返回列表
🧠 阿头学 · 💬 讨论题

如何打造让用户上瘾的产品

Wispr Flow 创始人的消费级产品三板斧——心理学、产品、团队——本质上是"一次只改变一个行为"的极致纪律。

2026-02-13 原文链接 ↗
阅读简报
双语对照
完整翻译
原文
讨论归档

核心观点

  • 一次只改变一个行为 这是全文最贵的一课。Humane Pin 同时要求用户戴新设备+用手掌投影交互，死了；Google Glass 同时要求戴摄像头+对空气说话，死了。Wispr 只要求一件事：用说的代替打字，其他一切不变。90%增长来自口碑，不是运气，是行为设计的结果。
  • 跟500个真人聊，不是发问卷 创始人亲自坐下来看用户用竞品时在哪一秒崩溃，所有设计决策都追溯到某次真实对话。大多数创始人造自己以为用户想要的东西，而不是用户真正痛的东西。
  • 产品要"学会"用户 个人词典、语气风格、自我纠错——这些不是花哨功能，是在告诉用户"我在乎你"。用户感受到被在乎，就会在乎你的产品。
  • 2人小队 + 招前创始人 200人团队仍然用2人pod制：最大ownership，最小沟通开销。大量招前创始人，因为他们有"3点钟不修好这个bug睡不着觉"的本能，这种ownership无法面试出来。
  • 招品味，不只招技能 消费级产品里，那个说"动画慢了50ms"但说不出为什么的人，比什么都值钱。

跟我们的关联

Neta 现在 DAU 10万+，跟 Wispr Flow 的 10万+ DAU 处于同一量级，但路径完全不同——我们是AI社交，他们是AI工具。几个直接碰撞点：

1. "一次只改变一个行为"原则：Neta 让用户跟AI社交，这本身就是一个巨大的行为改变。我们是否在某些功能上无意中叠加了第二个行为改变？比如同时要求用户学新交互方式+接受AI人格，需要审视。 2. 2人pod制 vs 我们的20人特种作战阵型：我们团队更小，但理念高度一致——极致ownership、最小沟通成本。这篇文章佐证了我们的编制逻辑。值得检查：我们的pod划分是否足够清晰，每个pod是否真的"full context, full accountability"？ 3. 招前创始人的策略：20人团队每一个人都是关键棋子，"招品味+招ownership"这个标准直接可以写进我们的招聘checklist。海外扩张需要的人更得是这种profile。 4. 500次真人对话：我们做海外产品时，有没有对海外目标用户做到这种深度的行为观察？还是在用国内经验推演海外需求？

讨论引子

  • Neta 要求用户做的"行为改变"到底是几个？如果超过一个，哪个可以砍掉或者藏起来，让用户先只改变一个？
  • 我们20人阵型里，2人pod制能直接搬过来用吗？现在的分工方式跟这个差距在哪？
  • 做海外产品之前，我们打算跟多少个海外真实用户"坐下来聊"？500这个数字对我们意味着什么？

