# Yc Insight Driven Bp

> Y Combinator风格的洞察先导商业计划书框架。极简5-8页格式，围绕一个反共识洞察构建。最适合pre-seed到seed、加速器、天使投资人和首次创业者。 Use when: "YC BP", "Y Combinator pitch", "洞察驱动BP", "极简BP", "pre-seed deck", "angel investor deck", "加速器申请", "insight-first BP", "YC风格融资". 投资人不投生意，他们投洞察——一个你看到而别人没看到的关于世界的真相。Cross-references: sequoia-structured-bp, raskin-narrative-bp, pitch-deck-creator. Built by UniqueClub 🌐 https://uniqueclub.ai

- Skill: `wulaosiji/yc-insight-driven-bp` (Agent Skill)
- Install (CLI): `npx skillmds add wulaosiji/yc-insight-driven-bp`
- Raw SKILL.md: https://api.skillmd.com/api/skills/wulaosiji/yc-insight-driven-bp/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: MIT
- Author: wulaosiji (https://skillmd.com/u/wulaosiji)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/wulaosiji/yc-insight-driven-bp

---


# YC Insight-Driven BP

> YC风格的洞察先导BP框架。极简、洞察先导、以创始人为中心。投资人不投生意，他们投洞察。

## When to Use

Use this skill when:
- 用户需要以反共识洞察为核心的极简BP框架
- 用户正在融pre-seed到seed轮，面对加速器或天使投资人
- 用户是首次创业者，需要创始人-centric的pitch deck
- 用户需要5-8页极简格式或1页memo
- 用户需要直接、对话式、略带不羁的语气

Do NOT use this skill if:
- 用户需要严谨的数据驱动框架（A轮及以后、机构投资人）→ use `sequoia-structured-bp` instead
- 用户需要故事驱动的现场演示框架（Demo Day/路演比赛）→ use `raskin-narrative-bp` instead
- 用户需要可编辑的PPTX文件 → use `pitch-deck-creator` instead
- 用户需要中文融资BP大纲 → use `fundraising-bp-planner` instead

Typical triggers:
- 「YC风格BP」「洞察驱动商业计划书」「极简pitch deck」「加速器申请材料」
- "Y Combinator pitch deck", "insight-first BP", "pre-seed deck", "angel investor presentation"

## Workflow

本框架描述"如何使用YC洞察先导BP框架"的步骤：

本工作流遵循六步推进法，共6个步骤：
1. **探查 (Probe)**：动手前先完整读取全部输入，确认目标和约束。
2. **约束 (Constrain)**：验证输入完整性，设定边界和不可降级的交付标准，受阻时换通道。
3. **证据 (Evidence)**：每个数字必须来自输入、具体信源或可复现计算，收集支撑数据。
4. **执行 (Execute)**：调用脚本/API/生成内容，先给影响与结论，再给行动和必要证据。
5. **验证 (Verify)**：用"可能失败"的动作验证——不同于生成路径的方式回读输出。
6. **交付 (Deliver)**：返回结果，清理临时文件。

### Step 1: 探查 (Probe)
完整读取用户的项目信息，确认融资阶段（pre-seed到seed为佳）、行业、创始团队背景和核心洞察。理解项目的独特世界观。

### Step 2: 约束 (Constrain)
设定框架使用边界：
- 输出长度5-8页幻灯片，或1页memo（最多500字）
- 语气：直接、对话式、略带不羁
- 一切内容服务于核心洞察
- 洞察必须是反共识的（聪明人会不同意）
- 不使用"AI将改变一切"等模糊表述
- 受阻时换通道，不降级交付物

### Step 3: 证据 (Evidence)
收集支撑洞察的证据：
- 洞察必须具体（不是"AI很大"）
- 问题必须是洞察的直接后果
- traction指标必须直接验证洞察（不只是增长）
- 客户引言必须确认洞察（不只是"产品很好"）
- 市场规模必须通过洞察重新定义（不是通用TAM）
- 团队故事必须解释"为什么你能看到这个洞察"

### Step 4: 执行 (Execute)
按照下方「Framework Content」中的逐页指南，先通过"一句话测试"确定核心洞察，再逐页撰写BP内容。先给影响与结论，再给行动和必要证据。每页遵循：Goal → Structure → Good/Bad Example → Check 的结构。

### Step 5: 验证 (Verify)
用Self-Review Checklist逐项验证——不同于生成路径的方式回读输出：
- 能否用一句话陈述洞察？
- 洞察是否反共识（聪明人会不同意）？
- 每页是否都服务于洞察？
- deck是否少于8页？
- 是否解释了"why now"？
- traction是否在验证洞察（不只是增长）？
- 团队故事是否关于"为什么我们能看到这个"，而非"我们做过什么"？
- ask是否直接且具体？

### Step 6: 交付 (Deliver)
返回完整的5-8页BP内容（Markdown格式，逐页指南），或1页memo格式。包含核心洞察、每页的目标、结构、示例和检查清单。清理临时文件。如需生成可编辑PPTX，使用 `pitch-deck-creator`。

