who has the problem into a customer, and the ideal problem is one people have to solve several times a day. A problem met once a year gives you one chance a year to be remembered. A problem met every working day gives you two hundred and fifty.

Frequency also decides what kind of company you can build. Weekly or daily problems support products that become habits and can be priced low per use. Quarterly or annual problems need a high price per occurrence or a long contract, because the customer will not remember you otherwise, and they tend to be sold rather than adopted. Neither is wrong. A founder who does not know which one they have will build the wrong sales motion for it.

Measure it by asking the person, in the [customer interview](/library/the-customer-interview-done-properly), when the problem last happened and the time before that. Two dates give a frequency. Ten interviews give a distribution. Write the median down; it is a number investors ask for and almost nobody has.

Intensity: the well and the puddle

Paul Graham’s image for intensity is the shape of the hole an idea digs in the graph of demand. A broad shallow hole means many people want something a little. A narrow deep one means a few people want it badly, and he tells founders to choose the latter, because nearly every good startup began as a well. The reason is practical: a product that a few people urgently need gets used, talked about and paid for, and can widen later. A product that many people mildly want gets a sign-up and a shrug.

Kunal Shah gives the Indian version of the same idea. His Delta 4 theory, first described publicly at TechSparks in 2016, scores the efficiency of the old way and the new way out of ten and says that when the new way is better by four points or more the behaviour change is irreversible: the customer never goes back, tolerates the product’s faults, and brags about it unprompted. Intensity is what makes a Delta 4 possible. You cannot deliver a four-point improvement on a problem that was only a one-point irritation to begin with.

Measure intensity in rupees. For a business customer it is money lost per occurrence, or hours spent times what those hours bill at, or a penalty paid. For a consumer it is what they already spend on a workaround, which is the most honest number because it is money that has left their account. A trader who spends three hours a month reconciling GST invoices at an accountant’s rate of ₹800 an hour has a ₹2,400 monthly problem. A commuter who takes an auto because the bus is unreliable has a ₹60 daily problem. Both are real. They are different companies.

Frequency times intensity is the number. Durability is whether you will still be collecting it in year ten.

A worked example: two ideas, scored

A founder in Jaipur is choosing between two ideas. The first is a tool that reconciles a small trader’s purchase invoices against the GST portal so that input credit is not lost. The second is an app that helps people in a city discover new restaurants.

Reconciliation. Frequency: monthly, because returns are monthly, so twelve occurrences a year. Intensity: interviews with fifteen traders put the cost at three to five hours a month plus input credit missed, roughly ₹3,000 a month when priced at what the accountant charges, so ₹36,000 a year per trader. Durability: the form will keep changing, but reconciling what you bought against what the seller declared is a structural feature of a value-added tax and will exist in 2036. People: about 1.5 crore active GST registrations as of April 2025, of which perhaps 50 lakh are small enough to feel it and large enough to pay. Annual pain ₹36,000, a well. Decade pool in the tens of thousands of crore. The company is a dull one and will be slow to sell. It is worth ten years.

Restaurant discovery. Frequency: the interviews say people try somewhere new perhaps twice a month, twenty-four occurrences a year. Intensity: the cost of a disappointing dinner is the dinner, perhaps ₹1,500, but the cost of the discovery problem itself is a few minutes of scrolling, call it ₹50 of attention. Annual pain ₹1,200. A puddle, and one that Google Maps, Instagram and the delivery apps already drain. Durability is high in the sense that people will always eat out, and irrelevant, because the pain was never deep enough to pay for. Many people, little pain, no well. This idea has a [history](/library/tarpit-ideas-why-they-keep-looking-good) and it is not a kind one.

The scorer above is set to the first idea. Move the intensity slider down to ₹50 and the frequency to twice a month and watch the point cross the green line going the wrong way. That crossing is the whole lesson.

Why the good ones look unattractive

The scoring tends to favour problems that founders find unglamorous, and this is not a coincidence. Graham’s essay on schlep blindness observes that the unconscious filters out ideas involving tedious, frightening work, and that a company is defined by the schleps it will undertake. His example is payments: for a decade every programmer knew taking money online was painful, and they built recipe sites instead, because dealing with banks and fraud looked like a nightmare. Deep wells are often surrounded by schlep. That is why they are still unfilled.

The same essay gives the correction: take yourself out of the picture and ask what problem you wish someone else would solve for you. In his piece on frighteningly ambitious ideas Graham adds the tactic for the very large ones: approach them obliquely, start with something deceptively small that works, and do not announce the decade-sized goal. A reconciliation tool for traders in one state is a deceptively small start on the problem of working capital for Indian small business. That is how a decade begins.

A high score is not a guarantee. It is permission to spend the next year finding out, with the first customers rather than with a slide.

A monthly ritual: the idea ledger

Keep one page for every idea you are seriously considering, and on the first Sunday of each month fill in four numbers from the past month’s conversations: the frequency the customers reported, the intensity in rupees they reported or revealed, the count of people you believe have the problem, and your honest estimate of the years it will last. Compute annual pain. Mark the page well, puddle or complaint. If an idea has been a puddle for three months of interviews it is not going to deepen; close the page. If an idea is a well and durable and you have not started, write down what is stopping you, because it is usually the schlep, and the schlep is the moat.

When you do choose, write the three numbers on the first slide of the deck you have not made yet. An investor who sees frequency, intensity and durability stated as numbers with sources will ask you better questions than one who sees a market size, and the questions are the point.

Nothing here is investment advice. It is a way of scoring a decision you will make anyway, so that you make it with numbers rather than with the feeling of an afternoon.