如何打造让人停不下来的产品

99% 的消费级产品都因同一个原因而失败。

做出一款人们愿意用的产品是一回事；做出一款人们离不开的产品，则是另一回事。

我最近的一家公司（语音转文字 AI）最近达到了 100,000+ 的并发日活用户，累计口述超过 100 亿个词。

打造一家成功的消费级公司，比你想象的要简单得多。

我们常常过度纠结于某些细枝末节（比如微妙的设计修补或新的营销活动），但真正驱动增长的因素其实要明显得多。

这篇文章想把我一路走来学到的一切分享给你。

背景

以我的经验，真正“起作用”的东西都可以归为三类：

心理学——理解是什么让一个人一天用你的产品 100 次，而不是只用一次。

产品——那些看不见的决定，如何把一个 demo 变成一种依赖。

团队——让 5 个人的交付速度超过 500 人公司的人员协作机制。

大多数创业建议会在这三类里偏重其一；但实际上，它们是同一个系统的三根支柱。

以下是完整打法：

I - 心理学

要做出任何成功的产品，你不能只是在工程上堆叠不同的用例，或把用户流程设计得无比精巧。你必须从底层出发，反向拆解他们的行为。

消费级产品不是靠功能赢，而是靠感受赢。而这两者之间的差距，正是大多数公司失败的地方。

沿着这个思路最重要的领悟是：一次只能改变一种行为。

这是消费产品里最昂贵的一课。

在 Wispr Flow 作为软件出现之前，我们在做的是一款硬件设备——一只可穿戴耳机，能够读取无声发声时的神经信号，并将其转成文字。

硬件运行得非常完美，我们甚至能把“默念”的想法转换成文本。

但它要求用户同时养成两种新行为：戴上一个新设备，并且用对着电脑说话来代替打字。

回头看看过去 5 年里每一个失败的硬件产品：Humane AI Pin 要求人们戴上一个新设备，并在手掌上的投影界面进行交互；Google Glass 要求人们把摄像头戴在脸上，并在公共场合对着空气说话。它们都在同时要求用户改变两种行为。

如果 Humane 当时只是做一个用于情境化 AI 的 App，或许就会有数百万人采用。硬件应该是一种升级，而不是一场信仰式的豪赌。

我们做 Wispr Flow 时，只要求一种行为改变：用说话替代打字。其他一切——你的应用、你的工作流程、你的屏幕——都保持不变。

正是这套无缝工作流，解释了为什么我们 90% 的增长来自口碑。这不是运气，而是我们死磕一个心理瞬间，并围绕它把一切工程化的结果。

II - 产品

有两个关键动作，让我们真正解决了用户的问题。

我们和 500+ 个人聊过。

在 Flow 诞生之前，我亲自与 500 多个人坐下来，了解他们到底卡在什么地方。不是问卷。不是 Google Forms。而是真实的对话：看着人们在市面上每一款听写工具上挣扎，记录挫败感击中的确切瞬间。

大多数创始人会去做他们以为用户想要的东西。可在那几百次对话之后，规律清晰得无法否认。Flow 的每一个设计决定，都能追溯到其中某次对话里某个人说过的话。

一个会学习你的产品

除了先进的听写能力与 AI 模型之外，我们还打造了一个能直接学习你正在做什么、以及你独特工作方式的产品。市面上大多数工具对某些名字或词总会拼错。我们刻意让用户看到：我们在持续学习与改进，尽量避免同样的错误再次发生。以下是一些我们交付的关键功能示例：

个人词典。Flow 把某个词弄错了，你改正后，Flow 会记住正确拼写。

语气风格。你可以根据自己喜欢的写作方式，主动设置合适的语气、标点和大小写风格。

自我纠错。当你说“我们明天见——不对，改成周五”，Flow 会输出“我们周五见”。

除了让产品在用户感受上无缝顺滑之外，这些独特的改变也在传递一件事：我们真正在意用户体验。这也是 Wispr 成功的关键因素之一——用户面对的问题是实打实的困扰，而不是被当作“额外工单”能拖就拖。

