创始人模式：Kashish Gupta，Hightouch创始人兼联合首席执行官

今天我和Carolyn与来自High Touch的Kashish Gupta一起，他于2019年参加了YC。欢迎欢迎。

Carolyn and I are here today with Kashish Gupta of High Touch, who went through YC in 2019. Welcome. Welcome.

谢谢邀请。

Thanks for having me.

真的很高兴见到你。我之前从未见过你。

It's really nice to meet you. I have never met you.

也很高兴见到你。我久仰二位大名了。

Good to meet you as well. I've heard so much about you both.

嗯，谢谢你今天来参加我们的特别系列节目。我们现在在Y Combinator创始人模式静修会上，所以我们想和创始人们聊聊他们在创始人模式方面的经历。但首先，我们很想了解更多关于你们公司是做什么的。

Well, thank you for coming on our special series today. We are here at the Y Combinator Founder Mode Retreat, and so we're talking to founders a little bit about their experiences with Founder Mode. But first, we'd love to hear more about what your company does.

HiTouch是一个面向营销人员的数据和AI平台。我们帮助营销人员做两件事。一是帮助他们访问客户数据，并利用这些数据运行任何广告或生命周期营销活动。二是帮助他们部署AI代理，这些代理可以更智能地运行这些营销活动，并从零开始构建这些活动。我们现在正在构建能够帮助编写高度个性化内容的代理，这样你就可以编写一个与网球活动截然不同的篮球活动，然后将这些内容发送给对这两个主题特别感兴趣的人。

HiTouch is a data and AI platform for marketers. So we help marketers do two things. We help them access their customer data and run any advertising or life cycle marketing campaign using their customer data. And two is we help them deploy AI agents that can help them more intelligently run those marketing campaigns and build those campaigns from scratch. So we're now building agents that can help you write content that's highly personalized for your customers so that you can write a basketball campaign that's very different from your tennis campaign, and then send those to people that are specifically interested in those two topics.

哦，哇。所以我感觉

Oh, wow. So I feel

这对我来说每天都在发生。非常精准的广告。

like that happens to me on the daily. Very targeted advertising.

可能你从我们这里收到了很多这样的营销。可能吧。那些使用high touch的品牌。

Probably getting a lot of that marketing from us. Probably. Brands that use high touch.

是的。是的。你们的一些大客户是谁？

Yeah. Yeah. Who are some of your big clients?

是的。我们主要与几家大型B2C品牌合作，通常是像PetSmart、Spotify、Men's Wearhouse这类公司。

Yeah. So we work with, several, like, large b to c brands primarily, usually companies like PetSmart, Spotify, Men's Wearhouse, these kinds of brands.

哦，哇。

Oh, wow.

而且我们一直专注于企业级客户。早期时，我们对大小公司都做了探索。大家都说卖给小公司更容易。没错。但我们决定不这么做，因为我们发现基于我们正在开发的产品，我们的产品市场契合度明显更适合企业级客户。

And, we've always been enterprise focused. So in the early days, like, we kind of did discovery on smaller and bigger companies. Everyone said sell to smaller companies because it's easier. Yep. And we just decided not to because we found out that our product market fit based on the product that we were building was strictly better for enterprise.

等等。有人告诉你们要卖给小企业？

Wait. People told you to sell to small businesses?

是的。包括Y Combinator也这么建议。

Yeah. Because they Including Y Combinator.

真的？是啊。因为这样销售周期会更短。没错。就像是，先取得一些小胜利，然后逐步升级到大企业。

Really? Yeah. Because this cycle was gonna be shorter. Exactly. So it's like, oh, get a few wins and then, like, graduate up to the big enterprises.

更容易获得反馈。正是这样。

Easier to get feedback. Exactly.

更容易测试、迭代等等。所以你们就直接决定，不，我们要一开始就瞄准大客户。

Easier to test and evolve and all. So then you're like, nope. We're gonna go for the big guns right away.

