Superhuman 成功的秘诀 | Rahul Vohra（CEO 兼创始人）
Superhuman 成功的秘诀 | Rahul Vohra（CEO 兼创始人）
逐字稿
精彩预告
Lenny Rachitsky： 我们来谈谈 Product-Market Fit（产品市场契合度）。
Rahul Vohra： 你必须刻意不去回应许多早期用户的反馈，与此同时你又要在认真倾听人们需求的基础上，打造人们想要的东西。我们存在的意义就是创造人们想要的产品，但不能是满足所有人。问题就变成了：你怎么听他们的话？即便听了他们说了什么，你关注什么、忽略什么？这里的诀窍是——
Lenny Rachitsky： 你并没有在做许多 CEO 认为自己应该花时间去做的事情。许多 CEO 认为需要在招聘或组织建设上花时间，而你刻意地决定：“我要把时间花在产品、营销和设计上。”
Rahul Vohra： 这是一种我称之为 Switch Lock 的技巧。它源于一个观察：你的日历记录的是你打算做什么，但只有你留下的工作痕迹才真正描述了你实际做了什么。我们如何捕捉这一点？于是我想出了下面这个主意——如果我就做自己想做的事呢？
Lenny Rachitsky： 你职业生涯和人生中最重要的转折点是什么？
Rahul Vohra： 我学到了病毒式传播背后真正的秘密。世界上不存在真正意义上的病毒式产品。那么真正的秘密是什么？它是——
嘉宾介绍
Lenny Rachitsky： 今天的嘉宾是 Rahul Vohra。Rahul 是 Superhuman 的创始人兼 CEO，也是我遇到过的最有思想、最有洞察力、表达最清晰的创始人之一。正如你将在我们的对话中看到的，很难不被 Rahul 的叙事能力所吸引，也很难不被他对如何打造伟大产品和团队所提出的深刻见解所打动。
这一期节目适合所有想要培养产品直觉、帮助团队加速推进、学习如何从第一性原理出发进行思考的人。同时也适合想了解 Superhuman 非常独特的公司建设方式的人——包括他们为什么多年来手动引导每一位新用户，又为什么决定停止这样做；为什么在寻找 Product-Market Fit 的过程中忽略了大部分客户反馈，以及你如何运用他的方法为自己的公司寻找 Product-Market Fit。还有游戏设计在打造伟大产品中的力量、一种非常反共识的定价策略、Rahul 在基于 AI 和 LLM 构建规模化产品方面的所学所得，以及更多内容。
非常感谢 Ed Sims、Conrad Irwin、Bell Trenchard 和 Gaurav Vohra 为本次对话提供的问题和话题建议。如果你喜欢这档播客，别忘了在你最喜欢的播客应用或 YouTube 上订阅和关注。另外，如果你成为我通讯的年度订阅者，可以获得一年免费的 Superhuman，立即开始使用。你还可以获得一年免费的 Notion、Perplexity Pro、Granola 和 Linear。详情请访问 lennysnewsletter.com。
话不多说，为你带来 Rahul Vohra。
Rapportive 与 LinkedIn 收购
Lenny Rachitsky： Rahul，非常感谢你来到这里。欢迎收听播客。
Rahul Vohra： 你好你好，谢谢你的邀请，Lenny。
Lenny Rachitsky： 我有很多问题想问你，我们会有很多话题要聊。我想先从你创立 Superhuman 之前的经历谈起。在准备这次对话的时候，我其实问过你：你职业生涯和人生中最重要的转折点是什么？你告诉我，除了创办 Superhuman 之外，就是把你之前的公司 Rapportive 卖给了 LinkedIn。那就从这里开始吧。那段经历是什么样的？人们对这个阶段的你有哪些不了解的事情？为什么它如此关键？
Rahul Vohra： 对于不太了解的人，Rapportive 是我的上一家公司。它是第一个扩展到数百万用户的 Gmail 扩展。简单来说，在 Gmail 的右侧，我们会显示人们的照片、工作单位、最近推文的链接、LinkedIn 个人资料以及他们在网上的其他动态。所以如果你在做招聘、市场营销、销售或 BD，超级有用。结果是我们不知怎么地把 LinkedIn 的大部分日活跃用户吸引到了这一个免费应用上，然后我最终把它卖给了 LinkedIn。正如你所说，那毫无疑问是我在创办 Superhuman 之前职业生涯中最关键的经历。
如果当时我知道我们已经把 LinkedIn 大部分活跃用户聚集到了一个应用上，我会卖一个高得多的价格。但真正关键的转折点其实是我与谁共事。因为我向 LinkedIn 的 Growth 负责人 Elliot Shmukler 汇报。他负责将 LinkedIn 从 2500 万用户增长到我加入时的超过 2.5 亿用户。在我和他的第一次一对一谈话中，我学到了病毒式传播背后真正的秘密——一个大提示：它与病毒式传播机制无关。那次收购经历整体上给了我时间去思考下一步做什么，也给了我真正放手一搏所需的资源。
Lenny Rachitsky： 好，我得接着你抛出的这个线索追问——病毒式传播的秘密到底是什么？你在那里学到了什么？
病毒式传播的真相
Rahul Vohra： 在我第一次一对一谈话中，我坐下来对 Elliot 说：“嘿，我是来学习的。请把你所知道的关于病毒式传播的一切都教给我。“他说：“好吧，不想戳破你的幻想，但根本不存在真正意义上的病毒式产品。“我说：“你什么意思？那你怎么解释 Facebook？怎么解释 LinkedIn？“他说：“我的意思是，没有任何一款应用能在任何真正持续的时间段内维持大于 1 的病毒系数。“即使是鼎盛时期的 Facebook，病毒系数也只有大约 0.7。他告诉我那大概持续了一年，也就是说一个人大约能带来 0.7 个新用户。
我继续追问：“那 Elliot，通讯录导入呢？“这是 LinkedIn 因此出名——或者说声名狼藉——的功能之一。你可以导入通讯录，然后它会邀请你通讯录里所有碰巧是 LinkedIn 成员的人，最终它会邀请所有人加入 LinkedIn。他说：“那是一个很棒的功能，但你得记住，不是每个人都会一直使用它。“所以即便是那个功能，终身病毒系数也只有大约 0.4，而这已经算不错了。
0.4 对于一个病毒式功能来说算不错，0.6 是很好，0.7 左右就非常惊人了。那已经是和当时 Facebook 并驾齐驱的级别了。所以我说：“好吧，那所有这些东西从定义上讲都会逐渐衰减。会有一个渐近线。没有任何病毒式传播机制能持续复合增长。这其实也说得通，如果东西就这么一直增长下去那才有点荒谬。那么病毒式传播真正的秘密到底是什么？“他说：“是口碑。是你无法衡量的、不属于机制、不属于功能的病毒式传播。是一个用户自发地把你的产品告诉另一个用户。“这深刻地塑造了我对增长和病毒式传播的思考方式。从那以后，它深刻影响了我们在 Superhuman 所做的一切，也深刻影响了我对品牌增长的思考方式。
Lenny Rachitsky： 哇，你真是个出色的故事讲述者。我在这里听着，完全被吸引了，“他接下来会说什么？“太精彩了。我恰好有一篇文章要链接过来，和你说的非常一致，标题大概是《病毒式传播是一个神话》。它是基于——我忘了——一本书，书中对真正的病毒做了大量研究。结果发现病毒实际上并不是以指数方式传播的，是某一个人把病毒传给了很多人，然后这种情况不断重复。
实际的数据显然显示的就是这样。我很好奇你是否也发现了同样的现象——当人们认为一款应用病毒式传播时，其实是某个拥有巨大平台的人分享了它，然后他的受众采用了它，这就是一对多，然后这种情况发生几次，看起来就像在病毒式传播，但实际上是个人对多人，而不是多人对多人。你怎么看？
“鲸鱼”用户与协作产品
Rahul Vohra： 是的，我们确实发现有”鲸鱼”用户——借用地游行业的术语——一个人会负责邀请 25、50、100 个人，他们可能有各种各样的动机。在 Superhuman，作为个人订阅用户，如果你推荐了其他人并且对方注册了，你们双方都能获得一个月的免费使用权，如果你是自掏腰包的话，这是一个很好的激励。
我们有用户发送了数百封邀请，有些人基本上因为邀请的人数太多，已经拥有了终身免费的 Superhuman。但当然，这种激励方式在公司或团队内部不一定奏效，因为最终是公司在为产品买单，所以你必须为这些人设计新的动机。在这一点上，没有什么能替代一个真正多人协作的产品。这正是过去大约两年里我们对 Superhuman 所做的重大进化之一。
去年年初我们发布了所谓的 Superhuman 2.0。基本思路是，我们看到几乎所有知名应用都默认变成了协作式的——Figma、Notion、Loom，这些默认都是多人协作的。然而电子邮件，这个我们所有人使用频率最高的工具，甚至比 Slack 还高，却仍然牢牢停留在它的单人使用起源上。
Lenny Rachitsky： 我想回到你刚才提到的一点，我认为它对你分享的内容至关重要——口碑如此重要。人们谈论各种病毒式功能、通讯录分享等等。而你的观点是，这些能带你去到某个阶段，但真正帮助一款偏消费端产品传播的是口碑，是人与人之间的分享。那么问题就来了，怎么做到这一点？我们会讨论你做的很多事情，让 Superhuman 成为一款人们愿意分享的产品，但归根结底就是做一款人们愿意分享的东西。这几乎就是定义本身。那么是什么让人愿意分享？因为它足够惊艳，因为它帮助了他们，因为它值得被谈论。
把增长要素写入公司价值观
Rahul Vohra： 事实证明，因为你提到了”值得被谈论”，这正是我们公司核心价值观之一。想想一家公司必须做什么——它必须增长。东西怎么增长？让我们采纳 Elliot 的建议——我相信它是正确的——就是创造人们会分享的东西。你提到了其中一种方式，就是创造人们愿意分享的东西。实际上还有另一种方式，就是简单地创造一些值得被谈论的东西，你用了那个词，而这正是 Superhuman 的核心价值观之一。
我们有”创造愉悦”——创造一些如此令人愉快的东西，真正给人们带来喜悦。我们有”交付卓越品质”——创造一些如此引人注目、如此有吸引力、如此值得关注的东西，让人们忍不住要告诉别人。然后我们有”构建非凡之物”——这是对我们想要构建的东西的有效性或创新性的衡量。这是另一个诀窍，就是把增长的这些原材料直接融入你的公司价值观里。
Lenny Rachitsky： 我不知道这也是你们的价值观之一。这完全说得通。好，我们会回到这个话题，因为我觉得……你如何看待产品、如何看待打造产品的公司，有太多值得学习的地方。但我想从这次对话的缘起开始说起。
Product Hunt 的 CEO Rajiv 几个月前发了一条推文，他发了这个，如果你在 YouTube 上看我们会展示这条推文：“Superhuman 的产品交付速度感觉最近像是换了一个档。有其他人注意到吗？“我看到后心想：“我完全注意到了。感觉就是各种功能接连发布，AI 这 AI 那的，感觉像是一家全新的公司。“我在推文里标记了你。我说：“嘿 Rahul，发生了什么变化？“你回答了几点，让我清楚地看到你做了很多值得学习的东西。
因为很多公司都处在这个阶段——“事情推进得不如我们期望的那么快。我们以前快多了，以前各种功能不断上线，现在不行了。“所以我觉得这是一个非常好的真实案例研究和说明性范例，我们可以来分析。那让我问你这个问题：你注意到了什么，让你意识到 Superhuman 需要做出改变？然后你实际做了什么改变，对你交付和加速迭代的能力产生了最大的影响？
感知到的放缓
Rahul Vohra： 我想我们注意到的是这样一种感受——我们自己最先感受到了，随后也开始从市场、从用户、从客户那里听到同样的声音：我们变慢了。作为创始人，作为 CEO，这是你最不想听到的话。毕竟，加速才是我们的职责。当我追问人们说”变慢了”到底是什么意思时，他们当然不是指产品本身——产品运行速度并没有变慢——而是指功能交付的节奏似乎放缓了。
要拆解这个问题，我认为首先要定义我们所说的”放缓”是什么意思。有一种放缓，在某些领域是不可避免的；还有一种放缓，则完全可以避免。我们实际上两种都经历了。先说不可避免的放缓。你可以把公司里构建的任何东西归入两个类别之一：方案深化和市场拓展。方案深化是指让产品对现有用户变得更好，但不扩大用户覆盖范围。而市场拓展是指让产品覆盖更多用户，但不改变产品本身。
方案深化与市场拓展
有些领域、有些市场、有些平台，市场拓展非常快、非常容易。但也有些领域——邮件就是其中之一——市场拓展非常困难、非常缓慢。但我们起步时非常专注：只支持 Gmail，只在 Web 端。在那几年里，我们可以把每一分研发精力、每一美元工程投入都倾注在方案深化上，为现有用户做得更好。用户当然喜欢。我们就是这样达到产品市场契合度的。大多数创业公司也是如此。
但到了某个节点，几乎每家公司都不得不开始投资市场拓展。比如，愿意使用一个新的 Gmail 前端但没有移动应用的用户群体，确实存在，但相对较小。这是每一个新邮件创业公司迟早都会学到的道理。为了持续增长，你必须添加 iOS 应用，然后是 macOS 应用，然后是 Windows 应用，然后是 Android 应用。接着你很快就会想支持 Office 365。但这不是一件事，实际上是三件事，因为你需要支持桌面端的 Office 365，然后是 iOS 端，然后是 Android 端。说起来容易做起来难。
我觉得 Superhuman 如今对这些 API 的了解，真的没有其他任何公司知道，而且我绝不希望这发生在哪怕最糟糕的敌人身上。快进到今天，Superhuman 可以在你工作所在的一切平台上运行——Gmail、Outlook、Mac、Windows、Web、iOS、Android 的每一种组合。这实际上形成了一道非常好的技术护城河。几乎没有其他邮件应用能做到这一点。这背后是多年密集的投入。我们后面可能还会聊到，但这是我们能够向企业销售的主要原因之一，因为我们现在知道每个人都能使用它。
但难点在于，当你在做市场拓展的时候，你并没有在做方案深化，所以外界感知到的产品迭代速度可能会下降。通过一些明智的技术决策可以规避其中一部分，但大部分时候你只能硬啃，而且熬过去是值得的。
可避免的放缓与管理结构重组
然后是那种可以避免的放缓。如果我没记错我在回复 Rajiv 推文时说的，我指的就是这种。在那情况下，问题出在我们的管理结构，也就是谁负责什么。
当我们招聘最初的执行团队时，我遵循了非常传统的做法。最终我手下有一批 VP，以及八个——我想是八个直接汇报线，也许甚至九个。我以为就该这样。我以为创业公司就是这样扩展的。但任何经历过的人都知道，八个直接下属非常多。大量的招聘，大量的目标设定，大量的 OKR，大量的绩效沟通，当然还有大量的解雇。没有哪个 CEO 能一次就把执行团队配齐。我本应花在那些我认为自己真正能做到世界级水平的事情上的时间——产品、设计、技术、营销——这些时间开始急剧减少，结果是整个组织开始变慢。
不巧的是，我一直在非常细致地追踪自己的时间，我有一种很疯狂的时间追踪方式。有一次我注意到，我每周花在这些我真正能做到世界级水平的领域上的时间只有 6% 到 7%。于是我有两个认识。第一，作为 CEO，一旦公司达到一定规模——我们当时显然已经到了——你其实可以定义 CEO 在你公司应该扮演什么角色。第二，Superhuman 的机会值得公司里每一个人尽可能多地待在自己的天才区里，包括我，也包括其他所有人。我做的第一件事是聘请了一位非常出色的总裁，我的直接汇报线从八个减少到两个，而我花在产品、设计、技术和营销上的时间从每周 6% 到 7% 提升到了大约 60% 到 70%。
时间追踪方法
Lenny Rachitsky： 我来复述确认几点。第一，人们可能觉得你没有以前发布得多，是因为你实际上在建的东西是他们在意范围之外的——比如支持 Office 以及各种他们不需要但企业需要扩展的东西，与 Microsoft、Android 的集成等等。我觉得这个观点非常好——看起来好像什么都没发生，实际上有很多好东西在为其他用户而发生。
然后就是这个关于授权的观点。一个领导者在授权、招聘了所有这些高管之后，会觉得”这不是我想要的。我为什么要这样做？“你以为会加速，结果反而变慢了。这里面有几个线索非常有趣。一个是这个时间追踪的事，我想知道你是怎么做的？你能精确到知道 7% 到 8% 或某个具体数字——你花在想做的事情上的时间竟然那么少——你是怎么做时间追踪的？不用展开太多，大致说说你的方法是什么？
Rahul Vohra： 这是一个我称之为 Switch Log 的技巧。它源于一个观察：你的日历显示的是你以为自己要做的事，但只有你工作的轨迹才能描述你实际做了什么。那么，我们怎么捕捉这个轨迹呢？更进一步，我们怎么创造一个不受日历束缚的工作系统，让你不必受制于某个时间表说你该做什么或不该做什么？于是我产生了这样一个想法：如果我干脆想做什么就做什么呢？如果每次我切换任务时，就在 Slack 上私信我的 EA（Executive Assistant）——发到 Slackbot 也行，只要有个去处——我私信我的 EA，写上”TS:“，然后几个词描述我正在做的事。
这会带来一些变化。你不再需要不断看日历，想着”哦，我是不是该停下手头的事开始下一个了”，你直接做你想做的就好。如果此刻我感觉的是，“天哪，我真的需要为 Lenny 的播客做准备”，那我就去做。如果八分钟后我厌倦了或分心了——有时候确实会这样，因为别的事情突然涌上心头——那我的身体把它涌上来是有原因的。我也修习超觉冥想，所以我非常认同保持觉知、倾听内心涌上来的东西这个理念。
时间去向的洞察
Rahul Vohra： 所以对我来说，去处理那个冒出来的念头是完全可以的，而不必把我的专注力、自律或意志力消耗在我原本以为自己该做的那个事情上。我要做的就是回到 Slack，发一条 “TS: 处理另一件事”。当然，你显然应该按时出席会议。我不是说要无视你的会议、不参加一对一沟通。那些一定要做。我想说的是，在你觉得该做某件事的时间里，做你觉得对的事。然后到周末，你就能看到自己的时间都花在了哪里。
有一段时间我意识到，我那时只花了 5% 的时间在招聘上，而我或许应该花 20% 或 30% 甚至更多的时间在招聘上。但最让我震惊的是，我发现自己花在产品、设计、技术和营销上的时间只有 6% 到 7%。这些都是我非常擅长的事情。我应该要么教别人怎么做，要么自己动手做，要么两者兼有。这可能对我来说是最有价值的事。它让我非常开心、非常愉悦，让我保持敏锐，同时也在推动组织规模化。这就是我们得到那种洞察的方式。一旦你有了这个 Slack 日志，你就可以绘图、制表，看看自己的时间实际上花在了哪里。
Lenny Rachitsky： 太酷了。这显然是一个 App 的机会，或者说 Agent 的机会——你只需要每次告诉它你在做什么。它本质上是在追踪上下文，而我们一直听到的是，要尽量避免上下文切换。
关于上下文切换