## Output

输出为Markdown格式的完整BP逐页指南或1页memo，包含：
- 核心洞察（通过一句话测试确定）
- 5-8页幻灯片，每页包含：Goal、Structure、Good/Bad Example、Check清单
- 或1页memo格式（最多500字）
- Self-Review Checklist
- Anti-Patterns to Avoid（5种常见反模式）
- YC-Specific Tips（5条YC特定建议）
- 所有示例基于ChainLedger作为统一案例

## Framework Content

### Meta

- **Style**: Minimalist, insight-first, founder-centric
- **Best For**: Pre-seed to Seed; accelerators; angel investors; first-time founders
- **Source**: Y Combinator Demo Day + partner office hours
- **Output Length**: 5–8 slides, or 1-page memo
- **Tone**: Direct, conversational, slightly irreverent

### Core Philosophy

Investors don't invest in businesses. They invest in **insights** — a truth about the world that you see and others don't.

> "The best founders have a secret. They know something the rest of the world doesn't." — Paul Graham

YC decks are short because the insight should be so compelling that everything else is obvious.

---

### The One-Sentence Test

Before writing a single slide, answer this:

> What do you believe about the world that is true, but that most smart people disagree with?

This is your **insight**. Everything in the deck serves this insight.

**Good Examples**:
- "The best software companies will be built outside Silicon Valley."
- "AI won't replace doctors — it will replace the medical billing industry."
- "Cross-border B2B payments aren't a fintech problem. They're a compliance data problem."

**Bad Examples**:
- "We make the best project management software." (Not contrarian)
- "AI will change everything." (Too vague)
- "We have a better UI than our competitors." (Not an insight)

---

### Slide 1: The Insight (The Hook)

**Goal**: Make the investor lean forward.

**Structure**:
1. **The counterintuitive truth** (1 sentence)
2. **Why most people are wrong** (1 sentence)
3. **What this makes possible** (1 sentence)

**Good Example**:
> "Most people think cross-border payments are slow because of banking infrastructure. The real bottleneck is compliance data — sanction lists, KYC, trade documentation are checked manually by 3 different parties. If you automate compliance data sharing, settlement drops from 7 days to 4 hours without changing a single bank."

**Bad Example**:
> "ChainLedger is a fintech platform for cross-border payments."

**Check**:
- [ ] The insight is contrarian (smart people would disagree)
- [ ] The insight is specific (not "AI is big")
- [ ] The insight implies a big opportunity

---

### Slide 2: Problem (From the Insight)

**Goal**: Show the insight creates a massive, painful problem.

**Structure**:
1. **The old world** (as seen through your insight)
2. **Who suffers most** (be specific)
3. **The cost** (quantified)

**Good Example**:
> "Because compliance is manual and siloed, every cross-border transaction requires 3–5 intermediaries to re-verify the same data. Mid-market manufacturers importing from Asia spend 7 days and $2,500 per transaction on compliance overhead alone. Their 2-person finance teams are drowning."

**Check**:
- [ ] The problem is a direct consequence of your insight
- [ ] The suffering party is narrowly defined
- [ ] The cost is quantified

---

### Slide 3: Solution (The Inevitable Answer)

**Goal**: Show that once you see the insight, your solution is the only logical answer.

**Structure**:
1. **The mechanism** (how you exploit the insight)
2. **The result** (what changes for the customer)
3. **Why now** (why wasn't this done before?)

**Good Example**:
> "We built a shared compliance data layer. Instead of 3 banks independently verifying the same KYC, our network verifies once and shares the attestation. Result: settlement in 4 hours, compliance cost down 90%.
> Why now? Open banking APIs + regulatory sandbox programs make data sharing legal for the first time."

**Check**:
- [ ] The solution directly exploits the insight
- [ ] The mechanism is simple (one sentence)
- [ ] The "why now" answers the "why hasn't this been done" objection

---

### Slide 4: Traction (Proof the Insight is Right)

**Goal**: Show early evidence that your insight is correct.

**Structure**:
1. **The metric that validates the insight**
2. **Trend**
3. **Customer quote** (if available)

**Good Example**:
> "In 6 months: 120 customers, $50k MRR, 12% MoM growth. But the key number: 94% of customers say 'we didn't know compliance was the bottleneck until we used ChainLedger.' This validates the insight."

**Check**:
- [ ] The traction metric directly validates the insight
- [ ] There's a growth trend
- [ ] A customer quote confirms the insight (not just "great product")

---

### Slide 5: Market (Size Through the Insight)

**Goal**: Show the insight opens a massive market.

**Structure**:
1. **Market size via the insight** (not generic TAM)
2. **Expansion path** (adjacent markets your insight unlocks)

**Good Example**:
> "If compliance data is the real bottleneck, the market isn't 'cross-border payments' ($180B). It's 'compliance automation for trade finance' ($42B). And once we own compliance data, we can expand into trade credit, insurance, and supply chain finance."