这是个特殊情况。我得说，对大多数创始人来说，那个建议可能仍然很好。但对我们而言，我们的架构完美适合企业级需求，对小企业来说反而不太合理。

And it was a unique scenario. I would say the feedback is still probably great for most founders. Yeah. But for us, our architecture was one that was perfectly suited for enterprise and didn't make that much sense for smaller businesses.

所以你们就凭直觉拒绝了那个建议。

And so you just went with your gut and said no.

这就是客户告诉我们的。我们基本上只听客户的，不管是谁告诉我们什么。

That's what the customers were telling us. We basically only listen to customers no matter who tells us what.

哇。这。也许这就是创始人模式的第一课。是的。第一课是

Wow. It. Maybe that's sort of founder mode lesson. Yeah. First lesson is

我认为大多数创始人模式的创始人都这样做。是的。就像，如果客户没有说过某件事，那么在客户说出来之前它就不是真的。甚至当我试图向我的联合创始人推销某个想法时，他们会问我，这是客户说的吗，还是只是你的想法，Kashish？如果只是我的想法，那么在客户说出来之前就不算数。

I think most founder mode founders do this. Yeah. Like, if the customer didn't say something, it's just not true until the customer says it. And even when I try to pitch my cofounders on something, they'll tell me, did the customer say that, or is that your thoughts, Kashish? And if it's my idea, then it doesn't count until the customer says it.

哇。你们对以客户为中心真是，怎么说，非常执着。

Wow. You guys are, like, intense about the customer centricness.

是的。有时候这非常严苛。

Yeah. Sometimes it's very harsh.

你们是怎么开始的？

How did you guys get started?

当然。我们最初是一家旅游公司。2019年我们就是带着这个业务加入Y Combinator的。

Sure. So we originally were a travel business. That's what we did Y Combinator with in 2019.

没开玩笑。是的。我们是

No kidding. Yeah. We were

基本上是旅游版的ChatGPT。你只需给我们发消息，我们就会使用当时的自然语言处理技术（在GPT模型出现之前）来为你预订旅行。

basically chat GPT for travel. You would just message us, and we would use natural language processing at the time before GPT models existed to book your travel.

好的。

Okay.

新冠疫情发生了。我们觉得这时候旅行毫无意义。

COVID happened. We decided it didn't make any sense to travel No.

当时正在旅行。

Was traveling.

紧接着第二年也是如此。于是我们开始转型尝试。我们做了很多不同的商业构想，每次都会问客户：你们需要这个产品吗？他们总是说需要。然后我们总会追问其他问题，比如：除了这个，你们最优先考虑的是什么？

Following exactly in the following year. And so then we were pivoting. We did a bunch of different business ideas and every single time we asked our customers, do you want this product? They always said yes. Then we always ask them other questions like, well, what are your top priorities other than this?

这时他们就会突然放松下来，告诉我们所有重要的优先事项——而这些往往与我们

And then they would just suddenly relax and tell us all their important priorities that were not what we

谈论的无关。你说客户时，我想到的是做旅游的人。在这个场景里，你说的客户具体指哪些人？

were talking When you say customers, like, I'm thinking, like, these were people doing travel. Like, who are the customers you're referring to in this scenario?

这些都是全新的客户群体。

These were completely new folks.

明白了。所以你们有了创意，找到客户验证，他们确实提供了很好的反馈信息。

Okay. So so you have an idea, you'd have customers, you'd run this by them, and then they would actually give you, like, really good information.

没错。关键是要问对问题。

Exactly. Okay. If you ask them the right questions.

确实。

Right.

要推动他们说出真心话，坦诚什么对他们才是最重要的。长话短说，我们最终找到了帮助企业在销售和营销中运用数据的创意。我们发现从'拥有数据'到'真正能用于实际销售营销'存在巨大鸿沟。在企业级市场，这个鸿沟更大，因为人们都把Snowflake、Databricks这类数据仓库当作数据库使用。我的两位联合创始人其实之前就在另一家YC公司Segment工作过。真的吗？

If you push them to, like, tell the tell you their honest truth about what is truly most important to them. So long story short, we eventually came to this idea that was helping companies use their data in sales and marketing. What we found is that there was actually a huge gap between I have data and I can now actually use it in production sales and marketing, And that in the enterprise, that gap was way larger because people were using data warehouses like Snowflake and Databricks as their database. And so we actually two of my cofounders used to work at a different YC company called Segment. We Really?