Rahul Vohra： 我觉得上下文切换是没问题的。确实有这样一个说法：每被打断一次，大脑平均需要大约 21 分钟才能恢复到被打断之前的工作效率。这当然很重要——毕竟我在做生产力软件，我们设计 Superhuman 的目标就是尽可能减少应用内的干扰和打断。但如果你正在做某件事，脑子里突然冒出一个念头，你总要以某种方式去处理它。有时候我会直接写下来，我的笔记本不在手边，但真的很大。我有一个超大号的，大概是 A4 两倍大的，应该是 A3 素描本，我总是随身带一支 4H 铅笔，所以每当那种念头冒出来，我就草草记下来。或者我干脆停下手头的事去处理那个任务，因为它此刻冒出来是有原因的。
Lenny Rachitsky： 我喜欢你连纸和笔的型号都知道——4H 铅笔、A3 纸。好吧，这会是一个主题。你提到了冥想，你说你练习 TM，所以是早上 20 分钟、20 分钟……你是按那种方式做的，还是做更长的时段？
超觉冥想的实践
Rahul Vohra： 我早上大约做半小时，包括休息时间。其中的身体休息部分对我来说非常重要。20 分钟是正式的超觉冥想，然后 10 分钟休息。我早上做一次，下午三点左右再做一次。
Lenny Rachitsky： 你就是在日历上把这个时间锁定。所有人都知道 Rahul 三点钟就不在了。
Rahul Vohra： 完全正确。我的 EA 知道，日历是他们安排的，他们会确保事情在应该发生的时候发生。他们还知道，没有任何事可以覆盖这个 TM 时间块。没有它，我真的会开始崩塌。但有了它，我能触达一些以前没有的深层能力。我已经坚持了大约四五年，最初我只是觉得更快乐了，有时候一次非常好的冥想之后甚至会感到更欣快。但随着时间推移，我发现自己的专注力在增强。我能把注意力维持在一件事情上更长时间，同时我也变得更有创造力、更有表达力。
这些都是 TM 很有名的附带效果——或者对一些人来说是预期的效果。有趣的是，如果把 TM 和其他形式的冥想做比较，后者在那么多执行功能上并没有产生同样广泛的影响。所以超觉冥想相比其他形式确实有某些特别之处，人们至今仍在试图解开和弄清楚。
Lenny Rachitsky： 如果有人受到启发，想尝试这种冥想方式，你有什么建议——去哪里学？
学习超觉冥想的建议
Rahul Vohra： 当然，有很多。但概括来说，找一个教练来教你。我自己在冥想上有很多次失败的尝试——试过各种 App，从书里学——都没有真正奏效。真正有效的是一对一的教学，教我的人本身也是被一对一地教授的，传承的是瑜伽（Yogic）或王瑜伽（Raja）传统。这个人恰好也做过多次获得风投融资的创始人，所以他非常了解我通常承受的那种压力。而且他的客户也大多在科技行业。如果你在湾区，这个人叫 Laurent Valasek。他运营一个叫 Peak Leadership Institute 的机构，专注于如何让我们过上一种更整合、更完整的生活——将冥想这样的健康实践整合进来，但目的是为了释放生活和事业中的巅峰表现。
Lenny Rachitsky： 谢谢你分享这些，非常实用。我们会在节目说明里附上链接。
招聘总裁
好，我来试着把我们拉回正题。你提到的另一件我觉得非常有趣的事是招聘了一位总裁。很多创始人和领导者听到这里可能会想，“从八个直接汇报人、做所有这些我不想做的事情，变成大部分时间花在产品、设计和营销上——太棒了。“这位总裁从你这里接走了什么？他们的职责是什么？这才让你能做自己想做的事？
Rahul Vohra： 最大的部分是接走了运营工作，以及高管团队和公司其余部分的管理。在 Superhuman，总裁这个角色在运营上极具挑战性，也是一个非常有成长空间的角色。它非常适合那些下一步想当 CEO 的人。原来需要我去招聘和解雇团队、管理和设定目标、进行问责对话——现在这些事由另一个人来做了。
除此之外——因为那不是全部工作——他们还是一个非常强的战略思考伙伴。比如，我们的第一阶段产品——邮件产品——该怎么走？多人协作这条路我们要走多远？我们应该多积极地投入 AI？在有 AI 的世界里，什么样的毛利率是合理的？从财务角度看，我们可以先往下探一探之后再回来吗？我们什么时候该开始做第二个产品？我们的研发战略该怎么想？我们应该继续在湾区招聘，还是像最近很多新员工那样继续在拉丁美洲招聘？要不要考虑其他时区？等等等等。我只是在随口举例，但这个清单真的是无穷无尽的。
另一种理解总裁角色的方式
Rahul Vohra： 另一种理解方式是，总裁这个角色几乎像一个成熟的联合创始人。与我共同创立公司的两个人，Comrade 和 Vivek，他们早就离开了 Superhuman。我们现在是一家有十年历史的公司，而我是那种罕见的、十年后仍然在坚持并且蓬勃发展中的创始人。话虽如此，这段旅程从未变得更轻松，只是变得不同了，而你仍然需要身边有那种联合创始人般的能量。我身边有几个核心人物，他们在自己的岗位上提供着那种能量、那种输入，并且乐在其中。总裁这个角色，绝对是其中之一。
Lenny Rachitsky： 非常有意思，信息量很大。我想分享几点观察，然后我们转向下一个话题。第一点是，帮助你加速前进、做你想做的工作的解决方案是组织设计——这一点很酷。这感觉是一件非常可行的事情。如果你发现自己没有把时间花在想花的事情上，事情推进的速度不如预期，本质上你可以找到人来接手你不想做的事情，调整组织的结构方式，这样就能解决很多问题。这正是你的做法。第二点我觉得也很有意思的是，这里有一个教训：作为创始人，如果你只是感到精力耗尽，或者身边没有你想要的搭档，你可以引入一个人来扮演那个角色。
Rahul Vohra： 完全同意。
对细节的极致关注
Lenny Rachitsky： 好。这个话题比我预期的丰富得多。我想把视角拉远一些。在我和你合作过的人、Superhuman 的投资者交谈时，有两个主题反复出现：一个是逆向思维（在打造公司方面），另一个是对细节的极度关注。我们先聊聊对细节的关注。正如我所说，这在我向人们问起你时被一次又一次地提到。我这里有一段 Ed Sim 的话，他可能是你的第一位投资人。他是你的第一位投资人吗？
Rahul Vohra： 是的，不过 Twitter 上会有一大堆人来抢这个头衔。但为了澄清事实，Ed Sim 确实给 Superhuman 写了前三张支票。
Lenny Rachitsky： 前三张支票？是在连续几轮中投的吗？
Rahul Vohra： 对。这里快速补充一下背景——他和他的搭档 Elliot Durbin 一起运营 Boldstart Ventures。他们对投资二次创业者有特别的兴趣，但也会投资首次创业者，而且他们喜欢应用层和基础设施领域的项目，Superhuman 正好属于这类，所以我们对他们来说是完美的投资标的。他之前还从他之前的基金给 Rapportive 投了一笔钱，我觉得我让他赚了五倍的回报。不算什么值得大书特书的业绩，但足以让他觉得”我要再投这个人一次”。所以我去找他，说：“听着，这听起来可能很疯狂。我要挑战 Gmail。“他说：“你有 deck 吗？“我说：“有，就一张幻灯片，给你。“那张幻灯片上是 Gmail 的截图，大部分内容被涂掉了，旁边写着”我要把那些部分建出来，而且会很棒。”
然后他说：“好，我们投了。我可以把钱打给你吗？“我说：“不行，我连银行账户都还没有。“两天后我带着一个银行账户回来了，他说：“好，我要给你打 75 万美元。“我说：“我甚至不知道拿这笔钱做什么。我不给自己发工资，短期内也不会。我们没有任何员工。我想不出要花钱的地方。这样吧，我就拿 25 万。“他愣了：“什么？“我说：“对，我就拿 25 万。“然后我们开始讨论风险投资的经济学问题。我说：“没事，我们会想清楚的。“几个月后我又拿了 25 万，又过了几个月再拿了 25 万——随着我开始发明各种方式、找到合适的渠道来有效部署资金。
Lenny Rachitsky： 我太喜欢这个故事了。你分享的这些故事我以前从来没听过，全都很精彩。顺便说一下，他来上我们的播客这件事本身就很棒，可能打破了我们不请 VC 的规矩。他跟我分享了一个具体的故事，可能很好地说明了你和你的细节控——他说你自己创建了一套字体，因为现有的字体都不够好。这是真的吗？
Rahul Vohra： 算是吧。我们现在使用的字体是 Adelle Sans 的修改版。故事是这样的：我把所有主要的字体家族都看了一遍，说实话，没有一个是我会称之为真正出色的。这听起来可能有点奇怪。所以，如果你允许我聊聊排版设计和邮件——
Lenny Rachitsky： 请讲。
Rahul Vohra： 我们做的第一件事是，把我们的 UI 用大约 15 种不同风格呈现出来，使用了各个主要字体家族的样例。我们实际上把这些打印了出来，放在办公室中央的一张桌子上。有时候做设计，你需要捕捉自己最本能、最直觉的反应；但有时候，你需要真正让一个设计方案慢慢沉淀、发酵。这次属于后者。我们让这些设计方案慢慢沉淀，让这些字体选择慢慢渗透。正如我所说，没有一个是真正出色的。
第一，我在找一款本身就足够漂亮的字体。第二，我在找一款能够传达任何类型的信息、却不会盖过信息本身情感的字体。比如，这封邮件是在邀请你参加派对——这个字体合适吗？很多字体，包括几乎所有的衬线字体，对于这种场景来说实际上太过庄重或过于严肃。或者换个极端，如果这封邮件是在通知某人的去世——这个字体合适吗？很多字体对于那种场景来说又太轻快了。比如你不会希望那种信息用 Comic Sans 来显示。第三，我在优化一款能让阅读速度和理解力都非常快的字体。第四，我在找一款能让邮件地址本身看起来很棒的字体。所以我把这 15 款全部淘汰了，因为它们都不够好。经过广泛搜索之后，我发现了一款叫 Adelle Sans 的字体，由一家叫 Type Together 的字体设计工作室设计，网站是 type-together.com。他们有很多精美的字体，值得去看看。
回到我的清单——第一，Adelle Sans 非常漂亮。我觉得每个字符都是一件艺术品，造型优美。第二，Adelle Sans 可以说是明快的、积极的，同时又足够严肃，能够传达任何类型的信息。它有恰到好处的个性，又不会个性过头。第三，Adelle Sans 还有一个特点，它的字形格外偏窄，这实际上特别适合邮件。我对 Gmail 的一个不满就是——它默认使用 Arial——文本行的宽度和你窗口一样宽。所以如果你在宽屏显示器上，行就会变得非常长。过宽过长的行会降低阅读速度。因为当你读完一行的末尾时，你的眼睛已经找不到下一行的起始位置了。而 Arial 本身的字符就比较宽，这进一步加剧了这个问题。
字体优化与邮件体验
所以在 Superhuman，如果你用过我们的产品就会知道，我们把行宽——也就是排版学上的行长——固定在阅读速度的最优值，根据字体的不同，大约在 90 到 120 个字符之间。而 Adelle Sans 相当窄，所以它实际上让我们在相当小的窗口里就能做到这一点，行也比较密。这样我们在较小的窗口里就能呈现大量信息，而不会导致行过长，从而优化了阅读速度和理解力。第四点，也是最后一点，Adelle Sans 对邮件地址中的 @ 符号做了非常独特的处理。它实际上把 A 的基线和 @ 的基线放在了与其余文本相同的基线上。
所以举例来说，如果你的名字里有 A——我的名字里有，Rahul at Vohra，三个 A，或者说两个 A 和一个 at——它们实际上都在同一条基线上。这是个小细节，但让邮件地址看起来极其自然。如果你把这个和用其他字体排列的邮件地址对比一下，那些字体会显得非常笨拙和别扭，因为 A 被挤来挤去，在我看来就是有点蠢。当然，Adelle Sans 也不是完美的。所以我们后来又和一位字体设计师合作，处理了一些具体的细节——有些字形会显得有点局促——我们今天使用的字体已经非常接近零售版的 Adelle Sans 了。
Lenny Rachitsky： 这是在产品上线之前还是已经上线之后的事？
Rahul Vohra： 当时大概有 10 到 15 个用户吧。
Lenny Rachitsky： 所以我觉得这相当反直觉且独特——在产品还没上线之前就对字体和字型如此执着。那时候可能还在想”这东西到底能不能成？有人会在乎吗？“我觉得这很大程度上说明了你做产品的方式。
Rahul Vohra： 哦对，那种念头从来没出现过。我们后面可能还会聊到，但”这东西会不会成不了？“这种想法是很危险的。我们不能开始那样想，因为一旦开始，你什么时候才能停下自我怀疑呢？
Lenny Rachitsky： 有意思。所以你是确信这一定会成功的——正因为我如此确信，所以我要把这些细节做到位。不过创始人也容易掉进另一个陷阱——花太多时间去完善一个根本行不通的东西。而且总有人建议”尽早发布、频繁发布”。你怎么看？你怎么找到那个平衡？你有什么建议？
发布时机取决于商业模式
Rahul Vohra： 上线前花多少时间，确实取决于市场、结构以及你商业模式的本质。比如说，假设你在一个全新领域做交易平台——想象一下巅峰时期的 Lyft 或 Uber。它有很强的网络效应，因为平台上的车越多，等待时间就越短，人们就会优先选择你的 app 而不是竞争对手的。那种情况下没有一分一秒可以浪费，那种情况下你大概连觉都不该睡。你要招揽最激进、最疯狂的拼命三郎。你要每周工作 120 个小时，因为每一个边际分钟确实都至关重要。在市场中的每一个边际分钟，持续复利增长，都会让你的下一年变得更好。
但这并不是所有初创公司的真实情况，Superhuman 这样的产品当然也不是如此。是的，更努力工作总是更好的，我们在 Superhuman 也确实极其拼命，但还没到值得发布一个不靠谱的产品的程度。我想起一位创始人的故事，他当时在 Y Combinator，给我讲了他的 demo day 经历。他用的是 Mailbox——有些人可能记得，这也是一家创业公司，Dropbox 后来以大约一亿美元的价格收购了它。它之所以出名，除了那次收购，还因为它是第一个普及”滑动归档”和”滑动标记”操作的产品——当然，现在这已经成了 Superhuman 和其他所有 app 的标配了。
这位创始人在 demo day 上用着 Mailbox，一切非常顺利。他在会场里四处交际，见投资人，推销他的摄影 app。那天晚上回到家，打开笔记本电脑，启动 Mailbox，发出了好几封跟进邮件。他等了一天，没收到回复；又等了两天，还是没有。第三天早上他意识到不对劲，于是打开 Gmail，查看已发送邮件——果然，里面一封都没有。Mailbox 出了问题。他一边骂自己一边安慰自己一切都会好起来的，然后用 Gmail 手动重新发送了所有邮件，这次都发出去了。
但后来有个投资人说，“对了，你可能想检查一下你的邮件客户端，因为我收到了一些你的重复邮件。“他回到 Gmail 一看，果然，原来在 Mailbox 里排队的那些邮件现在也确实发出去了，而不幸的是，他给大多数投资人发了两遍。好吧，这算世界末日吗？不算。投资人可以理解这种事，说不定你还因为尝试新 app 而加分了呢。但当时是不是惊恐万分、吓得要死？绝对是。
想象一下，如果面对的不是投资人而是客户——你正试图说服对方购买你的产品，让他相信你知道自己在做什么，你对细节有极致的关注，一切都在你的掌控之中、完美无缺。现在你丢脸了，你看起来像个傻瓜。这就是为什么当你有像邮件这样的关键任务型产品，你需要直接面对客户、候选人、投资人时，这些细节就真的很重要。邮件是关键任务型的，它不是你可以随便发布一个半成品就了事的东西。
工艺与速度的权衡
Lenny Rachitsky： 这是一个非常重要的、非常细腻的观点。一直都有这样的争论——应该花多少精力在工艺和用户体验上，花多少精力在发布速度和快速上线上。我在这里听到的，也是我完全同意的，就是：这取决于你所在的市场和你产品的关键程度。如果是邮件，它就是必须正常工作，你必须把这一点做到位，必须花所有需要的时间把它做对。
这让我想起了另一件事——你的早期投资人曾跟我分享过，来自 First Round Capital 的 Bill Trenchard。他谈到速度就是你把杠杆推到顶的东西。你说”我们要把这个作为核心焦点。速度、速度、速度。“我想也许这里的教训是，你选择那个你认为最能让你与众不同、让你远超现有方案的东西。那么你是如何决定速度就是你该执着的东西的？对于那些正在决定该把什么推到顶的人，你有什么建议？
速度与定位
Rahul Vohra： Bill 说得对，我同意他的看法——你必须选择一个东西。知道该选什么才是关键。在 Superhuman 早期，我读了一本关于定位的书，对我的思维方式影响很大。我记得书名叫《定位：争夺用户心智的战争》（Positioning: The Battle for Your Mind）。它让我印象深刻的是，最知名的品牌都代表着一个清晰的东西，它们拥有一个清晰的定位。所以为了让 Superhuman 令人难忘，我认为我们需要占据一个清晰的定位——独特的、未被占据的、并能强化我们产品战略的定位。
因此，在 Superhuman 的第一年，我采访了数百位潜在客户，了解他们使用 Gmail 和 Outlook 的体验。不出所料，几乎所有人都说电子邮件占用了太多时间。但有趣的是，也有很多人说 Gmail 和 Outlook 太慢了。正是从那时起，我第一次想到速度可能是一个有意思的定位。然后我问自己，“速度这个定位是否独特，是否可用？“答案是非常肯定的，因为几乎没有软件是以速度这个价值主张来销售或曾经销售过的。我能记得的上一次有人尝试这样做，是 Google 推出 Chrome 的时候，显然那次对他们来说取得了巨大的成功。你可能还记得他们拍过慢动作视频，对比 Chrome 渲染网页的速度，展示它比真正的闪电还快。从那以后，再也没有人做过这件事。
Rahul Vohra： 然后我又问，“速度是否能强化我们的产品战略？“答案同样是非常肯定的。我知道我们的竞争对手不会是创业公司，而是现有巨头。我也知道巨头通常在速度方面天生处于劣势，因为按定义它们拥有庞大的规模，通常还有根深蒂固的架构。最后，我做了一个我称之为鸡尾酒会测试的事情——就是去观察鸡尾酒会上的场景，看人们如何把你的产品推介给其他人。在我们的案例中，推介非常简单。人们会说，“兄弟，你必须试试，它真的他妈的快。“就这样。这就是他们的推介。我就是从这些地方知道速度对我们来说会是一个非常棒的起始定位。
反常规做法：手动引导每位新用户
Lenny Rachitsky： 接下来我想花时间聊聊的领域，我预计会有很多洞见——就是你在构建 Superhuman 的过程中，一些反常规的做法，很多公司从未想过要这样做，而你做了，而且奏效了。第一个就是手动引导每一位新用户。当然，创业公司也做过这种事，创始人拉来一些人，展示给他们看，然后就不再做了，接着就变成自助服务或交给销售团队。你把这个手动引导阶段在公司里做到了多大规模？有多少人在做引导工作？你手动引导了多少人？
Rahul Vohra： 对于不了解情况的人来说，在早期我们坚持一对一的专人引导，这绝对是正确的决定。你必须经过引导体验才能使用 Superhuman。现在几乎反过来了——几乎所有新的 Superhuman 客户都走自助服务流程。引导体验仍然存在，但同样，这也是绝对正确的做法。回答你的问题，在最高峰时我们大约有 20 个人做手动引导。
