Concise Summary简洁概述
YC distills a decade of partner advice into a compact set of principles: launch immediately, do unscalable things to learn from early customers, and find 90/10 shortcuts instead of chasing perfection.
The throughline is that growth, fundraising, and competitor-watching are all distractions unless the founder has first nailed the loop of shipping code and talking to users.
YC 把十年间合伙人的经验浓缩成一套精简原则:立刻发布产品、靠无法规模化的手段学习早期客户需求、用 90/10 捷径代替对完美的追求。
核心逻辑是:除非创始人先打磨好写代码与用户对话的循环,否则增长、融资、盯着竞争对手统统是分散注意力的干扰项。
Infographic信息图
Ship now, iterate with users
立刻发布,靠用户对话迭代
Launch before the product feels ready — the only way to learn what customers actually need is watching them use something real. "Minimum lovable" beats "eventually perfect" because feedback loops compound faster than polish does.
在产品还「不够好」时就发布,因为唯一能真正搞清客户需求的方式,就是让他们使用真实产品。「最小可爱化」胜过「终将完美」,因为反馈循环的复利速度远快于打磨速度。
Do things that don't scale
先做无法规模化的事
Airbnb's founders personally photographed listings — a tactic that breaks past 10 customers but taught them what an active marketplace needs. Manual, unscalable work in the earliest phase is a learning tool, not a shortcut to skip.
Airbnb 创始人亲自为房源拍照——这套办法服务到第十位客户就会崩溃,但它教会了他们打造活跃市场所需的一切。早期阶段手工、不可规模化的工作是学习工具,而非该绕开的权宜之计。
90/10 over 100/never
90/10 方案优于永远做不完的 100 分方案
Paul Buchheit's heuristic: find the version that captures 90% of the value for 10% of the effort. A shippable 90-point solution today beats a hypothetical 100-point solution that takes years and may never ship.
Paul Buchheit 的经验法则:用 10% 的精力找到能实现 90% 价值的方案。今天就能上线的 90 分方案,胜过耗时数年、甚至可能永远交付不了的 100 分方案。
Growth is a result, not a lever
增长是结果,不是可以先拉的杠杆
Growth without product-market fit just burns cash and exposes poor retention — PB's analogy: taking 80 cents from a customer and giving back a dollar is never smart. Growth follows building what's wanted, talking to users, and iterating — not the reverse.
没有产品市场契合的增长只会烧钱,并暴露出糟糕的留存——PB 的比喻是:从客户那里收 80 美分、退还一美元,从来都不明智。增长是打造用户想要的产品、与用户对话、持续迭代之后的结果,而不是可以提前触发的前提。
Detailed Summary详细解读
YC's launch-first doctrine rests on a falsifiable claim: you cannot fully understand a customer problem from the outside, only by watching real usage. The bar is "minimum lovable" rather than MVP in the weak sense — the product must clear a threshold of value that outweighs its flaws, not merely exist. This reframes perfectionism as a form of risk-avoidance that delays the only feedback that matters.
"Do things that don't scale" is often misread as a growth hack; it's actually a learning mechanism. The Airbnb photography example works because the founders weren't optimizing conversion at scale — they were extracting qualitative signal from direct contact that no dashboard could provide. The tactic is explicitly disposable: it's meant to break once the model is understood.
The 90/10 heuristic and the single-metric discipline are two sides of the same anti-sprawl argument. User interviews generate infinite backlogs; without a forcing function, founders default to building everything mentioned rather than the one thing blocking the most urgent problem. Anchoring decisions to one or two metrics is presented as a discipline against this drift, not just a reporting convenience.
The growth-follows-fit argument is the piece's sharpest critique of a common failure mode: startups that scale spend before validating retention or unit economics. The 80-cents-for-a-dollar analogy makes the mechanism concrete — growth that requires subsidizing usage below cost accelerates cash burn without building a durable business, regardless of top-line trajectory.
The piece closes on softer but load-bearing claims — nearly every startup looks broken internally, competitors rarely kill you (self-inflicted mistakes do), and co-founder conflict is a leading cause of early failure. These aren't tactical but psychological: they're meant to recalibrate what founders should actually worry about versus what merely feels urgent.