没错。

Exactly.

不会吧。是Segment团队的两个成员还是你的联合创始人？

No way. Two of the Segment guys or your cofounders?

所以他们是Segment团队的工程师。其中两位是Segment的

So they're engineers on the Segment team. Two of the Segment

好的。

Okay.

Segment的联合创始人最终确实投资了HITUCH，但这些人其实是Segment早期的工程师。我们原本以为Segment已经解决了所有数据问题，但当我们进军企业市场时，发现事实并非如此。这就是为什么我们认为我们的产品市场契合点特别适合企业级用户——因为我们发现Segment那种'把所有数据发给我，我来帮你整理并用于营销'的模式在企业级市场行不通。

Got the Segment cofounders did eventually invest in HITUCH, but, yeah, these are early engineers at Segment. Okay. And what we thought is that Segment actually solved all these data problems. When we went to enterprise, we found that it did not. And so that is why we felt like our product market fit was specifically for enterprise because we found out that the segment paradigm, which was send me all your data, I'll organize it for you, and I'll help you in marketing with it, broke an enterprise.

企业用户希望使用自己的数据库。嗯。这就是我们的核心创新点。我们决定帮助客户使用自有数据库，是的。然后让所有SaaS工具——整整100个——都能同步这个数据库的数据。

And enterprise, people wanted to bring their own database. Mhmm. And that was our core innovation. We decided we'll help you bring your own database Yeah. And we'll make all your SaaS tools, all 100 of them reflect that database.

这可能是个很小白的问题，但你们是大企业，拥有这么多数据。为什么让销售和营销团队与你们的大数据连接这么困难？就像，既然有全部数据，这个连接为什么不能简单点？

This is a really noob question, but you're a big business. You have all this data. Why is it so hard to connect your sales and marketing team with your big data where like, with all your data? Like, why is that connection not just easy?

我认为主要有两个原因。一是数据量太大，需要帮助人们理解哪些是正确的数据、能否使用、管理是否完善，还有各种企业级的数据使用限制。嗯。二是当你想实际完成某项任务时，比如我想运行广告活动，市场上根本没有现成的工具来实现这个。

I think there's two main reasons. One is you have so much data. You have to help people understand, like, what the right data is, can they use it, is it, like, governed well, like, all the kind of, like, enterprise, like, red tape that goes around data usage. Mhmm. And the second is when you wanna actually get something done, like, I wanna run an ad campaign, there's no tooling that exists for that.

所以我们发现市场上存在一个惊人的空白——数据库和营销平台之间竟然没有任何工具桥梁。实际上，营销平台也不希望存在这样的工具，因为他们想独占数据。于是我们决定打破这个局面，成为唯一不存储任何数据的SaaS工具，直到今天我们仍然不存储任何数据。我们根本就没有数据库。

So we just found this really surprising gap in the market that there was just no tool between the database and the marketing platforms. And, actually, the marketing platforms didn't want there to be because they wanted to own the data. And so we just said, we're gonna break that. We're gonna be the only SaaS tool that does not store any data, and we still to this day do not store any data. Like, we just have no database.

哦，哇。真有意思。那你们现在规模有多大？

Oh, wow. How interesting. So how big are you now?

我们大约有250名员工。好的。我们大概有，这么说吧，大约2000到3000名客户，我们可能和大约25家财富500强公司合作。

We're about 250 people. Okay. We are maybe, call it, like, 2,000 or 3,000 customers approximately, and we work with probably, call it, like, 25 of the Fortune 500.