Lenny Rachitsky： 所以人并不多。这很有意思。因为我一直以为是一个庞大的团队，但 20 个人就能处理很多事，这是关键结论。在什么规模时你停止了手动引导？给那些正在考虑这样做的人一些参考，什么时候该停下来？
Rahul Vohra： 我认为停下来的原因是，总会有某些性格类型的人不想经历一对一引导。到了某个阶段，这些人会变得非常重要，你需要准备一个世界级的自助服务选项。当我们开始构建自助服务时，感觉几乎不可能做到。事实上，这让人非常恐惧，因为很难高估整个公司的 DNA 有多少是围绕”手动引导用户”这个理念建立起来的。毕竟，我们在一对一引导中做了很多事情，而软件能做的终究有限。后来，经过大量的艰苦努力和坚持，我们最终找到了解决方案，如今我们拥有了一个世界级的自助服务体验，但当时并没有。
所以另一面的问题是，为什么一开始要这样做？我们发现了两个好处。第一，用户指标非常出色——无论是参与度、留存率、产品市场契合度评分、MPS、病毒式传播，所有这些指标，如果你愿意花精力一对一引导早期客户，我认为你会显著超越行业基准。当涉及到品牌建设时，拥有一批早期的超级粉丝会变得非常强大。如果大家还记得之前最上面的那个讨论，什么能创造真正的病毒式传播？不是病毒式传播机制，而是口碑。是品牌。这就是你启动品牌的方式。
第二，在一个你可以轻松快速融资的世界里——比如零利率现象那个时期——你实际上可以用资金来避免构建首次用户体验以及所有那些常规的增长循环。你可以把所有工程师的精力集中在一起去寻找产品市场契合度，或者做方案深化，或者做市场拓展，而不是例如去构建首次用户体验，不是例如去做激活，因为有人类在替你做激活。相比之下，我看到其他竞争对手公司往往把将近一半的工程投入花在这些事情上——花在自助服务流程上——而那些产品最终都没有找到产品市场契合度。所以，如果你真的想打造品牌——我认为所有偏消费类的公司都需要这样做——这件事就值得做。如果钱像从树上掉下来一样（不管什么原因），而我们确实有过那样一段时期，可以说现在的 AI 公司又重新拥有了这样的条件。所以如果你能把它融入你的战略，我认为你应该这么做，但你也应该知道什么时候该停下来。
Lenny Rachitsky： 非常有趣。我觉得有几个因素值得考虑，因为我本来就想问你，人们什么时候应该考虑这样做？如果有人听到这里觉得，“这太棒了，只要安排专人引导每个新用户，所有人都能被激活，太完美了。“那么你提到的一些变量包括——你是否有充足的低成本资金来投入，不一定是 20 个人，可以先从几个人开始。另外还有 LTV 和 ACV 的考量，即你从每个新客户身上能赚到足够的钱吗？我想这也是一个变量。你觉得创始人还应该考虑哪些其他因素？
Rahul Vohra： 当然。你不能在这件事上亏钱。需要说明的是，我们做引导一直都是赚钱的，只是在某个阶段，大众市场——对我们来说是企业客户或者全球所有的专业消费者——你的漏斗顶部需要足够宽，宽到手动引导不再有意义。
产品市场契合度
Lenny Rachitsky： 太好了。好，我们来聊聊产品市场契合度。我知道大家一提到 Rahul 就会想到产品市场契合度。你为 First Round 写了那篇重磅文章，描述了你们团队如何approach产品市场契合度。我们不打算花太多时间描述它的具体内容，大家可以自己去查阅。那我直接问你，关于寻找产品市场契合度、达成产品市场契合度，你觉得人们还有哪些没有理解的地方？考虑到这是创始人需要搞清楚的最重要的事情。如果你没有找到人们想要的东西，其他一切都无从谈起。有什么想分享的吗？
Rahul Vohra： 核心想法至今仍然足够反直觉，所以我就从这里讲起。第一，你可以量化产品市场契合度。第二，你可以优化产品市场契合度。第三，你可以系统性地，甚至用数字化的方式提升产品市场契合度。第四，你甚至可以让一个算法来为你撰写路线图，而且这个路线图是保证能提升产品市场契合度的。现在，如果这听起来很疯狂，我会第一个承认，这听起来不应该是对的，但去看看那篇文章吧。我认为它仍然是 First Round Review 上被分享最多的文章，标题是《How Superhuman Built an Engine to Find Product Market Fit》，或者直接 Google 搜索 Superhuman Product Market Fit Engine，你会看到完整的算法详解以及它为什么有效。
第二点我想说的是，要达到产品市场契合度，你必须刻意不去采纳许多早期用户的反馈。与此同时，你又需要认真倾听用户，构建他们想要的东西。这正是我们要做的——做出人们想要的产品。但不能是所有人。不能是每个人。问题变成了，你怎么听？以及即使听到了他们说的内容，你要关注什么、忽略什么？所有这些都在产品市场契合度引擎中有详细阐述。
用算法构建路线图
Lenny Rachitsky： 好的，我得顺着这个话题深入一下——用算法构建路线图来提升产品市场契合度。说说具体怎么做。
Rahul Vohra： 这其实就是引擎的核心部分。让我试试看能不能用一种非常容易理解的方式在这里浓缩讲解。假设——为了论证方便——你可以给产品市场契合度一个数值，而事实证明确实可以。方法很简单，你去问用户，“如果你不能再使用这个产品，你会有什么感受？“你给他们三个选项：非常失望、有些失望、不失望。非常失望的意思是，“我会崩溃的。我爱这个产品。我需要这个产品。”
Sean Ellis 的发现是这样的——Sean Ellis，如果你不认识他，就是那位创造了”增长黑客”这个词的人，他对这个初始问题进行了实施和基准化测试。他发现，增长困难的公司几乎总是”非常失望”的比例低于 40%，而增长最快的公司几乎总是”非常失望”的比例高于 40%。而且这个问题、这个指标比例如净推荐值之类的指标更能预测成功。
好，到目前为止还算简单。那怎么让这个数字上升呢？你希望更多人在没有你的产品时感到非常失望。这里的诀窍是，不要过多地根据”非常失望”的人给出的反馈来行动，因为他们已经爱你的产品了。同时也不要去理会”不失望”的人给出的反馈，因为他们离爱上你的产品太远了，基本上是没希望的。你要聚焦的是”有些失望”这个群体——他们有点喜欢你的产品，但有什么东西——我敢说是某个小问题——在阻碍他们。
然后你把他们分成两拨：一拨是对你产品的主要利益点产生共鸣的，另一拨是没有的。这是什么意思呢？你回到那些真正热爱你产品的人，基本上问他们为什么？我的产品有什么是你真正喜欢的？在 Superhuman 的早期，答案会是速度和键盘快捷键，以及整体的设计美学，还有我们为你节省的时间。然后你回到那些”有些失望”的用户，以 Superhuman 为例，我会直接问，“等等，你喜欢 Superhuman 是因为它的速度，还是因为别的什么？“如果是别的什么——这很难做到，但要礼貌地忽略这些人和他们的反馈。因为即使你按照他们的要求构建了所有功能，他们仍然在把你拉向一个不同的方向。而他们最喜欢你产品的地方，实际上并不是大多数人最喜欢你产品的地方。
这样你就明确了值得关注的子群体中的子群体，引擎中还有另一个问题来找出他们不喜欢产品的哪些方面。现在你有了一个人们喜欢的东西的清单，一个人们不太喜欢的东西的清单，你可以逐一攻克这个清单来让产品市场契合度分数上升。基本上在每个规划周期的开始，我建议把一半时间用来加倍投入人们真正喜爱的部分，另一半时间用来系统性地消除”有些失望”用户——具体来说是那些对主要利益点产生共鸣的用户——的反对意见。
Lenny Rachitsky： 这是一个非常精彩的总结。我知道我说过我们不会在这里花太多时间，但我很高兴我们这么做了。真的很有帮助。我再问你一个问题——我知道你最初在早期使用了这个方法，你现在还在以某种形式这样运作吗？
Rahul Vohra： 对于 Superhuman 整体，我们不再按原样运行这个引擎了。Superhuman 现在有足够多的子组件，它们几乎各自都是独立的产品。比如 Superhuman for Sales、我们的多人协作功能、我们如何看待企业市场、AI 也是一个独立的领域，但我们确实会在那些独立的部分上运行引擎。例如，我们会向销售人员提问产品市场契合度引擎的问题，专门针对 Superhuman for Sales。当我们考虑启动新产品时，我们绝对会部署产品市场契合度引擎。
Lenny Rachitsky： 太好了。你问这个问题的形式是产品内的插页式调查弹窗吗？
Rahul Vohra： 你可以用任何方式来做。Sean 最初基准化这个数值时是通过邮件调查。我觉得邮件调查完全没问题。关键在于——这适用于任何调查方法——如果你要改变调查的方式，你之前所有的数据就全部失效了。那只是一个新的基准线，从那之后重新开始。
Lenny Rachitsky： 明白了。我们之前请 Sean 上过播客，他详细描述了这个方法。如果大家想深入了解 Sean Ellis 测试，可以去听那一期。我们会在节目说明中附上链接。
游戏设计 vs 游戏化
接下来这个话题我很期待听听你的看法——游戏设计与游戏化的区别。这是你在产品设计方面非常独特的思维方式之一。人们听到你谈论这个时，往往会想：“哦，游戏化，就是把东西做得像游戏一样。哦，那是 Zynga、是 Farmville，我可不想那样做。“但你对为什么在产品设计时需要考虑游戏设计，有着截然不同的视角。来谈谈你的见解吧。
Rahul Vohra： 我坚信我们应该像做游戏一样来做商业软件。因为当我们像做游戏一样做产品时，人们会觉得它有趣。他们会告诉朋友，会爱上它。这其实从另一个角度回到了我们这次对话开头的主题——你在打造品牌，你在为口碑传播创造理由。这是一种完全不同的产品开发方式。
那我们怎么做呢？正如你所说，不是游戏化，游戏化行不通。游戏设计是有效的，但游戏设计不等于游戏化。它不是简单地在你的产品上加点积分、等级、奖杯或徽章。
要理解为什么游戏化行不通，我们必须从人类动机说起。斯坦福大学有一项非常有趣的研究，完美地展示了其中的差异。上世纪七十年代，斯坦福的研究人员招募了一批三到四岁的儿童，这些孩子普遍对画画有天然的兴趣。其中一部分孩子被告知会获得奖励——一张带有金色印章和丝带的证书。另一部分孩子则没有被提及任何奖励，他们甚至不期待也不知道有奖励。然后每个孩子被单独邀请到一个房间里画六分钟画，之后有些孩子会得到奖励，有些则不会。
在接下来的几天里，研究人员观察这些孩子自主画画的时间。没有奖励的孩子，花了 17% 的时间在画画。而那些期待奖励的孩子，遗憾的是，他们只花了 8% 的时间画画。奖励的存在直接把他们的动机减半了。这是怎么回事？研究人员区分了内在动机和外在动机。内在动机驱使我们去做那些本身就有趣、令人满足的事情，而外在动机则驱使我们去获取奖励、达成外部目标。奖励的问题就在于此——它们会严重削弱内在动机。这就是游戏化行不通的原因。而当游戏化确实有效时，那是因为底层体验本身已经被设计得像一个游戏了。
Lenny Rachitsky： 那什么让一个东西像游戏呢？我知道 Superhuman 在这方面做得很好，比如你在收件箱中追求收件箱清零的体验。为了让这个概念更具体一些，你觉得游戏设计到底是什么？对你来说意味着什么？是什么让它感觉像一个游戏？
Rahul Vohra： 也许大家不知道，在成为创始人之前——你们可能也看得出来——我实际上是一名职业游戏设计师。事实证明，游戏设计并没有一个统一的理论。要创造游戏，我们需要借鉴心理学、数学、叙事和交互设计的艺术与科学。在 Superhuman，我们梳理出了五个我们真正关注的核心领域：目标、情感、玩具、控制和心流。在这五个维度上，我们总结了许多游戏设计原则。举一个例子：先做出有趣的玩具，再把它们组合成游戏。
我喜欢问一个问题：玩具和游戏是一样的吗？它们看起来确实不同。比如，我们玩玩具，但我们玩游戏。球是一个玩具，但足球是一个游戏。事实证明，最好的游戏都是由玩具构建而成的。为什么？因为这样在两个层面上都很有趣——玩具本身的乐趣和游戏本身的乐趣。例如在 Superhuman 中，我们最喜欢的玩具之一就是时间自动补全器。如果你使用 Superhuman，这就是你按下 H 键时出现的东西——当你对邮件设置稍后提醒或 snooze 时使用的功能。你可以输入任何你想输入的内容，哪怕是乱码，它都会尽力理解你。比如输入 2D，它会变成两天；3H 变成三小时；输入 one MO 就变成一个月。时间自动补全器之所以有趣，是因为它迎合了你那种充满好奇的探索欲。
在引导（onboarding）过程中，我很快就开始看到有人问：“它能做什么？它在哪里会出错？它是怎么工作的？如果我持续输入一系列十会怎样？“结果是，那变成了十月十日晚上十点十分。那如果输入一系列二呢？那是二〇二二年二月二日下午两点。然后你开始尝试更复杂的输入，比如 in a fortnight and a day，这也能工作，而且会带来一种愉悦的惊喜。很快你还会发现更多令人愉快的惊喜——时区计算会自动完成，你根本不用费心去想。你只需输入 8:00 AM in Tokyo，它就自动换算成东部时间晚上八点，你再也不用自己算时区了。
大多数人后来非常高兴地发现，如果你真的想的话，你可以把邮件 snooze 到”永不”——也就是说，你真的可以输入 never，那封邮件就再也不会回来了。它还会附带一个小耸肩的表情符号。这个玩具能获奖吗？不能。但它有趣吗？出乎意料地，是的。所以我想鼓励大家去审视自己产品的功能：这些功能是否能引发充满好奇的探索？它们即使没有目标也依然有趣吗？它们是否能带来愉悦的惊喜时刻？如果是的话，你就拥有了一个玩具，然后你可以把它和其他玩具组合起来，真正开始构建一个游戏。
Lenny Rachitsky： 如果有人只听播客的这一段，绝不会猜到我们在谈论的是 B2B 软件和电子邮件，我很喜欢这一点。
定价策略
来聊聊定价策略以及你的定价方法吧。这也是你们采取的另一种非常反常规的做法——你们对电子邮件收取每月 30 美元的费用，而电子邮件在其他地方是免费的，人们根本不需要花钱。但这个策略奏效了，现在很多公司也在这样思考。你们最近甚至还提价了。关于定价策略，你有哪些心得可以分享给大家？
Rahul Vohra： 关于定价，我总是说同样的话：在你搞清楚定价之前，必须先搞清楚定位。Superhuman 是市场上最好的电子邮件工具。我们很幸运有数据可以证明这一点。销售电子邮件工具的一个好处是，你可以对比使用 Superhuman 之前三十天和使用之后三十天的情况，或者使用前一年和使用后一年的情况。我们当然做了这种对比。我们能够证明，人们使用 Superhuman 处理邮件的速度提高了一倍，回复速度加快了一到两天，每周节省四小时甚至更多。正因如此，我们非常有信心地说，Superhuman 是市场上最好的电子邮件工具，我们是为高绩效团队和高绩效个人打造的。换句话说，我们服务的是市场的高端部分。
一旦你搞清楚了自己的定位，就可以着手定价了。这方面最好的书之一是 Madhavan Ramanujan 写的《Monetizing Innovation》。Madhavan 在书中介绍了许多定价方法。我们用的是其中最简单的一种——Van Westendorp 价格敏感度测试。在早期，我们向大约一百位最早的用户问了以下四个问题。第一，你认为 Superhuman 的价格高到什么程度时，你会完全不考虑购买？第二，你认为 Superhuman 的价格低到什么程度时，你会担心它的质量而不敢购买？第三，你认为 Superhuman 的价格到什么程度时开始觉得偏贵——并非完全不能接受，但需要认真考虑一下才会购买？第四，你认为 Superhuman 在什么价格时算是非常划算——物超所值？
大多数初创公司会选择第四个价格点来定价。对于全新领域的市场机会和交易平台尤其如此，你必须把交易价格定在第四个点附近——基本上就是在漏斗顶端尽可能多地让人注册。但能够支撑我们”同类最佳、品类最优”定位的价格点，其实是第三个——让人觉得开始偏贵，但坐下来认真算一算自己花在邮件上的时间和投资回报率之后，还是会购买。第三个问题的中位数答案是每月 30 美元，这就是我们定价的由来。
定好价格之后，我们快速做了一个市场规模的压力测试。比方说，我们是一家风险投资级别的公司，但当时我们必须问自己的问题是：“我们能成长到十亿美元估值吗？“假设估值是 ARR 的十倍，那 ARR 就需要达到一亿美元。以每月 30 美元计算，那就是 30 万订阅用户——这还是保守估计，完全没有考虑其他提升 ARPU 的方式。你提到涨价，还可以往高端市场走，还可以销售新产品等等。我们问自己，不靠这些手段，我们有没有信心做到几十万订阅用户？我们的回答是肯定的，于是我们就按这个价格推出了。
AI 功能
Lenny Rachitsky： 好的，在我们剩下的时间里我还想聊几个话题，然后我知道你得走了。第一个是关于 AI，你们在这方面做的工作。我知道那是一个很大的突破。第二个是你们在企业市场的布局。如果还有时间，我还想问一个关于你们运营方式的有趣问题。
先聊 AI 吧。感觉你们像是占尽天时地利——你们在这方面已经构建了很长时间，然后 AI 又给你们在电子邮件领域能做的事情打开了一个新阶段。能聊聊你们做了什么、怎么看待 AI 与现有产品的整合、它是如何让你们再次腾飞的吗？
Rahul Vohra： 确实，有时候初创公司归根结底就是天时地利。我们最近刚进行了一次大规模的 AI 发布，大约两周前，但在此之前我们已经有了多个旗舰级 AI 功能。我们的第一个 AI 功能是”用 AI 写”（write with AI），写下几个关键词，我们就会把它变成一封完整的邮件。而且我们会匹配你已发送邮件的语气和口吻。所以与 Copilot、Gemini 以及几乎所有其他邮件应用不同，写出来的邮件听起来就像你本人。这个 AI 功能比我预想的受欢迎得多。现在平均每位用户每周使用它 37 次。
第二个 AI 功能是自动摘要（auto summarize），在每封对话上方显示一行摘要。随着新邮件到达，摘要会即时更新。同样，与 Copilot 和 Gemini 不同，它是预先计算好的。我们做的很多事情都是不遗余力地让这些功能真正高端、体验出色。之后的下一个 AI 功能是即时回复（instant reply）。想象一下，早上醒来收件箱里每封邮件都已经有了一份回复草稿——你只需要编辑然后发送，有时候甚至不需要编辑。我可以分享一下，我们刚完成了一项分析：在整个 2024 年，通过 Superhuman 发送的 AI 撰写邮件占比在一年内增长了四倍。
如果我没记错的话，之后的功能是”向 AI 提问”（Ask AI）。电子邮件当然是一个关键信息的宝库——项目状态、客户沟通、会议动态、交易进展等等。四十多年来，我们一直依赖那个可笑地被称为”搜索”的功能——你得记住发件人，猜测关键词，扫描主题行。现在你可以直接问：“Q1 线下会议在哪？“或者”我的航班详情是什么？“又或者”对 Ask AI 上线最积极的五条客户反馈是什么？“顺便说一句，这个任务以前需要我花二三十分钟阅读所有邮件然后生成报告，现在不到五秒就能完成。
