Most people think they are stuck because they do not have a brilliant idea. They are waiting for lightning to strike. It rarely does, and even when it does, a brilliant idea on its own is worth almost nothing. What matters is finding a real problem and proving someone will pay you to solve it, before you spend six months building the wrong thing. That is what this chapter is about.
Where good ideas actually come from
Paul Graham wrote the essay every founder should read here, “How to Get Startup Ideas.” The line to tattoo on your brain: the way to get startup ideas is not to try to think of startup ideas, it is to look for problems, preferably problems you have yourself.
Read that again. Stop brainstorming clever ideas at your desk. Go look for problems. Here is where they hide.
Scratch your own itch. The best ideas grow out of your own life. You feel the problem yourself, so you understand it better than any market research could teach you, and you are your own first customer. Almost every product I have built started as something I needed and could not find.
Live in the future. Get to the leading edge of something that is changing fast, then build what is obviously missing. People at the edge of a new technology see the gaps months or years before everyone else. In 2026 that edge is AI, but it is always somewhere. Put yourself there and the missing pieces become obvious.
Work the intersection of your skills and a real problem. You can see problems in a field you know that outsiders cannot even perceive. If you have spent years in logistics, or healthcare, or restaurants, you are sitting on problems other founders would kill for, and you do not even notice them because they feel normal to you.
Build for an audience you already have. If you have spent time building a following, a community, or a reputation in some niche, you already have distribution and you already hear their problems all day. Building for people who already trust you is one of the biggest unfair advantages there is.
Follow where money already changes hands. This is the most underrated one. Look for problems people already pay to solve, with clunky tools, with agencies, with virtual assistants, with duct-taped spreadsheets. Existing spending is proof of willingness to pay. You do not have to convince anyone the problem is worth money. They are already spending it.
How to find them on purpose
You do not have to wait for ideas to come to you. Go hunting.
Live in the communities where your future customers already complain. Indie Hackers is full of founders sharing what they built and what they earn, which is a searchable archive of small problems people pay to solve. The right subreddit, a niche Discord, a forum for an industry: read them and watch for the same complaint surfacing again and again. A problem that people gripe about repeatedly, in public, is a problem worth a closer look.
A useful lens here is jobs to be done, an idea from Harvard professor Clayton Christensen. People do not really buy products, they hire them to make progress on a job in their life. Stop thinking about features and start asking what job your customer is trying to get done, and what they currently hire to do it. The thing they hire today, however ugly, is your real competition and your biggest clue.
Now the hard part: do not build yet
You have a problem and an idea. Every instinct screams build it. Do not. First you go talk to the people who have the problem, and this is where almost everyone fools themselves.
Here is the trap. You are excited, so you describe your idea to friends and potential users, and they say “that is a great idea, I would totally use that.” You walk away validated and pumped. You have learned nothing. People lie to you, not out of malice, but out of politeness. Nobody wants to tell you your baby is ugly.
The fix is a small, brilliant book called The Mom Test by Rob Fitzpatrick. The name comes from a simple idea: you should be able to ask questions in a way that even your own mother could not lie to you, because you are asking about her life, not about whether she likes your idea. Three rules carry the whole thing.
- Talk about their life, not your idea. The moment you pitch, the conversation is contaminated and you start fishing for compliments. Keep your idea out of it for as long as you can.
- Ask about specifics in the past, not opinions about the future. “Would you use this?” and “Do you think this is a good idea?” are useless questions, because the answers are hypothetical and flattering. Ask “tell me about the last time you ran into this problem.” If they cannot remember a specific recent time, the problem is not urgent. If they have never tried to solve it, it is not painful enough.
- Talk less, listen more. You are there to learn, not to sell. Shut up and let them talk.
The signals you are listening for are not words, they are actions they have already taken. Have they already cobbled together a workaround? Have they already paid for something to solve this? Have they spent real time trying to fix it themselves? People who have hacked together a spreadsheet to manage a problem are telling you, with their behavior, that it hurts enough to pay for a real solution. That is gold. “Cool idea” is not.
Validate demand before you build
Talking to people tells you the problem is real. The next step is proving they will actually pay, and you can do most of this before you write a single line of product code.
Understand the ladder of signals, from weakest to strongest. At the bottom are the vanity signals: likes, “that is cool,” “I would totally use it,” a friction-free email signup. These feel like progress and mean almost nothing. At the top are the real signals: someone pays you, pre-orders, puts down a deposit, gives you a credit card, or hands over something genuinely scarce to them. The rule is simple. Money or a hard commitment is signal. Everything else is noise.
Here are the tests, from cheapest to most convincing.
A landing page and a waitlist. Put up a single page that describes the product as if it exists, with a clear price and a button. It costs almost nothing and tells you whether people are interested. Interest is weak signal on its own, so add friction to make it mean more.
The fake door test. Add a button, a plan, or a feature that does not exist yet, and measure how many people click it and try to proceed. You are testing real intent, what people do when they think they can actually get the thing, not what they say.
The concierge approach. Deliver the result by hand, manually, to your first few customers before you automate anything. No product, just you doing the work behind the scenes. You learn exactly what they value, and they pay you for an outcome while you still have zero code.
Charge before you build. This is the strongest test there is. A paid pre-order, a deposit, an upfront annual payment. The only thing that proves someone will pay is someone paying.
The famous examples all did exactly this. Drew Houston made a short demo video of Dropbox before the product really worked and posted it online, and the beta waitlist reportedly jumped from a few thousand to many times that almost overnight. The exact numbers get repeated and inflated in the retelling, but the lesson holds: he tested demand with a video instead of a finished product. Joel Gascoigne validated Buffer with a two-page test, a page explaining the idea and then a pricing page in the middle, to see not just whether people were interested but which plan they would click. Zappos started with the founder photographing shoes in local stores and buying them at retail to fulfill the first orders by hand, to test whether anyone would buy shoes online at all before holding a cent of inventory.
None of them started by building the whole thing and hoping.
The mistakes that kill ideas
- Building first, then looking for customers. You spend the money, then discover nobody wants it. Reverse the order. Find demand, then build.
- Asking “would you use this?” A hypothetical question that buys you a flattering lie. Ask about what they have actually done.
- Pitching in the middle of a customer conversation. The second you start selling, you stop learning. Save it.
- Building for a market you do not understand. If you are not close to the problem, you cannot see what is real and you cannot smell the polite nonsense people feed you.
- Falling in love with the solution instead of the problem. When you are attached to your clever solution, you defend it against every piece of evidence that says it is wrong. Stay in love with the problem. Be ruthless about the solution.
- Counting vanity signals as validation. Likes and “that is cool” are not customers. Until money or a real commitment changes hands, you have not validated anything.
Ideas are cheap and everywhere. The skill is not dreaming one up, it is finding a real problem, refusing to lie to yourself about whether people care, and proving they will pay before you build. Get good at that and you will never run out of things worth making.