lvndry/startup-brainstorm

Startup Brainstorm

Generate and evaluate startup ideas using the mental models of top founders (Thiel, Graham, Andreessen, Bezos, etc.), validated against current trends and competition.

When to Use

• User wants to brainstorm startup ideas (general or in a domain)
• User asks "what should I build?", "business ideas", or "startup opportunities"
• User wants idea validation with trend and competition analysis
• User wants to think like elite founders when evaluating concepts

Research Requirement

Before generating or evaluating ideas, use web search to:

• Trends: Macro, tech, regulatory, societal trends (use recency constraints: last 12–24 months)
• Competition: Direct competitors, adjacent players, substitutes
• Market signals: Funding rounds, acquisitions, analyst reports, "why now" factors

Cite sources and dates. Never invent market data.

Workflow

1. CLARIFY → 2. TRENDS → 3. BRAINSTORM → 4. COMPETITION → 5. EVALUATE → 6. PRESENT

Step 1: Clarify

• Domain/vertical: Industry, sector, or "anything"
• Constraints: B2B/B2C, geography, technical depth, team skills
• User's edge: Domain expertise, network, unfair advantage
• Preferred output: Single strong idea, list of 3–5, or full analysis of one concept

Step 2: Trends (Search)

Search for:

• Macro: Regulation, policy, economics, labor, demographics
• Tech: AI/ML, infra, dev tools, hardware breakthroughs
• Societal: Remote work, health, climate, privacy, creator economy
• "Why now": What changed in last 2–3 years that enables this?

Step 3: Brainstorm

Apply founder mental models (see references/founder-frameworks.md):

• Thiel: "What important truth do few people agree with you on?" Monopoly vs competition.
• Graham: Do things that don't scale. Make something people want. Talk to users.
• Andreessen: "Software is eating the world." Timing, market size, 10x better.
• Bezos: Customer obsession, flywheel, long-term.
• Musk: First principles. Solve the hardest problem. 10x, not 10%.

Generate 3–7 raw ideas. Mix problem-first and technology-first. Prefer contrarian or underrated angles.

Step 4: Competition (Search)

For each shortlisted idea, search:

• Direct competitors (same problem, similar solution)
• Indirect competitors (same outcome, different approach)
• Substitutes (status quo, manual process, "do nothing")
• Incumbents who could move in

Map: Who exists? Funding? Traction? Gaps they leave?

Step 5: Evaluate

Score each idea on:

• Problem: Real pain? Willingness to pay? Frequency?
• Market: TAM/SAM, growth rate, "why now" clarity
• Solution: 10x better or 10%? Defensibility? Moat potential
• Competition: Crowded vs white space? Can you win?
• Founder–market fit: User's edge, passion, unfair advantage
• Trend alignment: Riding tailwinds vs fighting headwinds

Step 6: Present

Use output format below. Lead with strongest ideas. Include trend and competition evidence.

Output Format

# Startup Brainstorm — [Domain/Theme] — [Date]

## Summary

[2–4 sentences: strongest idea(s), why now, key insight.]

## Trends (Sources: [date range])

- **[Trend 1]**: [What's happening + why it matters] — [source/date]
- **[Trend 2]**: [What's happening + why it matters] — [source/date]
- **Why now**: [2–3 bullets on what changed recently]

## Ideas

### Idea 1: [Name]

**One-liner**: [What it does in one sentence]

**Problem**: [Who has it, how painful, how often]
**Solution**: [Core value prop, 10x angle]
**Market**: [TAM/SAM, growth, geography]
**Competition**:

- Direct: [Player A], [Player B] — [gap they leave]
- Indirect: [Player C] — [how you differ]

**Moat**: [Network effects, data, brand, distribution, etc.]
**Founder fit**: [How user's edge helps]
**Trend alignment**: [How trends support this]

**Risk**: [Main risk]
**Why it could win**: [Key insight]

---

### Idea 2: [Name]

[Same structure]

## Competition Landscape (Summary)

| Idea   | Direct Competitors | Gap           |
| ------ | ------------------ | ------------- |
| Idea 1 | [List]             | [Opportunity] |
| Idea 2 | [List]             | [Opportunity] |

## Recommended Next Steps

1. [Validate X with users/customers]
2. [Search/research Y]
3. [Build MVP of Z]

Founder Questions to Ask

For each idea, mentally run through:

• Thiel: "What do I believe that's true but almost nobody agrees?"
• Graham: "Would 10 people pay for this today?"
• Andreessen: "Why couldn't this have been built 5 years ago?"
• Bezos: "What does the customer deeply want that we're not giving?"
• Hoffman: "Where's the network effect or blitzscaling potential?"

Anti-Patterns

• ❌ Inventing trends or competition without search
• ❌ Ignoring "why now" — every idea needs a timing thesis
• ❌ 10% better solutions; aim for 10x or contrarian
• ❌ Dismissing competition; map it and find the gap
• ❌ Generic ideas ("another AI wrapper"); push for specificity and edge
• ❌ No founder–market fit; tie ideas to user's strengths