最近我们宣布了有史以来最大的进化。Superhuman AI 持续为你提供帮助——它整理你的收件箱，同时确保你绝不会遗漏任何事情。我们有称之为”自动标签”（Auto Labels）的功能：你现在可以写一个简短的提示词，比如”求职申请”或”工作审阅请求”，然后当有邮件匹配该提示词时——有人申请职位或请你审阅工作——你能立刻看到。借助”自动提醒”（Auto Reminders），如果你的邮件需要回复，Superhuman 会自动设置提醒。你不用再记着这件事，再也不会遗漏。你只需要点击发送。借助”自动草稿”（Auto Drafts），Superhuman 会自动为你起草跟进邮件，很快就会为几乎所有需要回复的邮件起草回复。
最后，借助我们称之为”工作流”（Workflows）的功能，你现在可以将邮件变成可重复的自动化工作流。比如，我经常收到有兴趣在 Superhuman 工作的人发来的邮件，我通常会回复那位候选人，告知团队会查看。然后我会把原始邮件连同简历或求职信一起转发给我们的人力与运营负责人，请她在有兴趣时联系对方。有了工作流，我现在可以自动化这整个流程。你可以想象，它就像创建了一个小型流程图，规定了需要发生的一系列步骤。这不仅节省了大量时间，而且有了工作流，你甚至不需要待在收件箱里。事实上，你甚至不需要在工作。你可以正在度假，而 Superhuman AI 在为你工作。
Lenny Rachitsky： 这对我来说就是产品市场契合度。这一切听起来太棒了，完全合情合理。这就是我们一直被承诺的东西——水下城市、飞行汽车，然后就是能神奇运转、替我们回复的邮件以及所有这些功能。我喜欢你们在做的这一切。
对于正在用 AI 构建产品的人，我很好奇，在如此深度地依赖 AI 模型的过程中，最大的惊喜是什么——无论是好的还是坏的——你觉得有什么经验可以帮助大家注意或者关注的东西吗？
Rahul Vohra： 对我来说，最大的惊喜在于用户喜爱的东西如此难以预测——他们会喜欢什么，不喜欢什么。比如说，用 AI 写。这听起来像是一个同质化的功能，从表面上看也确实如此。每个邮件应用、每个写作界面都有用 AI 写的功能。我敢说我们的是邮件领域最好的——而令人惊讶的是，这正是我们的本职工作。但真正令人惊讶的是人们有多喜欢它，使用频率有多高。每用户每周 37 次，这个数字至今让我觉得难以置信。我完全没有预料到这一点，这是最让我意外的。
反过来，也有一些 AI 功能我原本预期会有大量使用，但我们并没有获得我们可能期望的那种使用量。希望我不是 AI 领域的 Kramer（总是判断反了的那个人），但基本上所有我认为会大受欢迎的功能，人们用得都比预期的少。而那些我心里想”不确定，但还是做出来吧”的功能，人们反而特别喜爱。
Lenny Rachitsky： 有意思。
Rahul Vohra： 也许我应该创造一个反我来做 AI 路线规划。
Lenny Rachitsky： 这就是一个简单的 agent——Rahul 说什么，做相反的事就对了。
Rahul Vohra： 是的。
从 PLG 转向企业销售
Lenny Rachitsky： 好，最后一个话题。我知道你们正在进军企业市场。人们想到 Superhuman 时，会觉得它是面向消费者的、给个人用户的产品，而你们正在做大量工作，把它打造成一个 B2B 企业级产品。对于那些开始考虑从 PLG 向销售主导的 B2B 企业模式转型的创始人来说，你学到了什么？走到那一步需要什么？你们的销售流程是怎样的？
Rahul Vohra： 在某些方面，这和向高级个人用户销售非常相似，只不过这些用户不是从 Gmail 来的（高级个人用户通常来自 Gmail），而是从 Outlook 来的。而 Outlook 用户与 Gmail 用户的期望截然不同。比如说，Outlook 用户期望他们的邮件应用同时也是一个功能完备的日历应用，而 Gmail 用户则完全不介意这两者是分开的。因此，我们在日历上投入了大量精力，而且仍在继续加大投入。我能说的有限，但确实非常令人兴奋。
Lenny Rachitsky： [听不清]。
Rahul Vohra： Outlook 用户还习惯于一些安全防护措施，比如如果你在企业中使用过 Outlook，当收件人不在你的域名内时会有警告提示，或者 Outlook 用户可能知道的敏感度标签。因此我们构建了对外部收件人指示器和敏感度标签的支持。但在某些方面，向企业销售又与向高级个人用户销售截然不同，因为涉及到其他利益相关方。比如说，我们通过实现 Microsoft Intune 来支持企业移动设备管理。
我们最近签约了三大战略咨询公司之一，这非常令人兴奋。我不能说是哪家，但他们非常喜欢 Superhuman，内部有数千人在使用。这是在经过一年的试点之后——他们试点了一整年，然后在最近几个月开始加速推广。信不信由你，我们才刚刚给他们提供移动应用。因为在那样规模的企业中，对合规移动应用能做什么和不能做什么有严格的管控要求。比如说，IT 部门需要能够控制哪些应用可以保存附件，或者你可以从邮件中向哪些应用复制粘贴文本。对于许多企业来说，这些管控措施至关重要。
Lenny Rachitsky： 哇，好的。所以听起来你们现在走的路线基本上就是构建大公司需要的所有这些功能。
Rahul Vohra： 没错。这里有两类利益相关方。一类是用户，他们其实与之前的用户有很大不同，因为他们是 Outlook 用户，日历需求是这种差异的主要体现之一。还有一整套其他利益相关方，IT 部门是其中之一，但还有其他的。比如说，这种规模的公司有职场管理团队，他们希望看到关于员工工作方式的分析数据，了解如何让团队更高效，所以这确实是一个涉及多个利益相关方的多线销售过程。
Lenny Rachitsky： 之前有 Linear 的人来播客做客，他——我不知道你有没有听那期——他谈到他们如何决定优先做什么，他们永远不会做的东西之一就是让中层管理者追踪下属表现之类的功能。这对你来说也许是一个有趣的取舍机会，可以砍掉一些东西。我不知道。
单一决定性理由（SDR）
总之，我想用最后一个精华来收尾。好，很高兴我们还有时间聊这个。你分享过 Superhuman 内部有一个做决策的系统，你称之为 Single Decisive Reason，SDR（单一决定性理由）。这是什么？
Rahul Vohra： SDR 是我在 LinkedIn 期间从 Reid Hoffman 那里学到的一个思维工具。其核心思想是，对于重要的决策，你应该能够找出一个——仅仅一个——独自就能支撑该决策的理由。它的依据是一个观察：我们太经常依赖一堆薄弱理由的集合来为决策辩护了。这种做法非常、非常容易陷入。想象一下，你正在斟酌一个决策，列出利弊清单。有 3 个优点，但假设有 10 或 15 个缺点。缺点的 sheer 数量、逐一思考它们所花的精力、把它们写下来所花的时间——这一切都会影响你，有意识地或更糟地、无意识地。在群体环境中尤其如此——群体天然就更倾向于规避风险、追求共识。
所以每当有人要做决策，把决策拿到我面前说”我们想做这件事，因为 X、Y、Z，有多个理由”时，我会问他们：“SDR 是什么？单一决定性理由是什么？“如果他们还无法把它提炼出来，这就告诉我他们还没有真正想清楚为什么要做这个决策。这并不意味着决策是错的，只是意味着他们还没有找到我们应该做这件事的那个唯一理由。然后他们可以逐一审视自己的理由清单，问自己：“仅凭这一条理由，就足以支撑这个决策吗？“也就是说，如果只有这一条成立，其他所有理由都不成立，我还会做这件事吗？有时候答案还是会的，但实际上有时候不会。我们会意识到，一堆薄弱理由的集合本身可能意味着——比如说——结果的可能性比我们以为的要低，或者它掩盖了决策另一侧一个真正强有力的理由。
Lenny Rachitsky： 这个太棒了。就是当有人拿着一个决策来找你时，你使用这个概念的方式是——你问他们单一决定性理由是什么？
Rahul Vohra： 差不多是这样。是的。当然他们做不到——这也是人之常情，人们很自然地会先列出三四个理由，这没问题。然后我会说：“好，但如果其中只有一条成立，而你仍然在为这个决策辩护，那会是哪一条？“我认为这就是一个好决策的门槛。
Lenny Rachitsky： 为什么这一点如此重要？因为你发现一堆低质量的理由加在一起也不等于一个好理由？
Rahul Vohra： 多个理由，这很讽刺。但这就是我对 SDR 之所以有效的 SDR——是的，多个低质量理由很少能叠加成一个做某事的高质量理由。但还有其他方面的考量：你做出的任何决策都有机会成本。你构建的任何一个功能，都是另一个你没有构建的功能。如果我们要为了一堆薄弱的理由去构建这个，而可以为一个强有力的理由去构建那个，我宁愿为那一个强有力的理由去构建。当然，这是在其他条件相同的前提下，而这些事情往往相当复杂，但 SDR 可以一层层地向下应用。你刚才就对我做了这件事——我对 SDR 的 SDR 是什么？
Lenny Rachitsky： 这不就来了嘛。Rahul，我们还有什么没聊到、你想聊的吗？在我们结束之前，有没有什么最后的智慧想留给听众的？
Rahul Vohra： 我感觉很好。我觉得我们聊了很多。谢谢你问了非常棒的问题。这次真的很愉快。
Lenny Rachitsky： 非常精彩。那好，我来问你——大家在网上哪里可以找到你？在哪里可以体验 Superhuman？在试用之前有什么需要了解的？还有，听众怎样能帮到你？
Rahul Vohra： 如果你想在网上找到我，我一般在 X 上。地址是 x.com/rahulvohra，R-A-H-U-L V-O-H-R-A。我的私信是开放的，随时可以找我。不过如果你私信我，我建议同时也给我发封邮件，rahul@superhuman.com，希望我能尽快看到你的消息。
如果你还没有试过 Superhuman，天哪，你在等什么？这是我给你的号召——去试试吧，因为你的时间比你以为的更值钱。去下载 Superhuman 试试看。邀请你的团队。那些数据是真实的。我知道它们听起来像是创业公司编出来的那种指标——但用一半的时间处理完邮件、提前一到两天回复、每周节省四小时甚至更多——这些都是真的。
说到这个，我之前提到的那家咨询公司，因为他们非常重视数据和分析，他们想自己验证这些数字，于是他们真的做了。他们在内部做了一个关于 Superhuman 的案例研究，结论是：“是的，Superhuman 为我们的合伙人每人每周节省了 3.3 小时。我们买过的工具里只有另外一个能做到这一点，那就是 ChatGPT。谢谢你们，我们很喜欢 Superhuman，正在全面推广。” 如果这对你或你的公司有吸引力，请一定要试试看。
Lenny Rachitsky： 太酷了。回想我最初设想这次对话会是什么样子，我当时就觉得会有很多逆向思维和对细节的极致关注——确实如此。Rahul，你太棒了。非常感谢你来做客。
Rahul Vohra： 谢谢。大家再见。
Lenny Rachitsky： 大家再见。
非常感谢你的收听。如果你觉得这期节目有价值，可以在 Apple Podcasts、Spotify 或你喜欢的播客应用上订阅。也请考虑给我们评分或留下评论，这真的能帮助其他听众找到这档播客。你可以在 lennyspodcast.com 找到往期所有节目或了解更多关于本节目的信息。下期再见。
术语表
| 原文 | 中文 |
|---|---|
| activation | 激活 |
| Adelle Sans | Adelle Sans（字体名称，保留原文） |
| Arial | Arial（字体名称，保留原文） |
| ARPU | ARPU（每用户平均收入，保留原文） |
| ARR | ARR（年度经常性收入，保留原文） |
| Ask AI | 向 AI 提问（Ask AI） |
| Auto Drafts | 自动草稿（Auto Drafts） |
| Auto Labels | 自动标签（Auto Labels） |
| Auto Reminders | 自动提醒（Auto Reminders） |
| auto summarize | 自动摘要（auto summarize） |
| Bill Trenchard | Bill Trenchard（人名，保留原文） |
| Boldstart Ventures | Boldstart Ventures（风投机构名称，保留原文） |
| Chrome | Chrome（Google 浏览器产品名称，保留原文） |
| Comic Sans | Comic Sans（字体名称，保留原文） |
| Comrade | Comrade（人名，保留原文） |
| concierge onboarding | 专人引导（concierge onboarding） |
| Copilot | Copilot（微软 AI 助手产品名称，保留原文） |
| demo day | demo day（创业孵化器展示日，保留原文） |
| Dropbox | Dropbox（公司名称，保留原文） |
| EA (Executive Assistant) | EA（Executive Assistant，行政助理） |
| Ed Sim | Ed Sim（人名，保留原文） |
| Elliot Durbin | Elliot Durbin（人名，保留原文） |
| enterprise mobile management | 企业移动设备管理（enterprise mobile management） |
| external recipient indicators | 外部收件人指示器（external recipient indicators） |
| first principles | 第一性原理 |
| First Round Capital | First Round Capital（风投机构名称，保留原文） |
| first-time user experience | 首次用户体验 |
| Gemini | Gemini（Google AI 产品名称，保留原文） |
| Google（公司名称，保留原文） | |
| Growth | 增长（指公司增长团队/职能） |
| growth hacker | 增长黑客 |
| growth loops | 增长循环 |
| instant reply | 即时回复（instant reply） |
| Kramer | Kramer（人名/角色名，指电视剧《宋飞正传》中总做错误判断的角色，保留原文） |
| Laurent Valasek | Laurent Valasek（保留原文） |
| Lyft | Lyft（公司名称，保留原文） |
| Madhavan Ramanujam | Madhavan Ramanujam（人名，保留原文） |
| Mailbox | Mailbox（公司名称，保留原文） |
| market widening | 市场拓展 |
| Microsoft Intune | Microsoft Intune（企业移动设备管理平台，保留原文） |
| Monetizing Innovation | 《Monetizing Innovation》（书名，保留原文） |
| MPS | MPS（用户满意度指标，保留原文） |
| net promoter score | 净推荐值（net promoter score） |
| Outlook | Outlook（微软邮件客户端产品名称，保留原文） |
| Peak Leadership Institute | Peak Leadership Institute（保留原文） |
| PLG | PLG（Product-Led Growth，产品驱动增长，保留原文） |
| Positioning: The Battle for Your Mind | 《定位：争夺用户心智的战争》（Positioning: The Battle for Your Mind，营销学经典著作） |
| Product Market Fit Engine | 产品市场契合度引擎（Product Market Fit Engine） |
| Product-Market Fit | 产品市场契合度 |
| prosumer | 高级个人用户（prosumer） |
| Raja | 王瑜伽（Raja）传统 |
| Rapportive | Rapportive（公司名称，保留原文） |
| Reid Hoffman | Reid Hoffman（人名，保留原文） |
| SDR (Single Decisive Reason) | SDR（Single Decisive Reason，单一决定性理由） |
| Sean Ellis | Sean Ellis（人名，创造了”增长黑客”一词） |
| sensitivity labels | 敏感度标签（sensitivity labels） |
| solution deepening | 方案深化 |
| Superhuman for Sales | Superhuman for Sales（产品名称，保留原文） |
| Switch Lock | （Rahul 自创的技巧名称，保留原文） |
| Switch Log | Switch Log（Rahul 自创的时间追踪技巧名称，保留原文） |
| technology moat | 技术护城河 |
| transcendental meditation | 超觉冥想 |
| Type Together | Type Together（字体设计工作室名称，保留原文） |
| Uber | Uber（公司名称，保留原文） |
| Van Westendorp Price Sensitivity | Van Westendorp 价格敏感度测试 |
| virality | 病毒式传播 |
| Vivek | Vivek（人名，保留原文） |
| Workflows | 工作流（Workflows） |
| workplace management groups | 职场管理团队（workplace management groups） |
| write with AI | 用 AI 写（write with AI） |
| Y Combinator | Y Combinator（机构名称，保留原文） |
| Yogic | 瑜伽（Yogic）传统 |
| zero interest rate phenomenon | 零利率现象 |
| zone of genius | 天才区 |
此文档由 AI 分片翻译（translate_long_document）
Superhumans secret to success | Rahul Vohra (CEO and founder)
A Quick Preview
Lenny Rachitsky: Let’s talk about product market fit.
