How do you choose a market for your startup?

The best market for a startup is one where a specific group of people desperately need something that doesn't yet exist — not one where millions of people might vaguely like what you're building. Choose depth over breadth: a small market with intense demand beats a large market with lukewarm interest every time. The founders who get this right almost always do so because they're living the problem themselves, not because they ran a market-sizing spreadsheet.

Depth beats breadth: why intensity of need matters more than market size

There's a mental model worth internalizing before you evaluate any market: picture demand as a two-dimensional graph. One axis is how many people want something; the other is how badly they want it. Most first-time founders chase the horizontal axis — they want a big addressable market — and end up building something a lot of people are mildly interested in. That's actually the harder path, because mild interest doesn't convert, doesn't retain, and doesn't generate word-of-mouth.

Paul Graham's analysis of this pattern in 'How to Get Startup Ideas' is clarifying: nearly every enduring startup began as something a small number of people needed intensely. Microsoft's first product served a few thousand Altair computer owners who were otherwise forced to program in machine language. Facebook launched exclusively to Harvard students. These weren't accidents or humble beginnings that later got lucky — the intensity of need in a narrow market is precisely what gives a startup the traction to grow.

In practice, this means when you're evaluating a market, ask yourself: 'If we disappeared tomorrow, would our target users be genuinely stuck?' If the honest answer is 'they'd find a workaround pretty easily,' you're probably in a broad-but-shallow market. If the answer is 'they'd lose something they actually depend on,' that's a real signal. The goal in the early stage isn't to serve everyone — it's to be irreplaceable to someone.

Start with a problem you actually have

The most reliable way to land in a good market is to notice a problem in your own life that's genuinely painful and unsolved. This isn't just inspirational advice — it's an epistemological shortcut. When you experience the problem yourself, you have ground truth. You know what workarounds people are tolerating, what the failure modes look like, and what a real solution would feel like versus a mediocre one. You're also far less likely to build something nobody wants, which Graham identifies as the most common startup mistake.

The alternative — picking a market by analysis alone — tends to produce ideas that are plausible on paper but disconnected from how people actually behave. Graham's own example of building a platform for art galleries illustrates this precisely: the idea looked reasonable, but because he wasn't embedded in the world of gallery owners, he missed the fundamental fact that galleries didn't want to operate that way. Six months of work preceded the realization.

This doesn't mean you can only work on problems you've personally experienced in your professional life. It means you should be able to get close enough to the problem — through your own experience, through deep relationships with people who have the problem, or through immersion in a domain — that your mental model of the user's world is accurate. The test is simple: can you predict, without asking, what your target user did when they ran into this problem last week?

Be at the frontier of something changing fast

Good market intuition isn't randomly distributed. It clusters around people who are deeply embedded in fields that are currently in flux — either because they're building at the frontier or because they're heavy users of tools that are evolving rapidly. If you're at that frontier, your hunches about what's missing are informed by real signal. If you're an outsider trying to reason your way in, you're working from a map drawn by someone else.

Graham makes this point directly in his writing on startup ideas: being at the leading edge of a fast-moving field is what sharpens your ability to spot what's worth doing. Importantly, 'leading edge' doesn't mean you have to be a researcher or an inventor. It can mean you're one of the heaviest users of an emerging tool — you're stress-testing it daily, running into its limits, and developing opinions about where it breaks. That usage-driven insight is exactly the kind that leads to markets others haven't identified yet.

For founders choosing markets right now, this suggests a practical filter: are you spending serious time in the space, or are you observing it from a distance? If you're a daily active user of AI coding tools and you keep hitting the same frustrating gaps, that's a live signal worth investigating. If you read that 'AI in legal tech is a big opportunity' and decided to build there, you're starting from a secondhand map. The former gives you a market insight that's genuinely yours; the latter puts you in competition with everyone else who read the same article.

Small and specific is a feature, not a bug

One of the most counterproductive instincts early founders have is the desire to appear large. This shows up in market selection as a tendency to define the target market broadly — partly to impress investors, partly because it feels safer than committing to a niche. But a vague market definition is actually a liability. It means you're optimizing your product for nobody in particular, your messaging resonates with nobody specifically, and your early users are too heterogeneous to generate useful feedback.

Garry Tan has pointed out that founders often imitate the behavior of big companies — including their indifference to individual users — because it feels more professional. In market selection, this manifests as targeting everyone and actually reaching no one. The opposite approach — picking a market so specific that it feels almost embarrassingly small — lets you go deep enough to build something that earns genuine loyalty from the people who matter most in the early stage.

Stripe's early trajectory is a useful reference point here. Patrick Collison has described the moment when Stripe went from feeling like something the founders were pushing uphill to something with its own momentum. That momentum came from solving a specific, acute problem for developers — not from trying to serve the entire payments ecosystem on day one. The narrow, deep approach is what creates the density of value that eventually tips into self-sustaining growth. A market that feels too small to matter is often exactly the right place to start.

How to pressure-test your market choice before you build

Once you've identified a candidate market, the fastest way to validate it isn't a survey or a market research report — it's direct contact with people who have the problem. Can you get ten of them in a room, or on a call, without relying on cold outreach? If you have no natural access to the people whose problem you want to solve, that's important information. It either means you're not embedded enough in the domain to build the right thing, or it means the problem isn't as acute as you believe (otherwise people with the problem would be easier to find).

If you're working on something where you genuinely can't reach initial users without cold calls — particularly in enterprise markets where you have no existing relationships — it's worth pausing to ask whether this is the right problem to be working on at this stage of your company. That's not a rule against enterprise software, but it's a real consideration: without direct access to users in the early stage, your feedback loop is broken, and a broken feedback loop in a new market is extremely costly.

The questions worth asking before you commit: Do I know ten people with this problem by name? Would any of them pay for a solution today, even a rough one? Can I get specific about who, exactly, is in my market — not as a demographic segment, but as an individual type of person with a specific workflow and a specific pain? If your answers are concrete and confident, you've found a market worth entering. If they're vague, keep looking.

“The way to get startup ideas is not to try to think of startup ideas. It's to look for problems, preferably problems you have yourself.”

— Paul Graham, source

The one thing to do

Identify ten specific people who desperately need what you'd build, confirm you can reach them without cold outreach, and make sure you understand their problem from the inside — that's your market.

Frequently asked questions

Should your initial market be large enough to build a big company?

Not necessarily at the start. A small, intensely underserved market is a better launching point than a large market with weak demand. Most large companies started in markets that looked too small to matter — the size comes later, from expanding out of a strong initial position.

How do you know if a market is too narrow?

A market is too narrow if there genuinely aren't enough people with the problem to build a sustainable business, even at full penetration. But most founders err in the opposite direction — they go too broad too early. If you can name the people in your initial market and reach most of them directly, you're probably at the right level of specificity.

Is it a red flag if your market doesn't exist yet?

Not always. Some of the best markets are ones that emerge because a new technology makes something newly possible or affordable. The question is whether you can identify people who are currently experiencing the underlying problem — even if they're solving it badly or not at all — rather than speculating that people will develop a need in the future.

What if two founders disagree on which market to target?

Treat it as an empirical question, not a debate. Each founder should identify ten real people in their proposed market and try to get one of them to pay for or commit to using a prototype. The market where you can actually get traction is the answer — not the one that sounds better in a pitch.

Sources

More playbook answers · Growth Prophet home