पाठशाला Pathshala · मन Man, The founder · Lesson 14 · Build
Saying no: the opportunities that would have killed the company
Startups rarely die of a bad idea taken on purpose. They die of good-looking offers accepted one at a time. A filter for partnerships, pivots and press protects focus without closing the doors you will need.
Pathshala, The Founder Library · 11 October 2026 · 7 min read
No founder kills a company by choosing a bad idea on purpose. They kill it by saying yes, one reasonable offer at a time, until the work that was going to make the company succeed has no one left to do it.
This lesson is about the opportunities that arrive once a company has a product and some customers: the large partner who wants a pilot, the adjacent market that suddenly looks bigger, the stage, the award, the feature in a national daily. It explains why these are dangerous in proportion to how good they look, gives a filter that sorts them in minutes, and shows how to decline in a way that keeps the door open for the day you will want it.
Why the dangerous offers arrive as good news
Paul Graham’s Startups in 13 Sentences gives one of its thirteen lines to the matter: avoid distractions. He adds that nothing kills startups like distractions, and that the worst are the ones that pay: consulting, day jobs, profitable side projects. He counts fundraising among them too. In How Not to Die he is blunter still: distraction is fatal to startups. The pattern is not that founders choose distraction. It is that distraction never looks like it. It looks like a bigger customer, a smarter strategy or recognition earned.
The mechanism is founder time. A company at this stage has one or two people whose hours set its speed. Every offer competes for those hours, and the offers that flatter most usually ask the most of them: the partner wants the founder in the meetings, the conference wants the founder on stage, the journalist wants the founder on the phone. An offer is not free because it brings money or attention. Its price is the founder hours it takes away from the one number the company is judged on.
Partnerships: the deal that eats the quarter
The large-company partnership is the classic killer because it combines every warning sign. It promises distribution the startup cannot get alone. It moves at the partner’s speed, which is slower than the startup’s by a factor of several. It requires custom work, a dedicated contact, legal review and steering meetings. And it often asks for something one-way: exclusivity in a category, a custom integration the startup must maintain, or an announcement that sets the startup’s position in public before it has chosen it.
Graham’s own experience is the useful corrective. In the same essay he recalls that his company had something like twenty deals of various types fall through, that after the first ten or so they learned to treat deals as background processes to ignore until they terminated, and that it is very dangerous to morale to start to depend on deals closing. A partnership that is worth doing can survive being treated that way. Put a small team on it, set a date by which it must show a number, and keep the founder’s calendar on the customers who are already paying.
Pivots: the new idea that arrives in a bad month
A change of direction is sometimes right. Graham lists obstinacy among the eighteen mistakes that kill startups, and advises against getting too attached to an original plan because it is probably wrong. The danger is not pivoting. It is pivoting for the wrong reason at the wrong time, and a founder’s mood is a poor guide to either. New ideas look most attractive in the month the current one is hardest, which is exactly when they deserve the most suspicion.
Two tests separate a pivot from an escape. Is it pulled by evidence or pushed by fatigue? A pivot pulled by evidence comes with customers already asking for the new thing, usage concentrated in an unexpected place, or a cohort that retains far better than the rest. A pivot pushed by fatigue comes with a story and a spreadsheet. Has the current plan had a fair trial? If the company has not run the current plan at full effort for a defined period with a number it agreed in advance, it does not yet know whether the plan failed. The [lesson on reversible and irreversible decisions](/library/decisions-reversible-irreversible-how-fast) applies directly: a pivot is usually a one-way door for the team and the customers, and deserves a rested decision.
Press, stages and awards
Attention is the opportunity founders most underprice, because it feels like marketing and costs nothing in cash. It costs a great deal in time: the application, the preparation, the travel, the day itself and the two days after spent answering messages from people who will never buy. Most early-stage press and most stages move vanity numbers and not the one that matters. A conference panel with five founders and a moderator rarely produces a customer.
The test is the same as for any other offer, with one addition. Brooke Hammerling, interviewed by First Round Review, argued that a startup should be able to state clearly what it does before pursuing press at all: if you cannot answer that question, do not do anything else until you can. Attention given to a company that cannot yet say what it does in a sentence is wasted at best. Say yes to the stage where your buyers sit in the audience, to the journalist who covers your customers’ industry, and to almost nothing else in the first two years.
An offer is not free because it brings money or attention. Its price is the founder hours it takes from the one number the company is judged on.
The filter, walked
A filter works only if it is quick enough to use on a Tuesday afternoon and strict enough to stop the flattering offers. The one below asks five questions in the order that saves the most time. The first is the strongest: an offer that does not move the quarter’s number within ninety days needs an unusually good reason to survive. The second protects the rare door that will not reopen. The third and fourth price the founder’s time, and the fourth is the one most people skip: if you cannot name what you will stop doing, the new thing will be done on top of everything else, badly. The fifth is the two-way door test from the lesson on decisions.
An example. A large retailer offers a pilot: integrate with its procurement system, give it category exclusivity for a year, and it will roll the product out to two hundred stores if the pilot works. It moves revenue, but not within ninety days. The door may not close, since the retailer will still buy software next year. If the founders decide it is rare enough to pursue, it takes two founder days a week, and nobody has named what stops. The likely verdict is a counter-offer: a ninety-day pilot in ten stores, no exclusivity, a fixed fee for the integration, and a review date.
Walk five offers from the last quarter through it, including some you accepted. Founders who do this usually find at least one yes they would now refuse and the hours it cost. That is the number that makes the next no easier.
Saying no without closing the door
Most founders keep bad opportunities alive because declining feels rude and might burn a relationship they will need later. The opposite is true. A fast, specific no is remembered as respect; a slow, vague maybe is remembered as being strung along. Three habits make the no easy to give and easy to receive. Answer within two days. The cost of an open maybe falls on both sides. Give the real reason in one line. “We are putting everything into our existing customers until March” is complete and true, and it tells the other side when to come back. Offer the smaller version where one exists. A pilot with a fixed end, an introduction to someone better placed, a written answer instead of a panel seat. The person who made the offer leaves with something and the door stays open.
Inside the company, make the no visible. When the founders decline something tempting, say so at the weekly meeting and say why. A team that sees its founders refuse attractive offers to protect the core work learns that focus is the company’s policy, and starts applying the same filter to the offers that reach them.
The monthly no list
Keep one running list, shared between the founders, of every opportunity declined or deferred: what it was, who offered it, the verdict from the filter and the date to revisit. On the first working day of each month, spend twenty minutes on it. Strike the ones that no longer matter. Move any whose door has started to close back through the filter. Check the yeses from three months ago against the result and the date you wrote down when you accepted them, and ask honestly whether each earned its hours. Then count founder hours last month spent on things that were not the quarter’s number. If that count is rising, the filter is being skipped, and next month starts with fewer yeses.
Sources were checked in October 2026. The filter is a discipline, not a rule: the founders’ judgement decides, and it decides better with the questions answered.
Sources
- Paul Graham, Startups in 13 Sentences, February 2009 — Avoid distractions; about twenty deals fell through; treat deals as background processes.
- Paul Graham, How Not to Die, August 2007 — Distraction is fatal to startups.
- Paul Graham, The 18 Mistakes That Kill Startups, October 2006 — Obstinacy; working on too many different things.
- First Round Review, Why Most Startups Don’t ‘Get’ Press, February 2015 — Brooke Hammerling: be able to say what the company does before pursuing press.