How to Get Startup Ideas

Based on the YC Library talk. The key insight running through all of this: most of the time you won't know if an idea is good until you launch it.

Evaluating an idea

Avoid tarpit ideas. These are ideas that seem attractive but have a graveyard of failed attempts behind them. Google your idea. If people have tried it and failed, reach out and find out why — it's the fastest way to learn whether the obstacle was execution or something structural.

Founder–market fit. The question isn't just "is this a good idea?" but "are we the right people to work on it?" Pick an idea that plays to your team's specific background and expertise.

Market size and trajectory. You need a real market. Ideally one that's already large, but a fast-growing market can work too — being early in a growing space beats being late to a crowded one.

Problem intensity. How much do people actually care? The strongest signal: the alternative to your product is doing nothing. If people are suffering and tolerating a broken status quo, you have a real problem.

Competition is fine — lack of insight is not. Entrenched competition doesn't disqualify an idea. Most good ideas do have competition. What you need is a genuine new insight that explains why you'll win where others haven't.

Personal desire. Do you want this? Do other people want this? Both matter.

Recent change. Has something shifted in the world that makes this idea newly possible, or that has created a new problem? Technology changes, regulatory changes, and cultural shifts all open windows.

Proxy companies. Is there a large company doing something similar in a different market — not a direct competitor, but a proof of concept? Mr D Food in SA, DoorDash in the US, and Deliveroo in the UK are all proxies for each other.

Long-term conviction. Are you prepared to work on this for years? Passion often follows success rather than preceding it — founders tend to grow into their idea — but you still need enough initial belief to push through the hard parts.

Scalability. Services businesses (agencies, dev shops, anything requiring high-skilled labour per customer) are hard to scale. Software and products that can serve more customers without proportional headcount are far easier to build into a venture-scale company.

Idea space quality. A good idea space has a reasonable hit rate for viable startup ideas and room for founder–market fit. If your first idea in the space doesn't pan out, are there adjacent ideas you could pivot to? A fertile space gives you multiple shots.

Counterintuitive places to look

These characteristics keep most people away — which is exactly why they're worth exploring:

• Ideas that seem hard. Difficulty is a moat. If it were easy, it would already be done.
• Ideas in boring spaces. Boring industries are chronically under-attacked. Nobody wants to work on payroll software until they see the margins.
• Spaces with existing products that nobody uses. If incumbents exist but have poor adoption, there's unmet demand waiting for a better solution.

How ideas actually happen

Don't sit down and try to think of startup ideas. That approach produces ideas that sound plausible but aren't grounded in real pain.

Ideas find you when you're embedded in a problem space. The ways to get there:

• Become an expert in a field — you'll start noticing broken things that outsiders can't see
• Work at a startup — proximity to the problem surfaces ideas naturally
• Build things you find interesting, even without a business goal in mind — products and tools often reveal themselves as businesses over time

A recipe for generating ideas

1. Start from your expertise. What is your team genuinely good at? Think of ideas that exploit those specific advantages. This gives you automatic founder–market fit from the start.

2. Start from a problem you've personally experienced. Your own frustration is one of the most reliable signals that a real problem exists.

Tip for (1) and (2): Go through every job you've held, every internship, every significant life experience. What problems did you run into? What do you know that most people don't? Where does your background put you in the best position to see something others miss?

3. Think of things you wish existed. This is the most tempting approach — and the most dangerous. It's the primary source of tarpit ideas. Use it, but be skeptical of the results.

4. Look for recent changes that create new opportunities. Regulatory shifts, new technologies, and cultural changes all open windows that didn't exist before.

5. Look for new variants of recently successful companies. If a model worked somewhere, ask whether it can be applied in a different vertical, geography, or customer segment.

6. Talk to people. Pick a fertile idea space first, then go deep into it: talk to potential customers, talk to founders who've worked in the space, talk to anyone with relevant knowledge. Build up a mental map of the problems worth solving. Most founders skip this step — the ones who do it seriously tend to find genuinely good ideas. It's unglamorous work with an outsized return.

7. Look for big industries that seem broken. Scale plus dysfunction equals opportunity.