故事 · Cooper 的「门」 Origin Story · Cooper's Gates
1980 年代,学者 Robert Cooper 研究了上千个新产品项目,发现一个惊人规律: 失败的项目往往不是死在技术上,而是死在「该砍不砍、一路烧钱到底」。 于是他提出 Stage-Gate:把开发切成几个阶段,每段之间放一道「门」,让高管像投资人一样定期复盘 —— 值得就放行、烧更多钱;不值得就果断砍掉,把钱留给更好的项目。门,不是官僚的关卡,而是止损的开关。 一个好的阶段门流程,敢于在早期杀死自己最爱的项目。 In the 1980s, the academic Robert Cooper studied thousands of new-product programs and uncovered a striking pattern: most failed projects didn't die from bad technology — they died from never being killed, burning cash all the way to launch. So he proposed Stage-Gate: split development into phases, plant a gate between each one, and have leaders review like investors at a regular checkpoint — fund the worthy ones, shut the rest down fast, and free that money for better bets. A gate isn't bureaucracy; it's a stop-loss switch. A healthy phase-gate process is one that is willing to kill its own favorite project early.

1 开发漏斗:四道门,层层淘汰 The Development Funnel: Four Gates, Layer by Layer

—

2 早杀早省:在哪道门砍,差别巨大 Kill Early, Save More: Where You Kill Makes a Huge Difference

阶段越往后,单个项目累计烧的钱越多。在第 1 道门砍只浪费一点点;拖到最后一道门才砍,几乎把全程的钱都白烧了。这就是阶段门要「早评审、早决断」的根本原因。 The later the phase, the more cash a project has already burned. Kill at gate 1 and you waste a sliver; drag it to the final gate and you've torched the entire program budget. That is precisely why phase-gate insists on early review, early decision.

3 现实里的阶段门 Phase-Gate in the Real World

研发投资组合:把有限预算像基金一样分散下注,定期复盘,赢家加注、输家止损。 R&D portfolio: Treat the R&D budget like a fund — diversify bets, review on a cadence, double down on winners, cut your losses on the rest.
go / kill 决策:每道门用统一标准(技术、市场、财务)评审,不靠拍脑袋,也不靠谁声音大。 Go / kill decision: Every gate uses the same criteria — technical, market, financial — so calls aren't made by gut feel or by whoever talks loudest.
漏斗淘汰:很多想法进、少数产品出。淘汰不是失败,是把资源从平庸项目腾给优秀项目。 Funnel culling: Many ideas in, few products out. A kill isn't a failure — it's how you free up resources from mediocre projects so the great ones can win.
资源聚焦:与其十个项目都半死不活,不如集中喂饱三个,让它们真正做成。 Resource focus: Ten half-starved projects beat nothing — but three well-fed ones actually ship. Concentrate the spend.
一句话In One Line
阶段门的本质,是把「要不要继续投钱」这个最难的决定,从感情用事变成有节奏、有标准的例行评审。 人天生舍不得砍掉投入过心血的项目(沉没成本谬误),而门强制你在每个节点冷静问一句:「如果今天才知道这些,我还会从头启动它吗?」 答案是否,就该 kill。记住那条成本曲线:越早砍,省得越多 —— 一个敢于早期杀死自己项目的组织,才有资源去赢真正的大仗。 At its core, phase-gate turns the hardest question — should we keep funding this? — from an emotional call into a scheduled review with shared criteria. Humans are wired to protect what they've poured effort into (sunk-cost fallacy); a gate forces you to ask, coldly, at every checkpoint: "If I were learning about this project today for the first time, would I start it from scratch?" If the answer is no, kill it. Remember the cost curve: the earlier you kill, the more you save. Only an organization brave enough to kill its own projects early ever has the resources to win the battles that really matter.
常见误用Common Mistakes
门变成走过场,从不 kill。一道从不砍项目的门等于没有门 —— 必须有真实的 kill 率才有止损效果。 Gates turned into rubber stamps that never kill. A gate that never kills isn't a gate at all — without a real kill rate, the whole stop-loss mechanism is dead.
因沉没成本不舍得砍。已花的钱回不来,评审只看「从现在起还值不值得投」。 Refusing to kill because of sunk cost. Past spend is gone — the only question that matters at a gate is "is it worth more money from this point forward?"
门设得太多太重,拖垮速度。门要轻量、聚焦关键判据;评审是为加速决策,不是制造官僚。 Too many gates, too heavy, killing speed. Keep gates light and focused on the decisive criteria — reviews exist to accelerate decisions, not to manufacture bureaucracy.

阶段门流程