你对用户用心，用户就会对你的产品上心。

III - 团队

世界上最好的产品，如果由不对的人来做，也没有意义。

如今我们团队大约 200 人。相较其他竞争对手，我们仍然很小，但我们的人均产出高得夸张。原因在这：

用 2 人小组来组织结构。

每个项目两名工程师 -> 最大化主人翁意识。最小化沟通成本。

当一个小组拥有某件事时，就没有模糊地带：不用开会讨论谁负责，也不会有 Jira 工单在 6 个人之间来回传递。两个人，完整上下文，完整责任。

这就是你在规模增长时依然能保持创业速度的方法。

你每多加一个人，沟通开销都会以指数级增长。

  1. 招前创始人，而不是前员工

我们团队里很大一部分人都是前创始人。

并不是因为创过业是一枚荣誉徽章，而是因为创始人有一种几乎无法通过面试筛出来的特质：他们会主动把问题解决掉，因为他们知道不会有人替他们修。

在大公司待过的前员工，往往习惯于看见一个 bug 就提个工单；而前创始人看见一个 bug，会熬到凌晨 3 点把它修好，因为他们在明知它坏着的情况下就是睡不着。

这种主人翁本能，就是“会持续迭代的团队”和“等指令的团队”之间的差别。

  1. 招看重品味，而不只看技能

在消费级产品里，品味就是一切。

所谓品味，是那种看着 200ms 的动效就说“这慢了 50ms”的人——他解释不出为什么，但他是对的。是那个会拒绝一种“技术上正确但情绪上不对”的颜色的设计师。是那个会因为心里别扭而重构一段永远没人看见的代码的工程师。

在招聘 Wispr 团队的每一位成员时，我们都遵循这些原则。

看起来这些要求很具体，但它们才是真正推动指标、让我们以极低开销高速前进的关键因素。

你在挑选顶尖团队上投入越多时间，之后就越不需要花时间去培训他们。

结语

打造一家能赢的消费级公司没有固定公式。但如果一定要把我见过的一切压缩成一句原则，那就是：

别做“人们想要”的东西，要做“人们离不开”的东西。

我们赢，不是因为我们的 AI 更好。我们赢，也不是因为我们更有钱。

我们做的是一款人们当初并不知道自己需要的产品。可一旦它到了他们手里，就成了工作流程的一部分，并创造了一个新习惯。

如果你在做消费级产品，把这篇文章收藏起来，问问自己是否走过了这三条原则。

希望这能帮你更清楚地理解：什么叫做打造一款能赢的产品。若你还有其他问题，欢迎在评论区留言。

祝你打造顺利 :)

链接：http://x.com/i/article/2020925806546124803

相关笔记

99% of consumer products fail for the same reason.

99% 的消费级产品都因同一个原因而失败。

It's one thing to build a product people will use, but it's another to build something people can't live without.

做出一款人们愿意用的产品是一回事；做出一款人们离不开的产品，则是另一回事。

My most recent company (voice-to-text AI) recently reached 100,000+ concurrent daily active users and over 10 billion words dictated.

我最近的一家公司（语音转文字 AI）最近达到了 100,000+ 的并发日活用户，累计口述超过 100 亿个词。

Building a successful consumer company is much more simple than you think.

打造一家成功的消费级公司，比你想象的要简单得多。

We often hyper-fixate on the intricacies of specific areas (subtle design fixes or new marketing campaigns), but the real drivers of growth are much more obvious.

我们常常过度纠结于某些细枝末节（比如微妙的设计修补或新的营销活动），但真正驱动增长的因素其实要明显得多。

This article is meant to teach you everything I learned along the way.

这篇文章想把我一路走来学到的一切分享给你。

Context

背景

Based on my experience, everything that actually moved the needle falls into three buckets:

以我的经验，真正“起作用”的东西都可以归为三类：

Psychology - understanding what makes someone use your product 100 times a day instead of once.

心理学——理解是什么让一个人一天用你的产品 100 次，而不是只用一次。

Product - the invisible decisions that turn a demo into a dependency.

产品——那些看不见的决定，如何把一个 demo 变成一种依赖。

Team - the people dynamics that let 5 people out-ship companies with 500.

团队——让 5 个人的交付速度超过 500 人公司的人员协作机制。

Most startup advice over-indexes on one of these categories. In reality, they're three pillars of one system.

大多数创业建议会在这三类里偏重其一；但实际上，它们是同一个系统的三根支柱。

Here is the full playbook:

以下是完整打法：

I - Psychology

I - 心理学

To build any successful product, you can't just engineer different use cases or make intricate designs for the user flow. You have to reverse engineer their behavior from the ground up.

要做出任何成功的产品，你不能只是在工程上堆叠不同的用例，或把用户流程设计得无比精巧。你必须从底层出发，反向拆解他们的行为。

Consumer products don't win on features. They win on feelings. And the gap between those two things is where most companies die.