YC 的「先发布」原则建立在一个可证伪的判断上:创始人无法仅凭外部推测彻底理解客户问题,只能通过观察真实使用来获取。门槛是「最小可爱化」而非弱意义上的 MVP——产品必须跨过一个价值阈值,让它的好处压过缺陷,而不是仅仅存在即可。这把完美主义重新定义为一种延迟唯一有效反馈的风险规避行为。
「先做无法规模化的事」常被误读成增长黑客技巧,实际上它是一种学习机制。Airbnb 拍照的例子之所以有效,不是因为创始人在规模化地优化转化率,而是因为他们通过直接接触获取了任何仪表盘都给不出的定性信号。这套战术本就是一次性的——它注定会在模式被摸清后失效并被淘汰。
90/10 法则和「只盯一两个指标」的纪律,其实是同一种反「摊子铺太大」论证的两面。用户访谈会生成无穷无尽的需求清单;如果没有强制收敛机制,创始人默认会去做被提及的每一件事,而不是解决那个最紧迫的问题。把决策锚定在一两个指标上,被呈现为对抗这种漂移的纪律,而非仅仅是汇报上的便利。
「增长是契合之后的结果」这一论证,是全文对一种常见失败模式最犀利的批评:在验证留存或单位经济效益之前就扩大投放的初创公司。「收 80 美分退一美元」的比喻让这套机制变得具体——需要靠低于成本补贴使用的增长,无论顶线曲线多好看,都只会加速烧钱,而不会构建出可持续的生意。
文章结尾抛出了几条更软但同样关键的判断——几乎每家初创公司内部看起来都一团糟、竞争对手很少能真正杀死你(自杀比他杀常见得多)、联合创始人矛盾是早期失败的主要原因之一。这些不是战术性建议,而是心理层面的校准:意在重新调整创始人「该担心什么」与「感觉紧迫但其实不重要」之间的判断。
FAQ常见问答
Isn't "launch immediately" bad advice for regulated or high-stakes products (medical, financial, infra)?对受监管或高风险产品(医疗、金融、基础设施)来说,「立刻发布」是不是个坏建议?
The essay doesn't address this — it implicitly assumes low-stakes software where iteration is cheap and reversible. For domains with safety, compliance, or irreversible-failure costs, "minimum lovable" needs a higher bar or a different launch strategy (limited pilots, staged rollout).
文章并未讨论这一点——它默认迭代成本低、可逆的软件场景。对于涉及安全、合规或不可逆失败成本的领域,「最小可爱化」需要更高门槛,或采用有限试点、分阶段上线等不同策略。
How do you reconcile "do unscalable things" with "do less, but excellently"?「先做无法规模化的事」和「少而精」这两条建议如何调和?
They apply at different stages: unscalable manual work is for learning what to build pre-fit; "do less" is about focus once you know the target. The common thread is avoiding premature infrastructure — neither scaling systems nor scattering effort across features before the core problem is validated.
这两条对应不同阶段:无法规模化的手工劳动用于在契合之前弄清楚该做什么,「少而精」则是在明确目标之后聚焦资源。共同点是避免过早建设——无论是过早搭建规模化系统,还是在核心问题被验证前把精力分散到一堆功能上。
Does "ignore competitors" hold once a market has an incumbent with strong distribution?当市场已有拥有强大分发渠道的巨头时,「别管竞争对手」这条建议还成立吗?
The piece itself hedges: it says competition matters eventually, just rarely in year one or two. It doesn't specify what changes that calculus (e.g. platform risk, incumbent copying), so founders in adjacent-to-incumbent markets should treat this as a default, not an absolute.
文章本身留了余地:它承认竞争终究重要,只是通常不会在头一两年内成为关键。但它没说明什么情况会改变这一判断(比如平台风险、巨头抄袭),因此身处巨头邻近赛道的创始人应把这条建议当作默认设置,而非绝对真理。
Is the 90/10 heuristic just a rebrand of scope-cutting, or does it add something new?90/10 法则只是「削减范围」的另一种说法,还是有新意?
It adds a search discipline: rather than cutting scope arbitrarily, you actively hunt for the specific version that preserves most of the value. The distinction is between subtraction (cutting features) and substitution (finding a cheaper mechanism for the same outcome).
它多了一层「主动搜索」的纪律:不是随意削减范围,而是主动去寻找那个能保留大部分价值的特定方案。区别在于「减法」(砍功能)与「替代」(为同一结果找到更廉价的实现机制)之间的不同。
Who is this advice not for?这套建议不适合谁?
Founders in capital-intensive or hardware-heavy businesses, where "launch fast and iterate" collides with long manufacturing or regulatory cycles, will find limited direct applicability — the essay's examples (Airbnb, Twitch, Dropbox) are all software-native.