Rahul Vohra: You have to deliberately not act on the feedback of many of your early users, and this is at the same time as listening to people intensely and building what people want. That’s what we’re here to do, is to make something that people want, but it can’t be all people. And the question becomes, how do you listen to them? And then even of what they say, what do you pay attention to and what don’t you? The trick here is-
Introducing the Guest
Lenny Rachitsky: You’re not doing what a lot of CEOs think they need to be doing with their time. A lot of CEOs think they need to spend time on hiring or org building and you intentionally, “I will spend time on product and marketing design.”
Rahul Vohra: This is a technique that I call the switch lock. It’s born out of the observation that your calendar says what you thought you were going to do, but it’s really only your trail of work that describes what you actually did. How can we capture that? So I came up with the following idea. What if I just did whatever the heck I wanted?
Rapportive and LinkedIn Acquisition
Lenny Rachitsky: What’s the most pivotal moment in your career, in your life?
The Truth About Viral Growth
Rahul Vohra: I learned the real secret behind virality. There is no such thing as a truly viral product. What then is the true secret? It is-
Lenny Rachitsky: Today my guest is Rahul Vohra. Rahul is the founder and CEO of Superhuman and one of the most thoughtful and insightful and articulate founders that I’ve met. As you’ll see in our conversation, it’s hard not to be captivated by Rahul’s storytelling skills and also his really insightful takes on how to build great products and teams.
This episode is for anyone who’s looking to build their product taste, help their teams move faster, learn how to think better from first principles. And also learn about Superhuman’s very unique approach to building their company, including why they manually onboarded every single new user for years and why they decided to stop. Why they ignored most of their customer feedback on their way to finding product market fit, and also how you can use his approach to finding product market fit for your own company. Also, the power of game design in building great products, a very contrarian take on pricing strategy, what Rahul has learned about building scaled products on top of AI and LLMs and so much more.
A huge thank you to Ed Sims, Conrad Irwin, Bell Trenchard and Gaurav Vohra for suggesting questions and topics for this conversation. If you enjoy this podcast, don’t forget to subscribe and follow it in your favorite podcasting app or YouTube. Also, if you become a yearly subscriber of my newsletter, you get a year free of Superhuman that you can start using immediately. You also get a year free of Notion, Perplexity Pro, Granola and Linear. Check it out at lennysnewsletter.com. With that, I bring you Rahul Vohra.
When I was at Airbnb, one of the things that I loved most was our experimentation platform where I could set up experiments, easily, troubleshoot issues, and analyze performance all on my own. Eppo does all that and more, with advanced statistical methods that can help you shave weeks off experiment time and accessible UI for diving deeper into performance. And out-of-the-box reporting that helps you avoid annoying prolonged analytic cycles. Eppo also makes it easy for you to share experiment insights with your team, sparking new ideas for the A/B testing flywheel. EPO powers experimentation across every use case, including product, growth, machine learning, monetization, and email marketing. Check out EPO at get epo.com/lenny and 10 x your experiment velocity. That’s geteppo.com/lenny.
The fund holds 10 by visiting fundrise.com/lenny. Carefully consider the investment objectives, risks, charges, and expenses of the Fundrise Flagship Fund before investing. Find this information and more in the fund’s prospectus at fundrise.com/flagship. This is a paid ad.
Rahul, thank you so much for being here. Welcome to the podcast.
Whale Users and Collaborative Products
Rahul Vohra: Hello, hello and thank you for having me Lenny.
Lenny Rachitsky: I have so many questions for you. We’re going to have so much to talk about. I actually want to start with your time before Superhuman. When I was preparing for this chat, I actually asked you, what’s the most pivotal moment in your career in your life? And you told me that other than starting Superhuman, it was selling your previous company Rapportive to LinkedIn. So let me just start there. What was that experience like? What do people not know about this phase in your life and just why was it so pivotal?
Baking Growth into Company Values
Rahul Vohra: So for folks that don’t know, Rapportive was my last company. It was the first Gmail extension to scale to millions of users. Basically on the right-hand side of Gmail, we would show you what people look like, where they work, links to their recent tweets, their LinkedIn profile and everything else that they were doing online. So if you were hiring, marketing, selling in BD, super useful. It turns out we somehow attracted most of LinkedIn’s daily active users onto this one free app, and I then ultimately ended up selling that to LinkedIn. That by far, as you said, was the most pivotal thing I’d done in my career, prior to starting Superhuman.
Now, had I known that we’d amassed most of LinkedIn’s active users onto one app, I would have sold it for far more. But the actual pivotal moment was really who I got to work with. Because I reported to LinkedIn’s head of Growth, Elliot Shmukler. He was responsible for scaling LinkedIn from 25 million members to when I joined, north of 250 million members. And during my first one-on-one, I learned the real secret behind virality and big hint, it’s not about viral mechanics. Overall that acquisition experience gave me the time to figure out what was next and the resources to truly swing for the fences.
Lenny Rachitsky: Okay. Well, I have to follow this thread that you put out there, what the secret is to virality. What did you learn there?
The Perceived Slowdown
Rahul Vohra: Well, in my first one-on-one, I sat down with Elliot and I said, “Hey, I’m here to learn. Please teach me everything that you know about virality.” And he said, “Okay. Well, hate to burst your bubble, but there is no such thing as a truly viral product.” I said, “What do you mean? How do you explain Facebook for that matter? How do you explain LinkedIn?” And he said, “What I mean is, no app has sustained a viral factor of greater than one for any real period of time.” Even Facebook in its heyday had a viral factor of about 0.7. And he told me that lasted for perhaps a year, so one person was creating about 0.7 new users.
I double-clicked again and I said, “Well Elliot, what about the address book import?” This is one of the things that LinkedIn got famous or infamous for. You could import your address book and then it would spam slash invite everyone who happens to be members of LinkedIn in your address book, and then eventually it would just invite everyone to LinkedIn. And he said, “That’s an amazing feature, but you have to remember not everyone is going to use it all the time.” So even that feature had a lifetime viral factor of about 0.4, and that’s considered good.
0.4 is good for a viral feature, 0.6 is great, something like 0.7 is absolutely incredible. You’re in the stratosphere up with Facebook at that time. So I said, “Well, okay, all of these things by definition are going to Peter out. There’s going to be an asymptote. None of these viral mechanics keep on compounding. Which actually makes sense, it would be a little absurd if things just kept on growing. What then is the true secret behind virality?” And he said, “It is word of mouth. It is the virality you can’t measure that isn’t a mechanic that isn’t in a feature. It is when one user spontaneously tells another user about your product.” That really colored how I think about growth and virality. Since then, it has shaped so much of what we do at Superhuman and so much of how I think about growing brands.
Lenny Rachitsky: Wow, you’re such a great storyteller. I’m just listening here, just captivated, “What is he going to say next?” That was fascinating. I actually have a post that I’m going to link to that, that very much aligns with what you’re talking about, which is titled, Virality is a Myth mostly. It’s based, I forget, on this book where they do all this research on actual viruses. It turns out they’re not actually spreading in this exponential way, there’s one person that spreads it to a lot of people and it keeps happening.
That’s actually apparently what the data shows. I’m curious if you found this same thing, which is, when people think of an app as going viral, it’s one person with a massive platform sharing it and their audience adopts it and that’s just one to many and then it just happens a couple of times and it looks like it’s going viral, but it’s a person to many people, not many people to many people. Thoughts on that?
Deepening Solutions and Expanding Markets
Rahul Vohra: Yeah, we’ve definitely found that there are whales, to use the gaming terminology, that one person is going to be responsible for inviting 25, 50, 100 people, and they may have various motivations for doing that. In Superhuman, as an individual subscriber, if you refer somebody else and they sign up, you both get a free month, which is a great incentive if you’re paying out of pocket.
We have people who send many, many hundreds of invites and there are some people who essentially have free Superhuman for life now due to how many people they’ve invited. But of course that incentive doesn’t necessarily work inside of a company or inside of a team where ultimately it’s the company paying for the product, so you have to then come up with new motivations for those people. That’s where there really isn’t any substitute to having a genuinely multiplayer or a genuinely collaborative product. That’s one of the huge evolutions we’ve taken Superhuman through over the last, probably about two years.
Early last year we launched what we call Superhuman 2.0. The basic idea is, we saw almost every single other app of note become collaborative by default, Figma, Notion, Loom. These are all multiplayer or collaborative by default. Yet email, the one tool that we all use more than anything else, even more than things like Slack, was still firmly stuck in its single-player origins.
Lenny Rachitsky: I want to come back to something that you mentioned that I didn’t come back to you that I think is really core to what you just shared, which is word of mouth being so important. People talk about all these viral features and sharing contact books and all these things. And your point is, that takes you to a place, but really what helps a consumer-ish product spread is word of mouth, people sharing with each other. Which, then the question is, how do you do that? We’re going to talk about a lot of things that you did to make Superhuman something people want to share, but in the end it’s just making something people want to share. That’s the definition almost. Then it’s like, what makes people want to share stuff? It’s amazing, it’s helping them, something that is remarkable.
Avoidable Slowdown and Management Restructuring
Rahul Vohra: Well, it turns out, because you mentioned remarkableness, that is one of our core company values. If you think about what a company has to do, it has to grow. How do things grow? Well, let’s take Elliot’s advice at face value, and I believe it’s true, it’s creating something that people share. You mentioned one way of doing it, which is something that people want to share. There’s actually another way, which is simply creating something remarkable, and you used that word, and that is one of the core values of Superhuman.
We have, create delight, create something that is so joyful that really truly brings people delight. We have deliver remarkable quality, something that is so striking, so compelling and worthy of attention that people can’t but help tell others about it. Then we have build the extraordinary, which is a measure of the efficacy or the innovativeness of what we want to build. That’s another trick, which is literally baking these raw ingredients for growth into your company values.
Time Tracking Methods
Lenny Rachitsky: I didn’t know that was one of your values. That makes so much sense. Okay, we’re going to come back to that, because I think that is… There’s so much to learn about how you think about product and how you think about building the company that builds the product. But I want to actually start here with how this conversation came to be.
The CEO of Product Hunt, Rajiv, tweeted months ago, he tweeted this and we’re going to show this if you’re on YouTube, “Superhuman’s product velocity feels like it’s kicked into another gear as of late. Does anyone else notice this?” I saw that, I’m like, “I completely have noticed this. It feels like there’s just feature shipping left and right, AI this, AI that. It feels like it’s just a new company.” And I tagged you on the tweet. I’m like, “Hey Rahul, what’s changed?” And you answered with a few things and it just made it clear there’s a lot to learn about what you did.
Because a lot of companies are in this phase of just, “Things aren’t moving as fast as we want. We used to be so much faster, we used to ship all these features and now we don’t.” So I think this is a really cool real case study illustrative example that we can analyze. So let me ask you this, what did you notice that told you something needed to change at Superhuman? And then what did you change that actually had the most impact on your ability to ship and move faster?
Insights Into Time Allocation
Rahul Vohra: I think what we noticed was this sentiment, and we felt it first ourselves, but we also started hearing it from the market, from our users, from our customers, that we’d slowed down. And as a founder, as a CEO, that’s the absolute last thing you want to hear. It’s our job after all to speed things up. When I ask people what do they mean by slowing down, they didn’t mean the product, of course, the product wasn’t working any slower, but that the pace of delivery seemed to have slowed down.
I think to break this down, it’s important to start by defining what we mean by a slowdown. There’s the kind of slowdown that is unavoidable in certain spaces, and then there is the kind of slowdown that is quite avoidable. We actually had both. So starting with unavoidable slowdown, you can classify anything that you build in a company into one of two categories, solution deepening and market widening. Now, solution deepening means making your product better for its existing users, but not making it available to more users. Whereas market widening means making your product available to more users, but not making the product itself any better.
There are some spaces, there are some markets, there are some platforms where market widening is really fast and really easy, and there are some spaces, email is one of them where market widening is really hard and really slow. But when we started we had a great deal of focus. We only supported Gmail, we were only on the web. In those early years, we could pour every ounce of R&D energy, every engineering dollar, into solution deepening, making the product better for existing users. And of course users loved it. It’s how we got to product market fit. It’s how most startups do.
But at a certain point, almost every company then has to start investing in widening the market. For example, the market of people who will use a new Gmail front end but without a mobile app, does exist, but it is relatively small. This is something that every new email startup is going to learn sooner or later. In order to keep on growing, you are going to have to need to add an iOS app and then a MacOS app, and then a Windows app, and then an Android app. Then you’ll soon want to support Office 365. But that’s not one thing, that’s actually three things, because you have to support Office 365 on desktop and then on iOS and then on Android. That’s all much easier said than done.
I think we at Superhuman now know things about these APIs that literally no other company knows, and I would not wish it upon my worst enemy. So fast-forward to today, and Superhuman now works wherever you do on every combination of Gmail, Outlook, Mac, Windows, Web, iOS, Android, and this actually turns out to be a really great technology moat. Almost no other email app can claim this. It’s taken many years of intense investment. I think we’ll touch on this later, but it’s one of the main reasons why we can sell into the enterprise, because we now know everyone can use it.
But this is the hard part, when you’re doing that market widening, you’re not solution deepening, so your perceived product velocity may decrease. You can avoid some of these things with some smart technology decisions, but mostly you just have to grind through it, and it is worth it to get to the other side. Then there’s the kind of slowdown that is avoidable. If I remember my answer to Rajiv’s tweet, that was the kind I was talking about. In that case it was our management structure, or who does what.
When we hired our initial executive teams, I followed very conventional wisdom. I ended up with a set of VPs and eight, I think direct reports, maybe even nine. I thought that’s what you were meant to do. That’s how startups are meant to scale. But as anyone who’s been there knows, eight direct reports is a lot. It’s a lot of hiring, it’s a lot of goal setting, it’s a lot of OKRs, it’s a lot of accountability conversations, and fortunately also it’s a lot of firing. No CEO ever gets their executive team right on the first try. That time I had for the things that I think I can genuinely be world-class at things like product and design and technology and marketing, that all began to rapidly disappear, and as a result the organization began to slow down.
Unfortunately, I was also tracking my time very closely, I had this crazy way of tracking it. At one point I noticed I was spending six to 7% of my week on these areas, these areas where I can truly be world-class at. So I had two realizations. Number one, as CEO, once you get to a certain scale, and we were definitely at that scale, you can actually define what you want the role of a CEO to be at your company. And number two, the Superhuman opportunity deserves everyone who works at the company to spend as much time as possible in their zone of genius, so that includes me as well as everybody else. What I did is, I hired a really great president, I went from eight direct reports to two, and the amount of time that I spend on product design, technology and marketing went up from six to 7% to about 60% to 70% of my week.
Lenny Rachitsky: Just to mirror back a few things. One is, people may feel like you are not shipping as much as you used to because you’re actually building things they don’t care about, which is support for office and all these things that they don’t need, but the business needs to expand, integrations with Microsoft and Android and all these things. I think that’s such a good point, that it looks like nothing’s happening when there’s a lot of good stuff happening for other users that aren’t you.