消费级产品不是靠功能赢，而是靠感受赢。而这两者之间的差距，正是大多数公司失败的地方。

The most important realization following this idea is this: You can only change one behavior at a time.

沿着这个思路最重要的领悟是：一次只能改变一种行为。

This is the most expensive lesson in consumer.

这是消费产品里最昂贵的一课。

Before Wispr Flow existed as software, we were building a hardware device - a wearable earpiece that could read neural signals from silent speech and turn them into text.

在 Wispr Flow 作为软件出现之前，我们在做的是一款硬件设备——一只可穿戴耳机，能够读取无声发声时的神经信号，并将其转成文字。

The hardware worked perfectly, and we could convert subvocalized thoughts into text.

硬件运行得非常完美，我们甚至能把“默念”的想法转换成文本。

But it required users to adopt two new behaviors simultaneously: wear a new device AND speak to their computer instead of typing.

但它要求用户同时养成两种新行为：戴上一个新设备，并且用对着电脑说话来代替打字。

Look at every failed hardware product in the last 5 years. The Humane AI Pin asked people to wear a new device AND interact with a projector on their palm. Google Glass asked people to wear a camera on their face AND talk to thin air in public. They were asking for two behavior changes at once.

回头看看过去 5 年里每一个失败的硬件产品：Humane AI Pin 要求人们戴上一个新设备，并在手掌上的投影界面进行交互；Google Glass 要求人们把摄像头戴在脸上，并在公共场合对着空气说话。它们都在同时要求用户改变两种行为。

If Humane had just made an app for contextual AI use, millions might have adopted it. Hardware should be an upgrade, not a leap of faith.

如果 Humane 当时只是做一个用于情境化 AI 的 App，或许就会有数百万人采用。硬件应该是一种升级，而不是一场信仰式的豪赌。

When we made Wispr Flow, we asked for exactly one behavior change: speak instead of type. Everything else - your apps, your workflow, your screen - stays identical.

我们做 Wispr Flow 时，只要求一种行为改变：用说话替代打字。其他一切——你的应用、你的工作流程、你的屏幕——都保持不变。

Creating this seamless workflow explains why 90% of our growth is from word of mouth. It's not luck. That's the result of obsessing over one psychological moment and engineering everything around it.

正是这套无缝工作流，解释了为什么我们 90% 的增长来自口碑。这不是运气，而是我们死磕一个心理瞬间，并围绕它把一切工程化的结果。

II - Product

II - 产品

There were two key actions that enabled us to truly solve the user's problems.

有两个关键动作，让我们真正解决了用户的问题。

We talked to 500+ people.

我们和 500+ 个人聊过。

Before Flow existed, I personally sat down with over 500 people to learn more about what they struggled with. Not surveys. Not Google Forms. Real conversations, watching people struggle with every dictation tool on the market, noting the exact moment frustration hit.

在 Flow 诞生之前，我亲自与 500 多个人坐下来，了解他们到底卡在什么地方。不是问卷。不是 Google Forms。而是真实的对话：看着人们在市面上每一款听写工具上挣扎，记录挫败感击中的确切瞬间。

Most founders build what they think users want. But instead, those hundreds of conversations later, the patterns were undeniable. Every design decision in Flow traces back to something someone said in one of those conversations.

大多数创始人会去做他们以为用户想要的东西。可在那几百次对话之后，规律清晰得无法否认。Flow 的每一个设计决定，都能追溯到其中某次对话里某个人说过的话。

A product that learns you

一个会学习你的产品

Besides the advanced dictation and AI models, we also built a product that learns directly what you're working on and your unique style of work. for most tools on the market, certain names or words are always misspelled. We made an effort to show we are constantly learning and improving so that doesn't happen again. Here are some examples of key features we shipped:

除了先进的听写能力与 AI 模型之外，我们还打造了一个能直接学习你正在做什么、以及你独特工作方式的产品。市面上大多数工具对某些名字或词总会拼错。我们刻意让用户看到：我们在持续学习与改进，尽量避免同样的错误再次发生。以下是一些我们交付的关键功能示例：

Personal dictionary. If Flow got a word wrong and you fixed it, Flow remembers the correct spelling.