对于资本密集型或硬件密集型企业的创始人来说,「快速发布、快速迭代」会撞上漫长的制造或监管周期,这套建议的直接适用性有限——文中举的例子(Airbnb、Twitch、Dropbox)全部是软件原生公司。
In-depth Analysis · Pros & Cons深入解读 · 优缺点
This is YC's own distilled canon of startup advice — not a single essay but a synthesis of a decade of partner wisdom compressed into a small set of load-bearing principles. It functions less as a how-to and more as a diagnostic checklist for founders who feel stuck or scattered.
这是 YC 官方对创业建议的自我提炼——不是一篇单独的文章,而是把十年间合伙人智慧压缩成的一小套「承重」原则。它更像是给感到卡壳或分心的创始人用的诊断清单,而非操作手册。
- Concrete, named examples具体、点名的案例Airbnb's photography hack, Twitch's focus on gamers over pirates, and PB's 90/10 heuristic give abstract advice testable, memorable shape rather than leaving it as platitude.Airbnb 的拍照战术、Twitch 聚焦游戏玩家而非盗版用户、PB 的 90/10 法则,让抽象建议变得可检验、可记忆,而不是停留在空泛口号层面。
- Internally consistent throughline内在逻辑一致Every principle traces back to one mechanism: prioritize direct customer signal over indirect proxies (conferences, fundraising, competitor-watching, big-company deals).几乎每条原则都能追溯到同一机制:优先获取直接的客户信号,而非会议、融资、盯竞争对手、大公司合作等间接代理指标。
- Names the counter-forces explicitly明确点出反作用力Sam Altman's "fake work" concept names the psychological trap of doing interesting-but-unimportant tasks — a rare moment of self-critique about why good advice is hard to follow.Sam Altman 提出的「虚假工作」概念,点出了做「有趣但不重要」之事的心理陷阱——这是文中少见的、反思「为何好建议难以执行」的自我批评时刻。
- Synthesizes a decade of essays into one map把十年文章合并成一张地图By citing 11 separate essays (Paul Graham, Michael Seibel, Sam Altman, Geoff Ralston), it functions as a curated index, saving readers from having to reconstruct the canon themselves.文章引用了十一篇独立文章(Paul Graham、Michael Seibel、Sam Altman、Geoff Ralston),相当于一份精选索引,省去了读者自行拼凑整套体系的功夫。
- Survivorship-biased examples幸存者偏差的案例选取Airbnb, Twitch, and Dropbox are used to justify "do unscalable things" and "low valuation is fine" — but these are the handful of outliers YC funded that worked; failed companies that followed the same playbook go unmentioned.用 Airbnb、Twitch、Dropbox 来论证「先做无法规模化的事」和「估值低没关系」——但这些只是 YC 投中的少数成功个例,同样遵循这套打法却失败的公司并未被提及。
- No falsification conditions given没有给出证伪条件Advice like "ignore competitors" or "growth follows fit" is stated as near-universal without specifying the boundary cases (winner-take-all markets, platform dependency) where the opposite might be true.「别管竞争对手」「增长跟随契合而生」这类建议被当作近乎普适的真理来陈述,却没有说明反例边界(赢家通吃市场、平台依赖等)何时会成立相反的结论。
- Software-centric blind spot软件中心的视野盲区Every example and the entire launch-fast framework assumes near-zero marginal iteration cost; the essay never addresses hardware, biotech, or regulated industries where the advice may actively mislead.所有案例和整套「快速发布」框架都默认迭代边际成本接近零;文章从未涉及硬件、生物科技或受监管行业,而在这些领域,这套建议可能会产生误导。
- Advice-giver incentive not addressed未处理建议方本身的激励问题YC benefits from founders who move fast and raise quickly (shorter time to next batch of metrics/exits); the essay doesn't acknowledge this structural incentive when recommending speed over caution.YC 本身受益于创始人快速行动、快速融资(能更早看到下一批指标或退出);文章在建议「快于稳」时,并未承认这一结构性激励可能带来的偏向。
Read this as a compact diagnostic checklist, not a universal law: it's most useful for early-stage, low-regulatory-risk software founders who are drowning in options and need a forcing function to focus on shipping and talking to users. Treat the specific examples as illustrations of a mechanism, not proof that the mechanism always works — and be skeptical of any advice here applied outside its software, low-capital-intensity context.