Then there’s this point about people delegate. Then a leader of delegates, hires all these execs and they’re like, “This is not what I wanted. Why have I done this?” And you think it’s going to speed up, but it slows down. A couple threads here that are really interesting. One is this time tracking thing, I need to know how do you do this? The fact that you knew seven to 8% or whatever the number is, to that granularity of your time you’re spending on things you wanted was that low, how do you do time tracking? Let’s not go super far, but just what’s your approach?
On Context Switching
Rahul Vohra: This is a technique that I call the Switch log. It’s born out of the observation that your calendar says what you thought you were going to do, but it’s really only your trail of work that describes what you actually did. So how can we capture that? And actually, how can we create a system of work that isn’t tethered to a calendar, where you aren’t at the behest of what some timetable says you do or you don’t have to do? So I came up with the following idea, what if I just did whatever the heck I wanted? What if every single time I change task I just Slack DM’d my EA, but this also works in Slackbot, it just has to go somewhere. I Slack DM’d my EA and I said, “TS:,” and then a few words for the task I was doing.
Well, that would create certain changes. Instead of having to constantly look at the calendar and think, “Oh, should I stop this task, start that task, I can just do what I want.” If what I feel right now is, “Oh boy, I really need to prepare for Lenny’s podcast, I’ll go ahead and do that.” And if I get bored or distracted eight minutes in, which sometimes happens because something else just bubbles up to the top of my mind, well, there’s a reason that my body is bubbling it up to the top of my mind. I also practice transcendental meditation, so I’m very keen on the idea of being aware and listening to what’s bubbling up.
So it’s okay for me to then go and attend to that thought as opposed to start to expend my focus points or my discipline or willpower on the thing that I thought I was meant to be doing. All I’d have to do is I’d go back to Slack, “TS: Dealing with this other thing.” And by the way, you should obviously turn up for your meetings. I’m not saying just blow through your meetings and not turn up for your one-on-ones. Definitely do those things. What I’m saying is, do what feels right for as long as it feels right to do. Then at the end of the week you can see where your time is going.
I realized at one point that I was spending only in those days 5% of my time on recruiting, whereas perhaps I should be spending 20 or 30% or more of my time on recruiting. But the biggest thing was, I saw I was only spending six to 7% of my time on product, on design, on technology and marketing. These are things where I know I’m really good at them. I should either be teaching people how to do them or doing them or some combination of both. That’s probably the best thing for me. It keeps me really happy, very joyful, it keeps me sharp, but it’s also scaling the organization. So that’s how we had that kind of an insight. Once you have this Slack Log, you can then graph it and chart it and see where your time is actually going.
Lenny Rachitsky: How cool. Clearly this is an app opportunity or an agent opportunity where you’re just telling this thing every time. It’s essentially tracking context, which we’re always hearing, try not to avoid context which switches.
Practicing Transcendental Meditation
Rahul Vohra: I think context switches are fine. There’s definitely this idea that, for every interruption you have, the brain does take roughly 21 minutes on average to recover, to get back to the efficacy before that you were disturbed. It’s a big deal, of course, I’m building productivity software, we designed Superhuman to minimize the amount of distraction and disruption that’s possible within the app. But if you are working on something and at the back of your mind something bubbles up, you have to attend to it in one way or the other. Sometimes I just write it down, actually, I don’t have my notebook with me, but it’s really big. I have a gigantic, whatever twice the size of A4 is, I guess A3 sketchbook and I always have a 4H pencil, so whenever one of those thoughts comes up, I just scribble it down. Or I actually stop what I’m doing and I attend to that task, because there’s a reason it’s bubbling up right now.
Lenny Rachitsky: I love that you know exactly the type of paper and pencil, 4H pencil, A3 paper, [inaudible 00:24:51]. Okay, this is going to be a theme. You mentioned meditation, you said you do TM, so you do 20 minutes in the morning, 20 minutes… Do you do it that style or you do a longer session?
Advice for Learning Transcendental Meditation
Rahul Vohra: I do about half an hour in the morning, including rest time. The physical rest component of it is very important to me. So it’s 20 minutes of the actual meditation, then 10 minutes of rest. I do that in the morning as well as in the afternoon at around 3:00 PM
Lenny Rachitsky: And you just carve that out in your calendar. Everyone knows Rahul at three o’clock, he’s going to be out.
Hiring a President
Rahul Vohra: Absolutely. My EA knows, they’re the one who’s organizing the calendar and making sure things happen when they need to happen. They also know that nothing can override this TM block. Without it I genuinely start to fall apart. But with it, I’m able to access some very deep competencies that I didn’t have before. I’ve been doing this now for about four or five years, and initially I simply felt happier, occasionally even more euphoric coming out of a really great meditation session. But over time I found that my ability to focus was increasing. I could hold attention on something for much longer, but I also was able to become much more creative and much more expressive.
These are well-known side effects, as it were, or intended effects for some people of TM. And interestingly about TM, if you compare it to other forms of meditation, they don’t have quite the same impact across quite as many executive functions. So there’s something particularly interesting that’s going on with transcendental meditation as opposed to other forms that folks are still trying to unravel and figure out.
Lenny Rachitsky: If folks want to, if they’re inspired and they want to check out this form of meditation, any advice on where they could go learn?
A Different View of the President Role
Rahul Vohra: Absolutely, a lot. But in summary, have a coach teach you. I had many false starts myself with meditation, trying the various apps, learning from books. None of it really worked for me. What worked was having one-on-one teaching from someone themselves who had been taught one-on-one the Yogic or the Raja tradition of teaching. This person in particular had also been a venture-backed founder multiple times over, so they’re very well aware of the kinds of stresses that I tend to be under. And all of his clients are mostly in technology as well. If you’re in the Bay Area, this person’s name is Laurent Valasek. They run an institution called the Peak Leadership Institute. And this is all about how we can live a more integrated and whole life. Integrating wellness practices like meditation, but for the purpose of unlocking peak performance in life and in business.
Extreme Attention to Detail
Lenny Rachitsky: Thank you for sharing that. That is very actionable. We’re going to link to that in the show notes.
Okay. I’m going to try to bring us back on course. The other thing you mentioned that I think is really interesting is hiring a president. A lot of founders and leaders might be hearing this and be like, “Going from eight reports and doing all these things I don’t want to, spending most of my time on the product and design and marketing, amazing.” What did this president take off your plate and what is their responsibility and that allowed you to do the stuff you wanted to do?
Rahul Vohra: The biggest thing was taking off the operations and the management of the executive team and the rest of the company. Think of the president role in Superhuman as an operationally extremely challenging and a very growthful role. It is perfect for someone who wants to go on to be a CEO in their next role. Instead of hiring and firing that team, instead of managing and setting their goals, instead of the accountability conversations, someone else who’s now doing that.
In addition, because that’s not the only job, in addition, they’re also a very strong thought partner when it comes to corporate strategy. When it comes to, where do we take act one, our email product? How far do we go down the multiplayer path? How aggressively should we lean into AI? What’s a reasonable gross margin in a world with AI? Are we from a financial perspective okay dipping now and then coming back later? When should we start building our second product? How do we think about our R&D strategy? Should we keep on hiring in the Bay Area, or as we’ve done for many of our recent hires, should we continue hiring in Latin America? Should we consider other time zones as well? And so on and so on and so on. I’m just randomly coming up with questions, but the list is truly endless.
Another way to think about it is, it’s almost like a grown-up co-founder. The two people I co-founded the company with, Comrade and Vivek, they’ve long since gone from Superhuman. We’re now a 10-year-old organization and I’m one those rare founders that is persisting and thriving actually 10 years in. That said, the journey never gets easier, it gets different and you still need that co-founding energy around you. I have a handful of people in the organization who are in their roles providing that kind of energy, that kind of input, and who thrive off doing so. Then the president role is definitely one of them.
Font Optimization and Email Experience
Lenny Rachitsky: Incredibly interesting. There’s so much there. One, just a couple of things I’ll share and then I want to move on to a different topic. One is just, it’s cool the solution to helping you move faster and do the work you want to do is org design. That feels like a really doable thing. If you’re finding you’re not spending time on things you want to spend time on and things aren’t moving as fast as you want, it’s essentially you can find people to take on things that you don’t want and shift the way that the org is structured and that could solve a lot of problems. That’s what it did for you. Then I think it’s also really interesting, there’s this lesson here of as a founder, if you’re just feeling depleted or just don’t have the partner you want, you could bring someone on that could be that person.
Launch Timing Depends on Business Model
Rahul Vohra: Absolutely.
The Craft Versus Speed Tradeoff
Lenny Rachitsky: Okay. There’s so much there. That was much more of a rich area than I even expected. I want to zoom out a little bit, and there’s a couple themes that came up again and again when I talked to folks that you’ve worked with, investors in Superhuman. The two themes are contrarian thinking, in terms of building the company, and strong attention to detail. Let’s spend a little time on attention to detail. Like I said, this is one of the things that came up again and again when I was asking people about you. So I have this quote from Ed Sims, and maybe your first investor. Were they your first investor?
Speed and Positioning
Rahul Vohra: Yeah, that there’s a bunch of people on Twitter who are going to fight for that. But to set the record straight, Ed Sim did actually write the first three checks into Superhuman.
Unconventional: Manually Onboarding Every User
Lenny Rachitsky: First three checks? At subsequent rounds.
Rahul Vohra: Well, yeah. Quick sidebar on that, he runs Boldstart Ventures alongside his partner Elliot Durbin. They have a particular interest in backing second-time founders, but they’ll also back first-time founders, and they love application and infrastructure areas like Superhuman, so we were like the perfect investment. He also wrote a check from his previous fund into a Rapportive, and I think I’d made him five X that money. Nothing to write home about, but definitely, “I’m going to back this guy again.” So I went to him and I said, “Hey listen, this is going to sound crazy. I want to take on Gmail.” He said, “Do you have a deck?” I was like, “Yeah, here it is one slide, here it is.” And there was a screenshot of Gmail with most of it scribbled out, “I want to build that and it’s going to be amazing.”
So he said, “Cool, we’re in. Can I wire you the money?” And I said, “No, I don’t even have a bank account yet.” I come back two days later with a bank account and he’s like, “Cool, I want to wire you 750 K.” And I said, “I don’t even know what I’m going to do with that money. I’m not paying myself, I won’t for a while. We don’t have any employees. I can’t think of anything I want to spend it on. Tell you what, I’ll just take 250 K.” And he was like, “What?” I’m like, “Yeah, I’ll just take 250 K.” We start having the conversation around venture economics. I’m like, “Yeah, it’s fine, we’ll figure it out.” Then a few months back I took another 250 K and a few months back I took another 250 K as I began inventing ways and finding channels to deploy capital properly.
Achieving Product-Market Fit
Lenny Rachitsky: I love this story. I love all these stories you’re sharing I’ve never heard before. And by the way, it is awesome. We’re talking about him coming on the podcast, maybe breaking our VC rule. So specifically the story he shared with me that is maybe an example of you and your attention to detail is, he said that you created your own font because existing fonts weren’t good enough. Is that true?
Rahul Vohra: Kind of. Okay. The font that we use today is a modified version of Adelle Sans. The story there is, I looked at all of the major font families, and honestly none of them was what I would call truly excellent. That may sound like an odd thing to say. So let’s, if you will permit me to talk about typography and email-
Building a Roadmap with Algorithms
Lenny Rachitsky: Please.
Game Design Versus Gamification
Rahul Vohra: The first thing we did was, we took our UI and we laid it out in about 15 different styles using examples of the major font families. We actually printed these out and we left them on a desk in the middle of our office. Sometimes with design, you want to tune in to your immediate most visceral response, but sometimes you want to truly let a design marinate. And this was the latter. So we let these designs marinate, we let these font choices percolate. Like I said, none of them was truly excellent.
Number one, I was looking for a font that was in and of itself gorgeous. Number two, I was looking for a font that could also convey a message of any kind, without overpowering the sentiment of that message. For example, does the font work when this is inviting you to a party? Many fonts, including almost all serif fonts, are actually too somber or too sober for that. Or to pick another extreme, does the font work if it is informing you of somebody’s passing, many fonts are just too jaunty for that. You wouldn’t want that kind of message in Comic Sans, for example. And number three, I was optimizing for a font that made reading speed and comprehension really fast. And number four, I was looking for a font that made email addresses themselves look great. So I discarded all the 15 because they weren’t good enough, and after searching high and low, I came across a font called Adelle Sans, which is designed by a foundry called Type Together, type-together.com. They have a whole bunch of lovely fonts, go check them out.
And if you go through my list, number one, Adelle Sans is gorgeous. I think each character is a work of art. It’s beautifully formed. Number two, Adelle Sans is, I would say upbeat, it’s optimistic, yet it’s serious enough to convey any kind of message. It has just the right amount of personality, yet not too much personality. Number three, Adelle Sans is also unusually narrow, and that actually fits email particularly well. One of my pet peeves with Gmail, which by default uses Ariel, is that the lines are as wide as your window. So if you’re in a wide screen, then the lines get really arbitrarily long. The problem with really wide and really long lines, is that they decrease reading speed. Because by the time you’ve reached the end of one line, your eyes have lost track of the start of the next line. And Ariel itself has fairly wide characters, which further exacerbates that.
So at Superhuman we, if you’ve used the product, you know this, we fix the line length or the typographical measure to the optimal length for reading speed, which depending on the font is around 90 to 120 characters. And Adelle Sans is quite narrow, so it actually lets us do this on quite small windows with fairly dense line. So we get a lot of information on fairly small windows without getting a very long typographical measure, optimizing for reading speed and for comprehension. Then number four, finally Adelle Sans has very unusual treatment of the at symbol in an email address. It actually puts the base of the A in the at on the same baseline as the rest of the text.
So for example, if your name has an A, my name does Rahul at Vohra, three A’s and or two A’s and an at, they’re all actually on the same baseline. It’s a small thing, but it makes the email addresses look incredibly natural. If you look at that and then you actually look at email addresses laid out in other fonts, those other the fonts look really clunky and awkward because the A is kind of shifted around and it just looks a bit silly in my opinion. Now Adelle Sans isn’t perfect. So we then worked with a type designer on some of the specific details that there are some of the glyphs, which get a little pinchy as it were, and what we use today is very close to retail Adelle Sans.
Lenny Rachitsky: And this was pre-launch or this was after you’d already launched?
The Pricing Strategy
Rahul Vohra: We’d probably had about 10, 15 users at the time.
Superhuman’s AI Features
Lenny Rachitsky: So I think that’s pretty contrarian unique to be this focused on the font and the typeface before you even launched. This was like, “Is this even going to be a thing? Will anyone even care?” And I think this says a lot about the way you think about product.
Rahul Vohra: Oh yeah, that thought never crossed my mind. I think we’ll probably come to it later, but the idea that, is this never going to be a thing? I think that’s a dangerous thought. We can’t start thinking that way, because at what point do you stop second-guessing yourself?
Shifting from PLG to Enterprise Sales
Lenny Rachitsky: Interesting. So you were confident this was going to work, so because I am so confident it’ll work, I need them to get this right. There’s also this trap founders fall into of just spending too much time perfecting a thing that never works and there’s always advice launch early, launch often. Thoughts there? How do you find that balance? What’s your advice there?
Single Deciding Reason (SDR)
Rahul Vohra: How much to spend time ahead of launch really does depend on the markets and the structure, the nature of your business model. For example, let’s say you are building a marketplace in a greenfield opportunity, so imagine the Lyft or Uber in their heyday. There’s a strong network effect, because the more cars you have on your platform, the shorter waiting times are, therefore people are going to preferentially use your app versus the other person’s app. That’s when there’s no time to spare, that’s when you probably shouldn’t even be sleeping. You’re going to hire the most aggressive maniacal people possible. You’re going to work 120-hour weeks, because every marginal minute actually does matter. Every marginal minute in the market, growing compounding is going to make your next year even better.
That’s actually not true of all startups and it certainly isn’t true of something like Superhuman. Yes, working harder is always better and we work tremendously hard at Superhuman, but not to the point where it made sense to release something that didn’t work. I’m reminded of a story of a founder that was in Y Combinator, told me about their demo day experience. They used Mailbox, which some folks may remember was also a startup, and Dropbox famously acquired them for about a hundred million dollars. The reason that they were well known, apart from the acquisition, is they were the first to popularize, swipe to archive or swipe to mark down, which of course is now standard in Superhuman and every other app.
This founder was using Mailbox and was having an amazing demo day. They’re working the room, they’re meeting investors, they’re pitching their photography app in this case. He went home that night and went to his laptop, fired up mailbox and sent off a bunch of follow-up emails. He waited the day, didn’t hear back, he waited two days, didn’t hear back. On the third morning he figured something was up, so he fired up Gmail, went to his sent mail, and you guessed it, there were no sent mails there. So something had broken with mailbox. So he’s cursing to himself trying to remind himself everything’s going to be okay. Sent all the same emails from Gmail manually and they all came through.
But then one of the investors said, “Hey, by the way, you might want to check your email clients, because I’ve been getting some of your emails twice.” Now he goes back into his Gmail, he sees that yes, actually the original emails that were queued up in mailbox have now indeed been sent, and some of the investors, and unfortunately most of the investors he actually pitched twice. Now, is this the end of the world? No, an investor can overlook that. Probably a good thing that you’re trying new apps. But was it horrifying and was it really scary? Absolutely.
Imagine this wasn’t investors, imagine this was a customer, someone who you were trying to convince to buy your thing and that you knew what you were doing and you had attention to detail and you had everything just buttoned up and under control. Well, now you’ve lost face, now you look foolish. That’s why when you have mission-critical products like email where you are interfacing with customers, with candidates, with investors, it turns out to really matter. Email is mission-critical. It’s not something where you can simply launch with a half-baked product.
Lenny Rachitsky: This is such an important nuance take on, there’s always this debate, how much to focus on craft and user experience, how much to focus on time to launch and get it out and speed. What I’m hearing here, which I completely agree with is, it depends on the market you’re in and the criticality essentially of your product. So if it’s email, it just needs to work and you need to get that right, you need to spend all the time, you need to get that right.