个人词典。Flow 把某个词弄错了，你改正后，Flow 会记住正确拼写。

Tone of voice. You can actively set the correct tone, punctuation, and style of capitalization based on how you enjoy writing.

语气风格。你可以根据自己喜欢的写作方式，主动设置合适的语气、标点和大小写风格。

Self-correction. When you say "Let's meet tomorrow, no wait, Friday instead," Flow outputs "Let's meet Friday."

自我纠错。当你说“我们明天见——不对，改成周五”，Flow 会输出“我们周五见”。

Beyond building a product that feels seamless to the user, these unique changes also show that we genuinely cared about the user's experience. This is one of the key factors that makes Wispr so successful: the problems users face are genuine concerns not extra tickets that are pushed off until as late as possible.

除了让产品在用户感受上无缝顺滑之外，这些独特的改变也在传递一件事：我们真正在意用户体验。这也是 Wispr 成功的关键因素之一——用户面对的问题是实打实的困扰，而不是被当作“额外工单”能拖就拖。

If you show care for your users, users will care for your product.

你对用户用心，用户就会对你的产品上心。

III - Team

III - 团队

The best product in the world doesn't matter if the wrong people are building it.

世界上最好的产品，如果由不对的人来做，也没有意义。

Today we're around 200 person team. We're still tiny relative to other competitors, but our output per person is absurd. Here's why:

如今我们团队大约 200 人。相较其他竞争对手，我们仍然很小，但我们的人均产出高得夸张。原因在这：

Structure by having 2-person pods.

用 2 人小组来组织结构。

Two engineers per project -> Maximum ownership. Minimum communication overhead.

每个项目两名工程师 -> 最大化主人翁意识。最小化沟通成本。

When a pod owns something, there's no ambiguity. No meetings about who's responsible. No Jira tickets passing between 6 people. Two people, full context, full accountability.

当一个小组拥有某件事时，就没有模糊地带：不用开会讨论谁负责，也不会有 Jira 工单在 6 个人之间来回传递。两个人，完整上下文，完整责任。

This is how you move at startup speed even as you grow.

这就是你在规模增长时依然能保持创业速度的方法。

Every additional person you add creates communication overhead that grows exponentially.

你每多加一个人，沟通开销都会以指数级增长。

  1. Hire ex-founders, not ex-employees
  1. 招前创始人，而不是前员工

A huge percentage of our team are former founders.

我们团队里很大一部分人都是前创始人。

Not because founding a company is a badge of honor. Because founders have a specific trait that's almost impossible to interview for: they can take the initiative to solve problems and no one else is going to fix it.

并不是因为创过业是一枚荣誉徽章，而是因为创始人有一种几乎无法通过面试筛出来的特质：他们会主动把问题解决掉，因为他们知道不会有人替他们修。

An ex-employee at a big company is conditioned to see a bug and file a ticket. An ex-founder sees a bug and stays up until 3am fixing it because they physically cannot go to sleep knowing it's broken.

在大公司待过的前员工，往往习惯于看见一个 bug 就提个工单；而前创始人看见一个 bug，会熬到凌晨 3 点把它修好，因为他们在明知它坏着的情况下就是睡不着。

That ownership instinct is the difference between a team that iterates and a team that waits for instructions.

这种主人翁本能，就是“会持续迭代的团队”和“等指令的团队”之间的差别。

  1. Hire for taste, not just skill
  1. 招看重品味，而不只看技能

In consumer products, taste is everything.

在消费级产品里，品味就是一切。

Taste is the person who looks at a 200ms animation and says "that's 50ms too slow" and can't explain how they know but they're right. It's the designer who rejects a color that's technically correct but emotionally wrong. It's the engineer who refactors code nobody will ever see because it bothers them.