把这篇文章当作一份精简的诊断清单来读,而非放之四海皆准的定律:它最适合处于早期、监管风险低的软件创始人,在选项过多、无所适从时用来强迫自己聚焦于发布产品和倾听用户。文中的具体案例只是机制的示例,而非证明该机制永远有效——一旦脱离软件、低资本密集度的语境,对这里的任何建议都应保持审慎。
Excerpt原文节选
This is a short excerpt, not the full piece — the complete essay belongs to its original author; please read it in full at the link above.
以下仅为节选，并非全文——完整文章版权归原作者所有，请点击上方链接阅读全文。
A lot of the advice we give startups is tactical; meant to be helpful on a day to day or week to week basis. But some advice is more fundamental. We've collected here what we at YC consider the most important, most transformative advice for startups. Whether common sense or counter-intuitive, the guidance below will help most startups find their path to success.
The first thing we always tell founders is to launch their product right away; for the simple reason that this is the only way to fully understand customers' problems and whether the product meets their needs. Surprisingly, launching a mediocre product as soon as possible, and then talking to customers and iterating, is much better than waiting to build the "perfect" product. This is true as long as the product contains a "quantum of utility" for customers whose value overwhelms problems any warts might present.
[…the source continues — read the rest at the link above]
[……原文更长，完整内容请点击上方链接阅读]
11. Mean People Fail by Paul Graham ↩
我们给予创业公司的建议大多属于战术层面，旨在提供日常或每周的具体帮助。但有些建议则更为根本。在此，我们汇总了 YC 认为对初创企业最为重要、最具变革意义的指导原则。无论这些建议是常识性的还是反直觉的，以下的指引将帮助大多数创业公司找到通往成功的道路。
我们总是告诉创始人的第一件事，就是立即推出产品。原因很简单：这是全面理解客户问题及产品是否满足其需求的唯一途径。令人惊讶的是，尽快推出一个尚不完美的产品，然后与客户沟通并快速迭代，远比等待打造"完美"产品要好得多。只要产品能为客户提供"最小可用价值"——其价值足以掩盖产品可能存在的任何瑕疵——这一原则就始终成立。
一旦启动项目，我们建议创始人做那些不可规模化的行动（参见保罗·格雷厄姆的《做无法规模化的事》）。许多创业顾问过早地劝说初创企业扩张。这将需要建立支撑扩张的技术与流程，若时机未到，只会徒耗时间与精力。这种策略常导致失败，甚至断送初创企业。相反，我们告诉创业者要不惜一切代价获取首批客户，即便依靠人工操作——这种模式连服务十位客户都难以维持，更不用说一百或一千位了。在此阶段，创始人仍在摸索应构建什么产品，而最佳途径就是直接与客户沟通。以爱彼迎创始人为例，他们最初提出为早期客户的房屋和公寓提供"专业"摄影服务，以提升房源对租客的吸引力。接着，他们亲自上门拍摄照片。网站上的房源展示因此改善，转化率得到提高，同时还与客户进行了宝贵的交流。这种做法完全无法规模化，却被证明是学习如何打造活跃市场的关键所在。
与用户交流后通常会得到一份冗长复杂的功能清单。YC 合伙人 Paul Buchheit (PB)在此情况下总会给出这样的建议：寻找"90/10 解决方案"。也就是说，寻找一种方法能以 10%的工作量/精力/时间实现 90%的目标。只要用心探索，几乎总能找到这样的捷径。最关键的是，能即刻上线的 90 分解决方案，远胜于耗时数年打造的 100 分方案。
随着企业开始成长，各种潜在干扰往往层出不穷。参加行业会议、商务晚宴、与风险投资人或大企业业务拓展部门洽谈（参见保罗·格雷厄姆的《别和业务拓展部交谈》 ），追逐媒体报道等等（YC 联合创始…
[…the source continues — read the rest at the link above]
[……原文更长，完整内容请点击上方链接阅读]
1. 保罗·格拉厄姆：《做那些无法规模化的事》 ↩ 2. 不要与企业开发部门交谈 by Paul Graham ↩ 3. 如何避免失败 by Jessica Livingston ↩ 4. 后 YC 低谷期 by Sam Altman ↩ 5. 你不想招揽的用户 by Michael Seibel ↩ 6. 《真正的产品市场契合度》作者：Michael Seibel ↩ 7. 《单位经济效益》作者：Sam Altman ↩ 8.