How to Get and Evaluate Startup Ideas

No one—not even YC—knows for sure which ideas will succeed. While startup success depends heavily on execution, starting with a promising initial idea stacks the deck in your favor. This framework draws on quantitative data from top YC companies, Paul Graham's classic essays, batch pivots, and analysis of thousands of rejected applications.

1. Why Do We Need This? (The Status Quo of Idea Generation)

Execution vs. Idea: While execution is paramount, certain ideas possess structural properties that make them vastly more likely to succeed.
The Danger of Bad Starts: Many capable founders waste months on flawed or structurally impossible ideas because they fail to evaluate them using rigorous conceptual tools.
The Goal: Move away from random guessing and toward a systematic approach to identifying, filtering, and generating high-potential startup ideas.

2. Core Mental Model: Fall in Love with a Problem, Not a Solution

The Solution in Search of a Problem (CISP): Founders often start with a cool technology (e.g., "AI is cool, what can I apply it to?") and look for a problem. This results in superficially plausible, made-up problems that users do not actually care about.
Avoid Abstract Societal Issues: Working on massive, abstract issues like global poverty is noble, but they are too broad to serve as viable starting points for early-stage startups.
The Correct Approach: Fall in love with a specific, tractable problem. If people care deeply about the problem, they will care about your solution.

3. The Four Most Common Founder Mistakes

Building a CISP (Solution in Search of a Problem): Creating products for issues users don't care about.
Getting Stuck on "Tar Pit Ideas":
- Definition: Common, highly tantalizing startup ideas that appear easy to solve but harbor hidden structural barriers (e.g., the college student weekend social-planning app).
- Why they happen: They stem from widespread, visible problems.
- How to avoid: Google your idea, find past attempts, and talk to people who tried and failed to understand the hidden friction.
Choosing the First Idea vs. Waiting for Perfection: Picking the very first thought without analysis is dangerous, but waiting around for a "perfect" idea causes paralysis. Treat your idea as a good starting point that will inevitably morph.

4. The 10-Question Framework for Evaluating Startup Ideas

When you have a startup idea, run it through these ten critical evaluation questions:

Evaluation Question	Key Consideration & Example
1. Do you have Founder-Market Fit?	Are you the right team for this specific idea? (Example: Plan Grid — construction worker + elite developer).
2. How big is the market?	Must be a billion-dollar market, either massive now or small but rapidly growing (Example: Coinbase in 2012).
3. How acute is the problem?	The alternative to your solution should be nothing or deep pain (Example: Brex providing corporate cards to startups when no bank would).
4. Do you have competition?	Counter-intuitively, competition is good; it validates market demand. You just need a new insight.
5. Do you want this personally?	Do you or people you know personally experience this pain?
6. Has it only recently become possible/necessary?	Changes in tech, regulation, or user behavior create new windows (Example: Checker providing API background checks for Uber/Doordash).
7. Are there proxies?	Look for large companies succeeding in other regions or similar verticals (Example: Rappi in LatAm looking at Doordash).
8. Is this an idea you'd want to work on for years?	Passion often grows as traction builds, especially in "boring" spaces.
9. Is this a scalable business?	Pure software scales infinitely; human-heavy services/agencies hit operational bottlenecks.
10. Is this a good "idea space"?	An idea space is a class of related ideas (e.g., fintech infrastructure vs. ad tech). High hit-rate spaces increase your odds of stumbling into success (Example: 5Tran pivoting within data analysis software).

5. Three Counterintuitive Signs That Make an Idea Good

Founders frequently shy away from the best opportunities because they look unappealing on the surface. Look for ideas that possess these traits:

Ideas that are hard to get started: Founders suffer from schlep blindness. If an idea requires dealing with messy initial friction (like Stripe getting special bank deals), most competitors are scared off, leaving a massive opportunity unprotected.
Ideas in boring spaces: Payroll (Gusto) or tax software lacks sex appeal, but they suffer from far less competition and enjoy much higher success rates than social apps or restaurant finders.
Ideas with existing competitors: If there are 20 competitors and users still struggle, it proves the market exists. (Example: Dropbox being the 20th cloud storage player, but winning via an OS-level sync UI insight).

6. Organic Discovery vs. Forced Generation

The Power of Organic Ideas: Over 70% of top YC companies came up with their ideas organically rather than sitting in a room forcing brainstorming sessions.
Playing the Long Game:
- Become an expert on something valuable (e.g., working at an early-stage startup).
- Build things you find interesting as a programmer, even if they aren't explicit business models at first.

7. Seven Recipes for Generating Startup Ideas

If you need to generate ideas actively right now, use these recipes ranked from most to least likely to succeed:

Start with your team's unfair advantage: Build at the intersection of your unique expertise and market needs (Example: Rezi founders using real estate/debt financing backgrounds).
Start with a problem you personally encountered: Leverage unusual visibility into an unsexy niche (Example: Vetcove — founders' dad was a vet struggling with old-school supply ordering).
Think of things you personally wish existed: Build tools you want to use daily (Example: DoorDash wanting food delivered to dorms).
Look for recent world changes: Capitalize on sudden shifts in society, behavior, or macroeconomics (Example: Gather Town capitalizing on remote work during COVID-19).
Look for successful companies and build variants: Systematically study growing models and adapt them to new markets or verticals (Example: Nuvo cargo acting as Flexport for Latin America).
Talk to people within a fertile idea space: Systematically interview practitioners or domain founders to map pain points (Example: A to B going to truck stops to understand trucker workflows before building fuel cards).
Look for big industries that seem broken: Target massive legacy sectors that have avoided software disruption.

Bonus Hack: Find a co-founder on platforms like Startup School Co-Founder Matching who already possesses a validated idea.

Summary and Key Takeaways

Fall in Love with the Problem: Avoid solutions in search of a problem (CISPs) or abstract societal ideals; focus on specific, tractable user pain points.
Beware Tar Pit Ideas: Tantalizing, common ideas hide deep structural traps; always research past attempts via Google and interviews.
Prioritize Founder-Market Fit: The best idea is not abstractly good—it is uniquely good for your specific team and background.
Embrace the Boring and Hard: High-hit-rate ideas hide in boring spaces (payroll, tax) and behind difficult initial hurdles (schlep blindness).
Competition Validates Demand: Do not fear competitors; presence of competitors proves market demand if you bring a step-function improvement in UI, technology, or workflow.
Organic Ideas Win: The majority of billion-dollar startups emerge organically from domain expertise or side projects rather than forced brainstorming.
Systematic Recipes Exist: Use targeted generation methods like leveraging past jobs, tracking macro changes, or embedding yourself in unsexy industries (like trucking).
Launch to Find Out: Conceptual tools help stack the deck, but the only definitive way to know if a startup idea works is to build it and put it in front of users.