所谓品味，是那种看着 200ms 的动效就说“这慢了 50ms”的人——他解释不出为什么，但他是对的。是那个会拒绝一种“技术上正确但情绪上不对”的颜色的设计师。是那个会因为心里别扭而重构一段永远没人看见的代码的工程师。

These are the principles that we live by when it comes to hiring each member of the Wispr Team.

在招聘 Wispr 团队的每一位成员时，我们都遵循这些原则。

While it may seem specific, these are the key factors that truly move the meter and enable us to move fast with little overhead.

看起来这些要求很具体，但它们才是真正推动指标、让我们以极低开销高速前进的关键因素。

The more time you spend choosing an elite team, the less time you'll need training them later on.

你在挑选顶尖团队上投入越多时间，之后就越不需要花时间去培训他们。

Closing Notes

结语

Building a winning consumer company isn't a formula. But if I had to compress everything I've seen into one principle, it's this:

打造一家能赢的消费级公司没有固定公式。但如果一定要把我见过的一切压缩成一句原则，那就是：

Don't build something people want, build something people can't live without.

别做“人们想要”的东西，要做“人们离不开”的东西。

We didn't win because we had better AI. We didn't win because we had more money.

我们赢，不是因为我们的 AI 更好。我们赢，也不是因为我们更有钱。

We built a product that people didn't know they needed. Once it was in their hands, it became part of their workflow and created a new habit.

我们做的是一款人们当初并不知道自己需要的产品。可一旦它到了他们手里，就成了工作流程的一部分，并创造了一个新习惯。

If you're building in consumer, save this article and ask yourself if you've gone through the three principles.

如果你在做消费级产品，把这篇文章收藏起来，问问自己是否走过了这三条原则。

I hope this helps you gain some clarity on what it means to build a winning product, and if you have any other questions, feel free to drop it in the comments.

希望这能帮你更清楚地理解：什么叫做打造一款能赢的产品。若你还有其他问题，欢迎在评论区留言。

Happy building :)

祝你打造顺利 :)

Link: http://x.com/i/article/2020925806546124803

链接：http://x.com/i/article/2020925806546124803

相关笔记

How to build a product people can’t stop using

  • Source: https://x.com/tankots/status/2022068761009041902?s=46
  • Mirror: https://x.com/tankots/status/2022068761009041902?s=46
  • Published: 2026-02-12T22:02:15+00:00
  • Saved: 2026-02-13

Content

99% of consumer products fail for the same reason.

It's one thing to build a product people will use, but it's another to build something people can't live without.

My most recent company (voice-to-text AI) recently reached 100,000+ concurrent daily active users and over 10 billion words dictated.

Building a successful consumer company is much more simple than you think.

We often hyper-fixate on the intricacies of specific areas (subtle design fixes or new marketing campaigns), but the real drivers of growth are much more obvious.

This article is meant to teach you everything I learned along the way.

Context

Based on my experience, everything that actually moved the needle falls into three buckets:

Psychology - understanding what makes someone use your product 100 times a day instead of once.

Product - the invisible decisions that turn a demo into a dependency.

Team - the people dynamics that let 5 people out-ship companies with 500.

Most startup advice over-indexes on one of these categories. In reality, they're three pillars of one system.

Here is the full playbook:

I - Psychology

To build any successful product, you can't just engineer different use cases or make intricate designs for the user flow. You have to reverse engineer their behavior from the ground up.

Consumer products don't win on features. They win on feelings. And the gap between those two things is where most companies die.

The most important realization following this idea is this: You can only change one behavior at a time.

This is the most expensive lesson in consumer.

Before Wispr Flow existed as software, we were building a hardware device - a wearable earpiece that could read neural signals from silent speech and turn them into text.

The hardware worked perfectly, and we could convert subvocalized thoughts into text.

But it required users to adopt two new behaviors simultaneously: wear a new device AND speak to their computer instead of typing.

Look at every failed hardware product in the last 5 years. The Humane AI Pin asked people to wear a new device AND interact with a projector on their palm. Google Glass asked people to wear a camera on their face AND talk to thin air in public. They were asking for two behavior changes at once.

