How to find and evaluate startup ideas: a YC framework

Executive overview

Most founders either jump at bad ideas or wait forever for perfect ones. The real skill is picking a promising starting point — one with enough signal to morph in the right direction.

A strong idea has founder-market fit, a real (not imagined) problem, and lands in a fertile idea space. The framework below covers how to evaluate an existing idea and how to generate new ones.

The best startup ideas aren't found by brainstorming — they're noticed organically by people with relevant expertise and genuine problems.

The four most common mistakes

• Building something that doesn't solve a real problem — a solution in search of a problem (SISP)
• Getting stuck on tar pit ideas: common, superficially plausible problems with hidden structural reasons they can't be solved
• Jumping at the first idea without reflection
• Waiting for the perfect idea and never starting

Ten questions to evaluate any startup idea

1. Founder-market fit — is your team obviously right for this problem?
2. Market size — is it a billion-dollar market now, or clearly growing toward one?
3. Acuteness — do people care enough? (Brex: startups literally couldn't get a corporate credit card elsewhere)
4. Competition — most good ideas have competitors; no competitors often means no demand
5. Personal desire — do you or people you know actually want this?
6. Timing — did something recently change (technology, regulation, new behavior) that created this opening?
7. Proxy — is there a large company doing something similar elsewhere that validates the idea? (Rappi used DoorDash as proxy)
8. Longevity — could you work on this for years? (Passion often develops as the business works, not before)
9. Scalability — software scales; service-heavy models often don't
10. Idea space — is the broader category fertile? Fintech infrastructure and vertical SaaS have high hit rates; consumer hardware and ad tech much lower

Three idea types that seem bad but are actually good

• Hard to start — schlep-blind ideas scare off competitors (Stripe: founders had to negotiate with banks and learn credit card infrastructure; thousands of developers saw the problem and still didn't try)
• Boring spaces — Gusto built payroll software; boring ideas get left on the table longer and have higher hit rates than "fun" ideas
• Existing competitors — Dropbox was the 20th cloud storage company; widespread non-use of existing products signals the problem is real but unsolved

Seven recipes for generating startup ideas

1. Start from your expertise — ideas from genuine domain knowledge have automatic founder-market fit (Rezzy founders searched only within real estate and fintech)
2. Start from a problem you've personally seen — especially one you're in an unusual position to notice (VetCove: founders grew up watching their veterinarian father order supplies by phone)
3. Think of things you wish existed — classic, but high tar-pit risk; ask why it doesn't exist yet (DoorDash)
4. Look for recent changes in the world — new tech, regulation, or behavior creates new gaps (Checker: background-check API emerged when gig economy hiring scaled)
5. Find successful companies and look for variants — new geography, adjacent market (Nuvo Cargo = Flexport for Latin America; Rappi = DoorDash for Latin America)
6. Talk to people in a fertile idea space — A2B (fuel cards for truckers) found their idea by driving to truck stops and interviewing drivers; also talk to founders in the space
7. Find a co-founder who already has an idea — gets you both a partner and a starting point

How to position yourself for organic ideas

• Become an expert at something valuable — work at a startup
• Build things you find interesting, even without a business model (Replika started this way)
• Ideas that surface organically are more likely to be good; at least 70% of top YC companies found their idea this way, not by explicit brainstorming