**Check**:
- [ ] Market size is reframed through the insight
- [ ] Expansion path is logical, not speculative
- [ ] Numbers are defensible

---

### Slide 6: Business Model (Keep It Simple)

**Goal**: Show you can make money. One model only.

**Structure**:
1. **How you charge** (one sentence)
2. **Unit economics** (if you have them)
3. **Why customers pay** (value captured)

**Good Example**:
> "0.15% per transaction + $2k/mo platform fee. We save customers $2,500 per transaction; they pay us $200. That's an 8x ROI."

**Check**:
- [ ] One primary revenue model
- [ ] Customer ROI is clear
- [ ] Pricing is simple enough to remember

---

### Slide 7: Team (Why You See This)

**Goal**: Show why your background gives you unique access to this insight.

**Structure**:
1. **Founder story**: How you discovered the insight
2. **Relevant experience**: Not credentials, but proof you can execute
3. **Key hire**: One person who validates the vision

**Good Example**:
> "I ran payments compliance at Ant Group for 8 years. I watched $2B in settlements get delayed because 3 banks were checking the same data. My co-founder built the compliance API at Stripe. We literally built this before — now we're doing it for everyone."

**Check**:
- [ ] The team story explains *why you see the insight*
- [ ] Experience is relevant, not prestigious
- [ ] There's a narrative, not a resume

---

### Slide 8: The Ask

**Goal**: Be direct. YC founders ask for money like they ask for coffee.

**Structure**:
```
Raising $X.
To achieve [3 milestones].
Current runway: X months.
```

**Good Example**:
> "Raising $1.1M pre-A.
> 3 milestones: $200k MRR, EU regulatory approval, 3 enterprise pilots.
> Runway: 8 months."

**Check**:
- [ ] The ask is a number, not a range
- [ ] Milestones are specific and time-bound
- [ ] Runway is disclosed (YC investors respect honesty)

---

### The YC One-Page Memo Format

For very early stage (pre-launch), YC sometimes prefers a memo over a deck:

```
[Company] — [One-line description]

Insight: [What you believe that others don't]

Problem: [What broken thing your insight reveals]

Solution: [How you fix it]

Traction: [What evidence you have]

Team: [Why you]

Ask: [How much, on what terms]

Use of funds: [3 specific allocations]
```

**Max 500 words.** If you can't explain it in 500 words, you don't understand it.

---

### Self-Review Checklist

- [ ] Can I state the insight in one sentence?
- [ ] Is the insight contrarian (smart people disagree)?
- [ ] Does every slide serve the insight?
- [ ] Is the deck <8 slides?
- [ ] Did I explain "why now"?
- [ ] Is traction validating the insight (not just growth)?
- [ ] Is the team story about *why we see this*, not *what we've done*?
- [ ] Is the ask direct and specific?

### Anti-Patterns to Avoid

1. **The "Me-Too" Insight**: "AI is changing everything." Not contrarian enough.
2. **The Feature List**: More than 3 features means you haven't found the core insight.
3. **The Credential Dump**: YC doesn't care about your MBA. They care about what you built.
4. **The Hand-Wavy Market**: "The market is huge." Reframe it through your insight.
5. **The Range Ask**: "We're raising $1–2M." Pick a number.

### YC-Specific Tips

1. **Demo if possible**: A 60-sec screen recording beats 10 slides.
2. **Talk about users**: YC partners want to hear "users love us because..."
3. **Be honest about risks**: "We don't know if enterprise sales cycle is 3 or 6 months. We're testing both."
4. **Show iteration**: "We started as X, but users kept asking for Y, so we pivoted."
5. **Know your numbers cold**: If you say 12% MoM growth, be ready to explain the monthly breakdown.

## Guardrails

**Anti-patterns**
- NEVER 使用"Me-Too"洞察——"AI将改变一切"不够反共识
- Do NOT 列出超过3个功能——意味着你还没找到核心洞察
- NEVER 堆砌学历——YC不在乎你的MBA，他们在乎你做过什么
- Do NOT 含糊地说"市场很大"——通过洞察重新定义市场
- NEVER 给出范围式ask——"我们融1-2百万美元"，选一个数字
- Do NOT 超过8页幻灯片——YC deck短是因为洞察应该足够有说服力
- NEVER 编造数据——每个数字必须有来源或可辩护

**Constraints**
- 输出长度5-8页幻灯片，或1页memo（最多500字）
- 语气：直接、对话式、略带不羁
- 一切内容服务于核心洞察
- 洞察必须反共识且具体
- traction必须验证洞察（不只是增长）
- 团队故事必须解释"为什么你能看到这个洞察"
- ask必须是具体数字而非范围
- runway必须披露（YC投资人尊重诚实）

## Related Skills

- **sequoia-structured-bp** — 红杉风格的严谨数据驱动框架，适合A轮及以后、机构投资人
- **raskin-narrative-bp** — Andy Raskin风格的故事驱动叙事框架，适合现场演示、Demo Day和路演比赛
- **pitch-deck-creator** — 框架内容完成后，从结构化JSON生成可编辑的10页PPTX文件

## About UniqueClub

Part of the UniqueClub toolkit.
🌐 https://uniqueclub.ai