不错。让我们快速回顾一下时间线。你们显然经历了COVID期间的困难低谷，然后开始与客户沟通。你们的关键年份是什么时候？比如，你们什么时候发现这个市场空白的？

Nice. Just let's go back to the timeline really quickly. You're you obviously had your COVID trough of sorrow, and then you started talking to, you know, the customers. What was your, like, your big year? Like, when did you figure you've discovered this gap.

事情真正开始蓬勃发展的年份是哪一年？

Like, what was the year things really started humming?

我们基本上每两年就会重新定义一次公司。所以我会说每两年，我们都会因为推出一个极其有趣的新产品而迎来需求的大幅增长。嗯。我们从数据连接产品转向了一个营销人员喜爱的用户界面。这带来了更多的需求。

We've somewhat, like, kind of redefined the company every two years. So I would say every two years, we have a huge surge of demand because we launch a new product that's, like, extremely interesting. Mhmm. So we went from this data connections product to a UI that marketers love. That caused a lot more demand.

像PetSmart和华纳音乐这样的公司，他们非常喜欢这个界面，并且每天都在使用。然后我们推出了一个AI决策产品，它使用强化学习为每个客户选择最佳营销方案。这可能就是你收到这么多短信的原因，因为人们只是把所有的短信都发送给他们的所有客户。是的。嗯。

Like, folks like PetSmart and Warner Music, they love this UI, and they use it every single day. Then we launched an AI decisioning product that uses reinforcement learning to choose the best marketing on a per customer basis. That's probably why you're getting so many text messages because people are just sending all their texts to all their customers. Yeah. Mhmm.

我们为每个客户决定最佳消息。这样我们可以减少发送给你的消息数量，比如一周一条，但内容实际上是相关的。

We decide the best message for every customer. And so we can send you less messages, like one a week, but it's actually relevant.

但这些消息质量更高，针对性更强。

But they're much better messages, much more targeted.

是的。所以，几乎每个B2C品牌都在思考：如何减少消息数量，减少退订，并增加实际参与度？这就是我们现在的工作。

Yeah. So, like, almost every b to c brand is, how do I reduce the number of messages, reduce unsubscribes, and increase actual engagement? That's, like, our job now.

所以你们的发展看起来像阶梯式上升。不是平滑的曲线。你们只是不断迭代，是的，是的。改变、迭代、再改变，然后变得越来越成功。

So you guys look like a stair step. You're not like a smooth curve. You just keep iterating Yeah. Yeah. Changing, iterating, changing, and getting more and more successful.

完全正确。

Exactly.

太有趣了。我和卡罗尔正在收集一些关于创始人模式的数据，比如它在实践中意味着什么。嗯。不仅仅是它的定义。你能告诉我们你的这段旅程有多长吗？你是否曾不得不进入创始人模式，或者你何时运用过其中的一些原则？

So fascinating. Carol and I are collecting a little data on founder mode, like, what it means in practice Mhmm. Besides just the definition of it. Can you tell us how long this your journey? Have you ever had to go into founder mode or when you've used some of the principles?

我非常喜欢这个概念，也很乐意分享一下我对它的理解。嗯。实际上我已经和我们整个高管团队以及所有销售领导分享了这一点，告诉他们会有时候看到我或我的联合创始人实践创始人模式，并且要期待它的发生，要对此感到投入和兴奋而不是害怕。

So I love this concept, and I'm happy to kinda share my take on what it means. Mhmm. And I've actually kind of shared this now with our whole exec team and all of our sales leaders as well, that there will be times where they see me or my cofounders practicing founder mode and to expect it to happen, to, like, be engaged and excited by it rather than scared.

哦，哇。那是个爱的警告。是的。就像是只是个警告。

Oh, wow. That's a love warning. Yes. It's like Just a warning.

我认为这对很多创始人来说是个很好的教训。比如，如果你提前告诉人们你要做什么，那么当你做的时候，他们就不会感到惊讶。他们不会认为你行为反常。他们不会认为你对他们生气。他们会想，哦，他说过他会这么做，现在他确实这么做了。

And I think this is, like, just a good lesson for a lot of founders. Like, if you tell people that you're gonna do something in advance, then when you do it, they're not surprised. They don't think you're acting out of character. They don't think you're mad at them. They're like, oh, he said he would do this, and now he is.