This reminds me of something else that when your early investors shared with me, Bill Trenchard from First-Round Capital. He talked about how speed was the thing that you just dialed up as a lever to 11. That’s where you just, “We will make this the focus. Speed, speed, speed.” I think maybe the lesson there is, you pick the thing that you think will most differentiate you, make you significantly better than what’s out there. So just thoughts on how you decided speed was the thing you were going to obsess with, and advice for folks that are trying to decide where to dial up things to 11?
Rahul Vohra: Bill is right and I agree with him, you have to pick something. Knowing what to pick is the trick. In the early days of Superhuman, I read a book on positioning that really influenced my thinking. It is, I believe called Positioning the Battle for Your Mind. It struck me how the most well-known brands have stood for one clear thing, they have a clear position. So in order for Superhuman to be memorable, I believed that we needed to occupy a clear position that was unique and which was available and which reinforced our product strategy.
In the first year of Superhuman, therefore, I interviewed hundreds of potential customers about their experience with Gmail and with Outlook. And predictably, almost everybody says that email takes way too much time. But interestingly, many people also said that Gmail and Outlook were way too slow. That was how I first thought that speed could be an interesting position for us. I then asked myself, “Is the position of speed unique and is it available?” And the answer was overwhelmingly yes, because almost no software was being sold or has ever been sold on the value proposition of speed. The last time I could remember anyone trying to do this, was when Google launched Chrome, and obviously that went incredibly well for them. You may remember they had slow-motion videos where they were comparing Chrome render webpages and showing that was faster than an actual strike of lightning. No one had done it since then.
I then asked, “Well, does speed reinforce our product strategy?” And again, the answer was overwhelmingly yes. I knew that our competition was not going to be startups, it was incumbents. And I also knew that incumbents generally struggle with speed, because by definition they have massive scale and usually entrenched architecture. Then finally I did what I call the cocktail party test, which is to look at the cocktail parties and to watch how people pitch your product to other people. In our case the pitches were simple. People would say, “Dude, you have to use it, it’s really fucking fast.” And that’s it. That was the pitch. That’s how I knew that speed would be a really great position for us to start with.
Lenny Rachitsky: I’m excited to chat with Christina Gilbert, the founder of OneSchema, one of our longtime podcast sponsors. Hi Christina.
Christina Gilbert: Yes, thank you for having me on, Lenny.
Lenny Rachitsky: What is the latest with OneSchema? I know you now work with some of my favorite companies like Ramp, Vanta, Scale and Watershed. I heard that you just launched a new product to help product teams import CSVs from especially tricky systems like ERPs.
Christina Gilbert: Yes, so we just launched OneSchema of FileFeeds, which allows you to build an integration with any system in 15 minutes, as long as you can export a CSV to an SFTP folder. We see our customers all the time getting stuck with hacks and workarounds, and the product teams that we work with don’t have to turn down prospects because their systems are too hard to integrate with. We allow our customers to offer thousands of integrations without involving their engineering team at all.
Lenny Rachitsky: I can tell you that if my team had to build integrations like this, how nice would it be to be able to take this off my roadmap and instead use something like OneSchema. Not just to build it but also to maintain it forever.
Christina Gilbert: Absolutely, Lenny. We’ve heard so many horror stories of multi-day outages from even just a handful of bad records. We are laser focused on integration reliability to help teams end all of those distractions that come up with integrations. We have a built-in validation layer that stops any bad data from entering your system and OneSchema will notify your team immediately of any data that looks incorrect.
Lenny Rachitsky: I know that importing incorrect data can cause all kinds of pain for your customers and quickly lose their trust. Christina, thank you for joining us and if you want to learn more, head on over to OneSchema.co. That’s one OneSchema.co.
The next area I want to spend time on and I imagine we’ll have much insight is, some of the contrarian ways you approach building Superhuman that a lot of companies never thought about doing that you did that worked out for you. So the first is manually onboarding every single new user. Sure, startups have done this, founders bring on some folks and then cool, show it to them and then they stop doing that and then it’s self-service or sales teams. How far did you scale this manual onboarding phase of your company? How many people did you have onboarding people, how many people did you manually onboard?
Rahul Vohra: So for folks that don’t know, in those early days we insisted on one-to-one concierge onboarding, and it was absolutely the right thing to do. You couldn’t use Superhuman unless you went through the onboarding experience. Now it’s almost the reverse. Almost every new Superhuman customer goes through self-service. The onboarding experience is still there, but again it is absolutely the right thing to do. To answer your question, at peak we had about 20 people doing manual onboarding.
Lenny Rachitsky: Okay, so it’s not that many people. That’s really interesting. Because I always imagined it was like a massive team, but 20 people can handle a lot, is the takeaway there. What was the scale where you stopped manual onboarding, just for folks that are thinking about doing this and then when to stop?
Rahul Vohra: I think the reason to stop is that there will always be certain personality types who do not want to go through a one-on-one onboarding. At a certain point those people will become very important, and you’ll need to be ready with a world-class self-service option. When we started building self-service, it seemed nearly impossible. In fact, it was terrifying, because it’s difficult to overstate how much the entire DNA of the company was built around this idea that we would onboard users manually. After all, we did so much in our one-to-one onboardings and there’s only so much that software can do. Now, we did after a lot of grind and persistence eventually figure it out and we have a world-class self-service experience today, but we did not at the time.
So the flip side is, why would you even do this to begin with? What we found is two things. Number one, the user metrics are excellent for things like engagement, retention, product market fit score, MPS, virality, for all of those metrics. I think you you’ll significantly beat your industry benchmarks if you go to the effort of one-on-one onboarding your early customers. It becomes so powerful to have that early cohort of super fans when it comes to things like building a brand. If folks remember that conversation from way up at the top, what is it that creates true virality? It’s not viral mechanics, it’s word of mouth. It is brand. This is how you can kickstart a brand.
And number two, in a world where you can easily and quickly raise funding, like for example the zero interest rate phenomenon era, you can actually use dollars to avoid building a first-time user experience and all of the normal growth loops that you would then have to build. You would then instead focus all of your engineers on finding product market fit or in solution deepening or in market widening, but not for example on a first-time user experience, not for example on activation, because you have humans doing activation for you. By contrast I saw other companies often competing spend almost half their engineering dollars on those things, on self-service flows for products that ultimately did not find products market fit. So makes sense to do if you really want to create that brand, which I think all consumer-ish companies need to do. And if there is money falling off trees, for whatever reason, which we did have for a period of time, arguably AI companies have that again today. So if you can weave this into your strategy, I think you should, but you should also know when to stop.
Lenny Rachitsky: Super interesting. I guess some factors to think about, because I wanted to ask you when should people consider doing this? If they’re hearing this and they’re like, “This is awesome, so many problems solved if I just have somebody onboarding every new user, everyone’s activated. Amazing.” So some of the variables you’re sharing is, do you have like cheap cash to invest in say, it doesn’t have to be 20 people, it could be a few people to start. Then if there’s an LTV, ACV element of just are you going to make enough from a new customer? Imagine that’s a variable. Is there anything else you think founders should think about?
Rahul Vohra: Absolutely. You don’t want to lose money doing this. We always made money doing onboarding to be clear, it’s just that at a certain point the mass market, whether it for us it’s enterprise or all of the prosumers in the world, you hit a top of funnel width, it needs to be wide enough where manually onboarding no longer makes sense.
Lenny Rachitsky: Awesome. Okay, let’s talk about product market fit. I know that everyone, when they think of Rahul, they think product market fit. You wrote this epic First Round post that described the way you guys approach product market fit. We’re not going to spend a lot of time on describing it, because people can look it up. So let me just ask you this, what are a couple of things that you think people still don’t understand about finding product market fit, getting to product market fit? Considering it’s the most important thing you got to figure out as a founder. If you don’t find something people want, nothing else matters. Anything there you want to share.
Rahul Vohra: The core ideas are still weird enough that I’ll start there. Which is number one, you can measure product market fit. Number two, you can optimize product market fit. Number three, you can systematically, even numerically increase product market fit. And number four, you can even have an algorithm write your roadmap for you, and that is a roadmap that is guaranteed to increase product market fit. Now, if that sounds crazy, I would be the first to admit it doesn’t seem like that should be true, but go check out that post. I think it is still the most widely shared post on First Round Review, it’s called How Superhuman Built an Engine to Find Product Market Fit, or just Google the Superhuman Product Market Fit Engine. And you’ll see the algorithm laid out there fully explained and why it works.
I’d say the second thing is to get to product market fit, you have to deliberately not act on the feedback of many of your early users. This is at the same time as listening to people intensely and building what people want. That’s what we’re here to do, is to make something that people want. But it can’t be all people. It can’t be everybody. The question becomes, how do you listen to them? And then even of what they say, what do you pay attention to and what don’t you? All of that’s covered in the Product Market Fit Engine.
Lenny Rachitsky: Okay, I got to follow this thought on algorithmically building your roadmap to increase product market fit. Talk about how one would do that.
Rahul Vohra: Well, that’s really the meat of the engine. Let’s see if I can condense it here in a very easy to grok fashion. Let’s assume for the sake of argument, that you can put a number on product market fit, and it turns out you can. Very simply, you’re going to ask people, “How would you feel if you can no longer use this product?” You give them three responses. One of them is very disappointed, the other is somewhat disappointed, and the other is not disappointed. Very disappointed means, “I’d be devastated. I love this product. I need this product.”
What Sean Ellis found, Sean Ellis, if you don’t know him, is the guy who coined the term growth hacker, and he instrumented, benchmarked this initial question. What he found, is that the companies that struggled to grow almost always had less than 40%, very disappointed. Whereas the companies that grew the fastest almost always had more than 40%, very disappointed. And this question, this metric is way more predictive of success than something, for example, like net promoter score.
Okay, so far so easy. How do we make this number go up? Well, you want more people to be very disappointed without your product. The trick here is not to act too much on the feedback that the very disappointed people are giving you, because they already love your product. Also, not to act at all really on the feedback that the not disappointed people are giving, you because they’re so far from loving your product that they’re essentially a lost cause. But to focus on the segment of the somewhat disappointed people, they kind of love your product, but something, and I would wager something small, is holding them back.
You then divide them into two camps, the camp for whom the main benefit of your product resonates and the camp for whom it doesn’t. What do I mean by that? Well, you go back to the people who really love your product and you basically ask them why? What is it about my products that you really love? In the early days of Superhuman, it would have been speed and keyboard shortcuts and the overall design aesthetic as well as the time that we were saving you. You then go back to the somewhat disappointed users, and in the Superhuman example, I would simply ask, “Wait, do you like Superhuman because of its speed or for something else?” And if it’s something else, well, and this is hard to do, but politely disregard those people and their feedback. Because even if you built everything that they asked for, they’re still pulling you in a different direction. And the thing that they like the most from your product isn’t actually what the people who en mass love it the most for, is.
You have then articulated the subsegment of the subsegment that it makes sense to pay attention to, and there’s another question in the engine to figure out what they don’t like about the product. Now you have a list of things people love, you have a list of things people don’t love, and you can work down that list to make the product market fit score go up. And basically at the start of every planning cycle, I advise spending half your time doubling down on what people really love and half your time systematically overcoming the objections of the somewhat disappointed users, but specifically those for whom the main benefit resonates.
Lenny Rachitsky: That was an excellent summary. I know I said we wouldn’t spend a ton of time here, but I’m really glad we did. That was really helpful. Let me ask you this, I know you used this initially in the early days, are you still operating in this way in some form?
Rahul Vohra: We don’t run the engine as is for Superhuman as a whole. There are enough subcomponents of Superhuman now that are almost individual products. For example, Superhuman for Sales, our multiplayer and collaboration features, how we think about the enterprise, AI is its whole thing, but we do sometimes run it on those individual pieces. For example, we’ll ask a salesperson, the Product Market Fit Engine, as it relates to Superhuman for sales. As we think about starting new products, we would absolutely deploy the product market fit engine.
Lenny Rachitsky: Awesome. The way you ask this question is an in-product interstitial sort of survey pop-up thing?
Rahul Vohra: You can do it however you want. The way Sean initially benchmarked the number was via email surveys. I think email surveys work just fine. The key thing is, and this applies to any survey methodology, if you’re going to change the method of surveying, all of your old numbers are invalidated. So it’s just a new baseline going forwards.
Lenny Rachitsky: Got it. We had Sean on the podcast and he describes this method in detail. So if folks want to explore the Sean Ellis test, listen to that podcast. We’ll link to it.
Okay, next topic that I’m excited to get your take on, is game design versus gamification. This is one of the more unique ways you think about designing product. When people hear you talk about this, they think it’s like, “Oh, gamification making things like games. Oh, it’s Zynga, Farmville, I don’t want to do that.” But you actually have a really different perspective on why you need to think about game design as you design products. Talk about your insights there.
Rahul Vohra: Well, I strongly believe that we should make business software like we make games, because when we make products like we make games, people find them fun. They tell their friends, they fall in love with them. It’s another way actually of backing into where we open this conversation, which is you’re making a brand, you are giving reason for word of mouth. It’s actually an altogether different kind of product development. So how do we do this? Well, as you’ve said, it’s not gamification, that doesn’t work. Game design works, but game design is not gamification. It’s not, for example, simply taking your product and adding points, levels, trophies or badges.
To understand why gamification does not work, we actually have to start with human motivation. There’s a very interesting study from Stanford that demonstrates the difference perfectly. In the 1970s, these Stanford researchers recruited children who were aged three to four years old, and all of these kids were generally pre-interested in drawing. Some kids were told they would get a reward, a certificate with a gold seal and a ribbon. And some kids were not told about any reward and they did not even expect one or didn’t know of one. Now each child was then invited into a separate room to draw for six minutes and afterwards they would either get the reward or not.
Over the next few days, the children were observed to see how much they would continue to draw by themselves. So the children with no reward, they spent 17% of their time drawing, but the children who expected a reward, sadly they only spent 8% of their time drawing. The very presence of a reward halved their motivation. So what’s happening? What’s happening here, is researchers differentiate intrinsic motivation and extrinsic motivation. With intrinsic motivation we do things because they are inherently interesting and satisfying, and with extrinsic motivation, we do things to earn rewards and to achieve external goals. That’s the problem with rewards, is they just massively undermine intrinsic motivation. That’s why gamification doesn’t work. And when gamification does work, it’s because the underlying experience was already designed like a game.
Lenny Rachitsky: What makes something like a game? I know Superhuman is really good at this, of just your inbox zero quest that you’re on. Just to make that a little more real, what is game design? What does that mean to you? What makes it feel like a game?
Rahul Vohra: Well, maybe folks don’t know this, but before I was a founder, you can probably tell, I was actually professionally a game designer. And as it turns out, there is no unifying theory of game design. To create games, what we need to do is draw upon the arts and the science of psychology, mathematics, storytelling, interaction design. And at Superhuman we’ve identified five key areas that we really care about, goals, emotions, toys, controls and flow. And across these we’ve identified many principles of game design. One example principle would be, make fun toys and then combine those into games.
A question I like to ask is, are toys the same as games? They do seem different. For example, we play with toys, but we play games. A ball is a toy, but football is a game. As it turns out, the best games are constructed out of toys. Why? Because then they are fun on both levels, the toy and the game itself. So for example, in Superhuman, one of our favorite toys is the time auto-completer. If you use Superhuman, this is the thing that appears when you hit H, when you snooze or set reminders on emails. You can type whatever you want, it can be gibberish and it does its best to understand you. For example, if you type in 2D, that becomes two days, 3H is three hours, one MO is one month. The time auto-completer is fun because it indulges your playful exploration.
In onboardings, it wasn’t long before I saw people asking, “What can it do? Where does it break? How does it work? What happens if I keep on typing in a series of tens? Well, it turns out that’s October the 10th at 10:10 PM. Well, how about a series of twos? Well, that’s February the second, 2022 at 2:00 PM.” Then you start trying more complex inputs like in a fortnight and a day, and that works, which is a pleasant surprise. And it’s not long before you find more pleasant surprises like time zone math happens without you thinking about it. You can just type in 8:00 AM in Tokyo and it turns out that’s 8:00 PM Eastern Time and you no longer have to do the time zone math.
Then most people were really delighted to find out that if you really want, you can snooze emails until never, i.e. you can literally type in never, and the email will never come back. It had like a little shrug emoji at the same time. Is this toy going to win awards? Nope. But is it fun actually, surprisingly yes. So what I would encourage people to do is, think about the features of their product. Do those features indulge, playful, exploration? Are they fun even without a goal? And do they elicit moments of pleasant surprise? If so, you have a toy and you can combine that with other toys and actually start to build a game.
Lenny Rachitsky: If people were to listen to this segment of the podcast, they would never guess we’re talking about B2B software and email, which I love. Let’s talk about pricing strategy and your approach to pricing. Another very contrarian approach that you guys took where you charge $30 a month for email that was free, that people don’t need to pay for anywhere. And it’s worked and now a lot of companies are thinking of it this way. You’ve even raised your prices recently. What have you learned about pricing strategy that you think might be helpful to folks?
Rahul Vohra: I always say the same thing when it comes to pricing, which is before you figure out pricing, you must first figure out positioning. Superhuman is the best email tool on the market. We fortunately have the metrics to show this. One of the cool things about selling an email tool, is you can compare the 30 days prior to using Superhuman to the 30 days after, or the year before to the year after. We do that obviously. We’re able to show that people get through their email twice as fast with Superhuman, that they respond one to two days faster, and that they save four hours or more every single week. Because of that, we’re very confident in saying that Superhuman is the best email tool on the market and that we’re building it for high performing teams and high performing individuals. In other words, we serve the high end of the market.