If Humane had just made an app for contextual AI use, millions might have adopted it. Hardware should be an upgrade, not a leap of faith.

When we made Wispr Flow, we asked for exactly one behavior change: speak instead of type. Everything else - your apps, your workflow, your screen - stays identical.

Creating this seamless workflow explains why 90% of our growth is from word of mouth. It's not luck. That's the result of obsessing over one psychological moment and engineering everything around it.

II - Product

There were two key actions that enabled us to truly solve the user's problems.

We talked to 500+ people.

Before Flow existed, I personally sat down with over 500 people to learn more about what they struggled with. Not surveys. Not Google Forms. Real conversations, watching people struggle with every dictation tool on the market, noting the exact moment frustration hit.

Most founders build what they think users want. But instead, those hundreds of conversations later, the patterns were undeniable. Every design decision in Flow traces back to something someone said in one of those conversations.

A product that learns you

Besides the advanced dictation and AI models, we also built a product that learns directly what you're working on and your unique style of work. for most tools on the market, certain names or words are always misspelled. We made an effort to show we are constantly learning and improving so that doesn't happen again. Here are some examples of key features we shipped:

Personal dictionary. If Flow got a word wrong and you fixed it, Flow remembers the correct spelling.

Tone of voice. You can actively set the correct tone, punctuation, and style of capitalization based on how you enjoy writing.

Self-correction. When you say "Let's meet tomorrow, no wait, Friday instead," Flow outputs "Let's meet Friday."

Beyond building a product that feels seamless to the user, these unique changes also show that we genuinely cared about the user's experience. This is one of the key factors that makes Wispr so successful: the problems users face are genuine concerns not extra tickets that are pushed off until as late as possible.

If you show care for your users, users will care for your product.

III - Team

The best product in the world doesn't matter if the wrong people are building it.

Today we're around 200 person team. We're still tiny relative to other competitors, but our output per person is absurd. Here's why:

Structure by having 2-person pods.

Two engineers per project -> Maximum ownership. Minimum communication overhead.

When a pod owns something, there's no ambiguity. No meetings about who's responsible. No Jira tickets passing between 6 people. Two people, full context, full accountability.

This is how you move at startup speed even as you grow.

Every additional person you add creates communication overhead that grows exponentially.

  1. Hire ex-founders, not ex-employees

A huge percentage of our team are former founders.

Not because founding a company is a badge of honor. Because founders have a specific trait that's almost impossible to interview for: they can take the initiative to solve problems and no one else is going to fix it.

An ex-employee at a big company is conditioned to see a bug and file a ticket. An ex-founder sees a bug and stays up until 3am fixing it because they physically cannot go to sleep knowing it's broken.

That ownership instinct is the difference between a team that iterates and a team that waits for instructions.

  1. Hire for taste, not just skill

In consumer products, taste is everything.

Taste is the person who looks at a 200ms animation and says "that's 50ms too slow" and can't explain how they know but they're right. It's the designer who rejects a color that's technically correct but emotionally wrong. It's the engineer who refactors code nobody will ever see because it bothers them.

These are the principles that we live by when it comes to hiring each member of the Wispr Team.

While it may seem specific, these are the key factors that truly move the meter and enable us to move fast with little overhead.

The more time you spend choosing an elite team, the less time you'll need training them later on.

Closing Notes

Building a winning consumer company isn't a formula. But if I had to compress everything I've seen into one principle, it's this:

Don't build something people want, build something people can't live without.

We didn't win because we had better AI. We didn't win because we had more money.

We built a product that people didn't know they needed. Once it was in their hands, it became part of their workflow and created a new habit.

If you're building in consumer, save this article and ask yourself if you've gone through the three principles.

I hope this helps you gain some clarity on what it means to build a winning product, and if you have any other questions, feel free to drop it in the comments.

Happy building :)

Link: http://x.com/i/article/2020925806546124803

📋 讨论归档

讨论进行中…