是的。没错。所以我知道为什么会发生这种情况。

Yeah. Yes. So I know why this is happening.

你没有打破角色。

You're not breaking character.

是的。完全正确。是的。所以你提前获得许可，这之后会给你更多的信誉去真正实践它。是的。

Yeah. Exactly. Yeah. So you ask for permission in advance, and it gives you so much more credit afterwards to actually practice that. Yeah.

很久以前有人教过我一些东西，我当时很难内化，但那就是，CEO的工作是什么？最终，我学到的是，CEO的工作是让公司成功，而不幸的是，并不是让周围的每个人都开心。

So I was taught something a long time ago that was really difficult for me to internalize, but it was like, what is the job of a CEO? And, ultimately, what I was taught is that the job of a CEO is to make a company successful, and it is unfortunately not to make everyone around them happy.

那是真的。那是

That's true. That is

真的。在早期我为此挣扎了很久，因为我总是试图倾听周围每个人的意见，接受他们的反馈，做他们认为应该做的事情。

true. And I I struggle with this so much in the early days because I would try to, like, listen to everyone around me, like, take their feedback, like, do the things that they thought should be done.

建立共识。

Consensus building.

没错。而且是以相对民主的方式。是的。那是一种糟糕的公司领导方式。说实话，非常坦率地说，我觉得我在后来，大概2024年左右，成为了一个更好的联合创始人和联合CEO，那时我停止了谈判。

Right. And in a relatively democratic way. Yeah. That is a terrible way to lead a company. I think I became a significantly better cofounder and co CEO later on around, like, call, like, really 2024, to be honest, very candidly, when I stopped negotiating.

所以人们会说，哦，但是，我们别那样做。我们做那个的一个版本吧。但作为创始人，你对事情有如此强烈的信念，实际上你应该完全地，就像，勇往直前，就像，让我们以正确的方式来做这件事，而不是某种妥协的方式。所以我对创始人模式的定义基本上是，你的工作是提供结果或让公司成功，而那些为了公司愿景或使命在那里的人，总是会更喜欢这样，而不是你对他们说‘是’，让他们开心，甚至提供指导或职业成长。就像，他们总是会为了你做出让公司成功的决定而牺牲一些那些东西。

So people would say, oh, but, like, let's not do that. Let's do, like, a version of that. But you as a founder have so much conviction in the thing that you should actually just take full kind of, like, go ahead on, like, let's actually do this the right way rather than some compromise type of way. So my definition of founder mode is basically your job is to provide results or to make the company successful, and the people that are there for the vision or the mission of the company will always prefer that versus you saying yes to them, making them happy, even providing the mentorship or career growth. Like, they will always sacrifice some of those things for you making the decisions that make the company successful.

是的。你知道，这很有趣。所以，你知道，有时候我们在这个播客上会进入父母模式。

Yeah. You know, it's funny. So, you know, sometimes we go into parent mode when we're on this podcast.

是的。我们确实会。

Yes. We do.

而这个，你所说的让我想起这个，有一种心理学认为，就像，孩子们实际上在父母更严格时感觉更好。嗯。因为界限实际上让他们感到安全。嗯。这有点——我不知道。这不完全可比，但这就是我脑子里想的。

And this, what you're saying reminds me of this, there's a psychology that, like, children actually feel better when their parents are more strict Mhmm. Because the boundaries actually make them feel safe. Mhmm. This is a little I don't know. It's not exactly comparable, but that's what I'm in my head.

我在想，就像，也许这可以转化到工作场所。

I'm thinking, like, maybe that translates to the workplace.