Once you understand your positioning, you can then move on to pricing. And one of the best books on this is a book called Monetizing Innovation by Madhavan Ramanujam. And Madhavan covers a lot of ways to develop pricing. We used one of the easiest methods, which is the Van Westendorp Price Sensitivity [inaudible 01:07:08]. In the early years, we asked, I think it was around a hundred of our earliest users, the following four questions. Number one, at what price would you consider Superhuman to be so expensive that you would not consider buying it? Number two, at what price would you consider Superhuman to be priced so low that you’d be worried about its quality and you wouldn’t buy it? At number three, what price would you consider Superhuman to be starting to get expensive, so that it’s not out of the question, but you’d have to give some thought to buying it? And number four, at what price would you consider Superhuman to be a bargain? A great buy for the money?,
Now most startups orient around price point number four. This is especially true for greenfield opportunities, marketplaces, you’ve got to set the transaction value around price 0.4. Basically when you want as many people to sign up as is humanly possible, at the top of the funnel. But the price point that supports our best in class, best in category position, is actually the third one. It starts to feel expensive, but then you sit down and you think about the time that you spend in email, the ROI, and you still buy it anyway. It turns out that the median answer for the third question was $30 per month, and that’s how we picked our price.
And once we picked our price, we then do a quick gut check on market size. For example, we’re a venture scale company, but at the time the question that we had to ask is, “Could we grow into a billion dollar valuation?” Well, let’s assume that at that point our valuation is 10 times our ARR, so our ARR would have to be a hundred million dollars. Well, that would be 300,000 subscribers at $30 per month. That is conservatively assuming no other ways to increase ARP. You mentioned price increase, you can also go up market, you can sell new products and so on. We asked ourselves, without those tricks, do we think we can get to hundreds of thousands of subscribers? And we answered emphatically, yes, so we went ahead with that price.
Lenny Rachitsky: Okay, there’s a couple more things I want to chat about in the time that we have and then I know you have to run. One is around AI and the work you guys are doing there. I know that’s been a big unlock. And then two, the stuff you’re doing in the enterprise. Then if we have time, there’s a question I want to ask that I think is a really interesting way you guys operate.
Let’s talk about AI first. It feels like there’s this being in the right place at the right time. It feels like you guys have been building this for a while, and then AI just unlocked another stage in what you’re able to do with email. Just talk about what you’ve done and what how you think about AI integrating into what you’re doing, how it’s enabled you to kind of take off again?
Rahul Vohra: It’s true that sometimes startups boil down to being in the right place at the right time. We actually had a massive AI launch recently about two weeks ago, but even before then we had multiple flagship AI features. Our first AI feature was write with AI, jot down a few words and we’ll turn them into a fully written email. We actually match the voice and tone in the emails you’ve already sent. So unlike Co-pilot, unlike Gemini, unlike basically every other email app, the email sounds like you. This AI feature is way more popular than I expected it to be. On average today, users are using it 37 times per week.
Number two, our next AI feature was auto summarize, which shows a one line summary above every conversation. And as new emails arrive, it updates instantly. Again, unlike Co-pilot and Gemini, it’s pre-computed. One of the things we do is, we go above and beyond to make these features really premium and feel amazing. The next AI feature after that was instant reply. Imagine waking up to an inbox where every email already has a draft reply. You would simply edit and then send, and sometimes you wouldn’t even need to edit. I can share because we just finished this analysis, that over 2024, the percentage of emails that are AI written and sent with Superhuman has grown four times just in one year.
Then if I remember correctly, the feature after that was Ask AI. Email of course, is this treasure trove of critical information, things like project statuses, customer communication, meeting updates, deal execution, and so much more. And for over 40 years we’ve had to rely on what we hilariously call, search. You have to remember senders, guess keywords, scan subject lines, and now you can just ask, “Where is the queue one offsite?” or, “What are my flight details?” Or, “What is the top five most positive customer responses to the Ask AI launch?” A task by the way, which previously used to take me 20 or 30 minutes to read through all the emails and then create that report now happening in less than five seconds.
Recently we, like I said, announced our biggest evolution yet. Superhuman AI is constantly helping you. It’s organizing your inbox. It’s also making sure you never drop the ball. We have things that we call Auto Labels. You can now write a short prompt like job applications or requests to review work, and you can then immediately see when emails match that prompt, when people apply for a job or they ask you to review work. With Auto Reminders, if your email needs a response, Superhuman will now automatically set a reminder. You don’t have to remember to do that and you’ll never drop the ball again. All you need to do is hit send. With Auto Drafts, Superhuman will now automatically draft your follow-up emails for you and will soon be drafting replies to basically every email that needs a response.
And finally, with what we call Workflows, you can now turn email into repeatable automated workflows. For example, I often get emails from people who are interested in working at Superhuman, and I would normally reply to that candidate and I would let them know that the team will take a look. I’ll then forward to the original message, including any resume or any letter to our head of people and operations and ask her to reach out, if interested. With Workflows, I can now automate this entire process. It’s, you can imagine, creating a little flowchart of what has to happen. Not only does that save a huge amount of time, with Workflows you don’t even have to be in your inbox. In fact, you don’t even have to be working. You could be on vacation while Superhuman AI is working for you.
Lenny Rachitsky: This sounds like product market fit to me. This all sounds wonderful. It just makes sense. This is the stuff we’ve been promised, our underwater cities and flying cars and then just email that just works magically and replies for us and all these things. I love all these things you’re doing.
For folks that are building with AI. I’m curious, what’s maybe been the biggest surprise, either good or bad, building so deeply on top of AI models that you think might be helpful for folks to just, “Watch out for this,” or, “Hey, check this out.”?
Rahul Vohra: I think for me the biggest surprise has been how unpredictable the user love has been in terms of what they love and what they don’t love. For example, write with AI. This sounds like a commodity feature and on all surface level it is. Every email app, every writing surface has a write with AI feature in. I would wager ours is the best at emails and surprisingly that’s what we do. But the surprising thing was just how much people love it and how often it gets used. 37 times per user per week is still mind-blowing to me. I had not expected that, so that’s the most surprising thing.
And on the flip side, there were certain AI features where I did expect a ton of usage, but we didn’t quite get the usage that we were perhaps hoping for. Hopefully I’m not AI Kramer, but basically everything I thought would work out well, people use it less than they thought they did. And everything where I was like, “I don’t know, but let’s build the thing,” people love that.
Lenny Rachitsky: Interesting.
Rahul Vohra: Maybe I should just create an anti-me to do AI road-mapping.
Lenny Rachitsky: That’s in a simple agent right there, whatever Rahul says, do the opposite.
Rahul Vohra: Yeah.
Lenny Rachitsky: Okay. Another maybe a last topic. I know that you guys are starting to move into the enterprise. When people think of Superhuman, they think of it’s consumer-y, it’s for people, and you guys are doing a lot of work to make it a B2B enterprise product. For founders maybe that are starting to think about this transitioning from PLG to sales led and B2B enterprising, what have you learned about just what it takes to get to that point and what does that sales motion look like for you guys?
Rahul Vohra: In some ways, it’s very like selling to prosumers, except these users are not coming from Gmail where prosumers would normally come from, they’re coming from Outlook. And Outlook users have very different expectations to Gmail users. For example, Outlook users expect their email app to also be a fully featured calendar app, whereas Gmail users are happy with those two things being entirely different. As a result, we’ve invested in calendar very heavily and we continue to do so. There’s only so much I can say, but it’s pretty exciting.
Lenny Rachitsky: [inaudible 01:16:49].
Rahul Vohra: Outlook users are also used to certain safeguards, like if you’ve used Outlook in an enterprise, warnings when a recipient is external to your domain or what Outlook users might know as sensitivity labels. And as a result we’ve built support for external recipient indicators and sensitivity labels. But in some ways it’s very different to selling to prosumers because there are other stakeholders involved. For example, we’ve built support for enterprise mobile management by implementing Microsoft Intune.
We recently sold one of the big three strategy consulting firms, which is super exciting. I can’t say which one, but they love Superhuman and they have thousands of people internally using Superhuman. This is after a year… They’ve been piloting for a year and then accelerating over the last few months. We only just got them the mobile app, believe it or not. Because, at an enterprise like that, there are significant controls on what a allowed compliant mobile app can and cannot do. For example, IT needs to be able to control which apps can save attachments or which apps you can copy and paste text into from email. And for many enterprises, those controls are super important.
Lenny Rachitsky: Wow, okay. So it sounds like essentially just building all these features that large companies need, is kind of the road you’re on right now.
Rahul Vohra: Exactly. And there’s two stakeholders. There’s the users, which are actually quite different because they’re Outlook users and Calendar is one of the main ways that manifests. There’s a whole bunch of other stakeholders, IT is one of them, but there are others as well. For example, companies this large have workplace management groups who want to see analytics of how people are working, how they can make their teams more efficient, so it truly is a multi-threaded sale with multiple stakeholders.
Lenny Rachitsky: They had a product from Linear on the podcast [inaudible 01:18:51], and he actually, I don’t know if you heard that episode, but he talks about how they decide what to prioritize, the thing they never build is middle managers needing to track how their reports are doing and things like that. That’s an interesting opportunity for you guys maybe to cut stuff. I don’t know.
Anyway, I want to end on one more nugget. Okay, I’m glad we have time for this. You shared that you have this system internally at Superhuman for making decisions. You call it Single Decisive Reason, SDR. What is that?
Rahul Vohra: SDR is a thinking tool that I picked up from Reid Hoffman during my time at LinkedIn. The idea here is that for important decisions, you should be able to identify one, one reason that on its own supports the decision. It’s based on the observation that all too often we rely on a collection of weak reasons to justify decisions. It’s very, very easy to do this. Imagine you are contemplating a decision, you write a list of the pros and the cons. There are three pros, but let’s say there are 10 or 15 cons. The sheer number of cons, the effort of thinking them through, the time it took to write them down, is going to affect you, consciously or worse subconsciously. This is especially true, I’ve seen in group settings, which just in general are a little bit more risk averse and a little bit more consensus driven.
So whenever anyone is making a decision and they bring that decision to me and they say, “Well, we want to do this because of X, Y, Z, and there are multiple reasons.” I ask them, “What’s the SDR? What’s the single decisive reason?” If they can’t yet isolate it, that tells me they haven’t yet figured out why they want to make the decision. It doesn’t mean the decision is wrong, it just means that they haven’t figured out the singular reason why we should do the thing. They can then go through their list of reasons and ask, “Is this alone enough to support this decision?” Meaning if this was true and all the other things were not true, would I still do it? And sometimes we still do, but actually sometimes we don’t. We realize that a collection of weak reasons alone means that, for example, the outcome is less likely than we thought it was, or it was hiding a really strong reason on the other side of the decision.
Lenny Rachitsky: That is very cool. This is just when someone comes to you with a decision, the way you use this idea is, you ask them what’s the single decisive reason?
Rahul Vohra: Pretty much. Yeah. And what they can’t do, obviously this happens, people are human and natural, they’ll usually start mentioning three or four things, and that’s fine. And then I will say, “Okay, but if only one of those was true and you’re still advocating for this decision, what is it?” I think that’s just a bar for a good decision.
Lenny Rachitsky: Why is that so important? Because you found that a bunch of low quality reasons just don’t add up to a good reason to do something?
Rahul Vohra: Multiple reasons, which is ironic. But that’s my SDR for why SDRs work. Which is yes, multiple low quality reasons rarely add up to a high quality reason to do something. But there are also other things as well, which is, any decision you take has an opportunity cost. Any feature you build is another feature that you didn’t build. If we’re going to build this for a collection of weak reasons, whereas we could build that for one strong reason, I’d much rather build that for one strong reason. Now this is all other things being equal, and these things often end up being quite complicated, but you can apply SDR all the way down. You just did that to me, what’s my SDR for SDR?
Lenny Rachitsky: There we go. Rahul, is there anything that we haven’t covered that you wanted to cover? Is there any last piece of wisdom you want to leave listeners with before we let you go?
Rahul Vohra: I feel good. I think we covered a lot. Thank you for asking amazing questions. This was really fun.
Lenny Rachitsky: This was incredible. Okay, so let me just ask you this then. Where can folks find you online? Where can they check out Superhuman? What should they know before they try it out? And then just how can listeners be useful to you?
Rahul Vohra: If you want to find me online, I am generally on X. That is x.com/rahulvohra, R-A-H-U-L V-O-H-R-A. My DMs are open, so feel free to ping me. If you’re going to do that, I would suggest also emailing me, that’s rahul@superhuman.com, and hopefully I’ll see your message soon.
If you haven’t tried Superhuman, then gosh, what are you doing? This is my call to you to do so, because your time is worth more than whatever you think it might be. So go download Superhuman and give it a shot. Invite your team. The metrics are real. I know they sound like the kind of metrics that startups make up, but getting through your email twice as fast, responding one to two days sooner, saving four hours or more every single week, they’re all real.
Actually, speaking of which, the consulting firm I mentioned earlier, because they’re so into data and into analysis, they wanted to corroborate those numbers for themselves, and so they did. They ran their own internal case study on Superhuman, and they were like, “Yeah, you’re saving our partners 3.3 hours per person per week. And there’s only one other tool that we’ve bought that does that, which is ChatGPT. So thank you. We love Superhuman. We’re rolling it out.” If that sounds interesting to you or your company, please do give it a shot.
Lenny Rachitsky: That is super cool. Reflecting back on what I imagine this conversation would look like, a lot of contrarian thinking and attention to detail, I think that’s exactly what it was. Rahul, you’re awesome. Thank you so much for being here.
Rahul Vohra: Thank you. Bye everyone.
Lenny Rachitsky: Bye everyone.
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. See you in the next episode.
Glossary
| English | 中文 |
|---|---|
| activation | 激活 |
| Adelle Sans | Adelle Sans（字体名称，保留原文） |
| Arial | Arial（字体名称，保留原文） |
| ARPU | ARPU（每用户平均收入，保留原文） |
| ARR | ARR（年度经常性收入，保留原文） |
| Ask AI | 向 AI 提问（Ask AI） |
| Auto Drafts | 自动草稿（Auto Drafts） |
| Auto Labels | 自动标签（Auto Labels） |
| Auto Reminders | 自动提醒（Auto Reminders） |
| auto summarize | 自动摘要（auto summarize） |
| Bill Trenchard | Bill Trenchard（人名，保留原文） |
| Boldstart Ventures | Boldstart Ventures（风投机构名称，保留原文） |
| Chrome | Chrome（Google 浏览器产品名称，保留原文） |
| Comic Sans | Comic Sans（字体名称，保留原文） |
| Comrade | Comrade（人名，保留原文） |
| concierge onboarding | 专人引导（concierge onboarding） |
| Copilot | Copilot（微软 AI 助手产品名称，保留原文） |
| demo day | demo day（创业孵化器展示日，保留原文） |
| Dropbox | Dropbox（公司名称，保留原文） |
| EA (Executive Assistant) | EA（Executive Assistant，行政助理） |
| Ed Sim | Ed Sim（人名，保留原文） |
| Elliot Durbin | Elliot Durbin（人名，保留原文） |
| enterprise mobile management | 企业移动设备管理（enterprise mobile management） |
| external recipient indicators | 外部收件人指示器（external recipient indicators） |
| first principles | 第一性原理 |
| First Round Capital | First Round Capital（风投机构名称，保留原文） |
| first-time user experience | 首次用户体验 |
| Gemini | Gemini（Google AI 产品名称，保留原文） |
| Google（公司名称，保留原文） | |
| Growth | 增长（指公司增长团队/职能） |
| growth hacker | 增长黑客 |
| growth loops | 增长循环 |
| instant reply | 即时回复（instant reply） |
| Kramer | Kramer（人名/角色名，指电视剧《宋飞正传》中总做错误判断的角色，保留原文） |
| Laurent Valasek | Laurent Valasek（保留原文） |
| Lyft | Lyft（公司名称，保留原文） |
| Madhavan Ramanujam | Madhavan Ramanujam（人名，保留原文） |
| Mailbox | Mailbox（公司名称，保留原文） |
| market widening | 市场拓展 |
| Microsoft Intune | Microsoft Intune（企业移动设备管理平台，保留原文） |
| Monetizing Innovation | 《Monetizing Innovation》（书名，保留原文） |
| MPS | MPS（用户满意度指标，保留原文） |
| net promoter score | 净推荐值（net promoter score） |
| Outlook | Outlook（微软邮件客户端产品名称，保留原文） |
| Peak Leadership Institute | Peak Leadership Institute（保留原文） |
| PLG | PLG（Product-Led Growth，产品驱动增长，保留原文） |
| Positioning: The Battle for Your Mind | 《定位：争夺用户心智的战争》（Positioning: The Battle for Your Mind，营销学经典著作） |
| Product Market Fit Engine | 产品市场契合度引擎（Product Market Fit Engine） |
| Product-Market Fit | 产品市场契合度 |
| prosumer | 高级个人用户（prosumer） |
| Raja | 王瑜伽（Raja）传统 |
| Rapportive | Rapportive（公司名称，保留原文） |
| Reid Hoffman | Reid Hoffman（人名，保留原文） |
| SDR (Single Decisive Reason) | SDR（Single Decisive Reason，单一决定性理由） |
| Sean Ellis | Sean Ellis（人名，创造了”增长黑客”一词） |
| sensitivity labels | 敏感度标签（sensitivity labels） |
| solution deepening | 方案深化 |
| Superhuman for Sales | Superhuman for Sales（产品名称，保留原文） |
| Switch Lock | （Rahul 自创的技巧名称，保留原文） |
| Switch Log | Switch Log（Rahul 自创的时间追踪技巧名称，保留原文） |
| technology moat | 技术护城河 |
| transcendental meditation | 超觉冥想 |
| Type Together | Type Together（字体设计工作室名称，保留原文） |
| Uber | Uber（公司名称，保留原文） |
| Van Westendorp Price Sensitivity | Van Westendorp 价格敏感度测试 |
| virality | 病毒式传播 |
| Vivek | Vivek（人名，保留原文） |
| Workflows | 工作流（Workflows） |
| workplace management groups | 职场管理团队（workplace management groups） |
| write with AI | 用 AI 写（write with AI） |
| Y Combinator | Y Combinator（机构名称，保留原文） |
| Yogic | 瑜伽（Yogic）传统 |
| zero interest rate phenomenon | 零利率现象 |
| zone of genius | 天才区 |
Reformatted by reformat_english.py