我必须承认我当时想的是这个关于在我孩子面前打破角色设定的想法。就像，我总是试图——如果你，就像，失控并表现得疯狂，那会很糟糕，因为你打破了妈妈的角色。

I have to admit what I was thinking is this idea of breaking character with my kids. Like, I always tried it was bad if you, like, lost it and acted crazy because you're breaking character of mom.

你懂吧？是的。

You know? Yeah.

是的。所以我们俩都进入了我们的育儿模式。就像这样。这是对一切的隐喻。是的。

Yeah. And so we both went to our parenting mode. Like this. It's metaphor for everything. Yeah.

这是个好建议。

It's good advice.

是的。而且我认为，很多人会聘请经验远超自己的高管。那个人在那个领域确实应该比你更出色。但你拥有来自公司其他部门、市场和客户的视角。

Yeah. And I think, like, a lot of people, for example, will hire executives that have significantly more experience than them. That person should be better than you at that domain than they are actually. But you have perspective both from all the other parts of the company as well as from the market and from the customer.

但这很难调和吧？这个人可能在某个领域经验更丰富，但你希望他们听从你的意见。

Is it hard to reconcile that, though? This person might have more experience in x, but you want them to listen to you.

非常困难。是的。尤其是当你经常是房间里最年轻的人时。

So difficult. Yeah. Especially when you're, like, often the youngest person in the room.

没错。是的。

Right. Yes.

但我认为，接受你拥有市场和客户视角这一事实会有所帮助。因为你最终会意识到，作为企业CEO，你总是比其他人掌握更多背景信息。所以如果你有某种感觉，很可能是有原因的，你有责任坐下来思考、写作，直到最终弄清楚为什么自己如此坚持某种做事方式。是因为我抗拒改变，还是因为有实际的客户背景支持这种做法？

But I think, like, embracing the fact that, like, you have perspective on the market and the customer is what helps. Because you eventually realize that as a CEO of a business, you're always gonna have more context than everyone else. So there if you're feeling something, there's probably a reason why, and you owe it to yourself to sit down and think and write until you eventually identify why was I so strongly held to a certain way of doing things. Is it because I'm resistant to change or because there's actual customer context that enforces why this should be done this way?

你会通过大量写作来帮助思考吗？

Do you do a lot of writing to help you think?

是的。这是我过去几年学到的。我几乎每天每小时都在写。哇。倒不总是写日记，但几乎每个想法我都会写下来。

Yeah. So that is, I guess, what I learned in the last couple years. I write almost every hour of every day. Wow. So I'm not always journaling, but pretty much every thought I have, I write down.

你会分享其中的多少内容？

How much of it do you share?

我会尽量分享所有到周末仍然成立的内容。就像

I try to share everything that is left that is still true by the end of the week. So, like

这真是个复杂的过程。

That is quite a process.

你知道，就像，每个人都有，嗯，波动性。对吧？你可能在一周内失去信念又重拾信念。到周末甚至当天结束时，如果某件事仍然存在且你仍然相当确定，这时候如果其他人不知道，我就会开始感到不安。所以我还学到一点：如果你知道某些事却还没告诉公司里的其他人，特别是你的高管或领导层，那么，基本上你就是在失职或撒谎，因为你在向他们隐瞒信息。

You know, like, everyone has, like, volatility. Right? You might lose conviction and gain conviction throughout a week. By the end of the week or even end of the day, if something is still there and you're still quite sure about it, that is when I start feeling uncomfortable if other people don't know. So another thing I was taught, if there's anything that you know that you've not told other people in your company, especially your executives or, like, your leadership, then, like, you're basically not doing your job or you're lying because you're hiding information from them.

而如果你大致想明白了某件事，或者你的直觉在某种程度上指引你去某个方向，你有责任尽快与他们分享这些信息。

And if you kind of figure something out or your intuition is kinda leading you somewhere, you owe it to them to share that information as soon as possible.

对你来说，这个时间线是一周。就像，你给自己一周时间

And for you, that timeline is a week. Like, you give yourself a week

理想情况下是的。

Ideally.

如果你还在考虑它，那就是时候公开了。

And if you're still thinking about it, it's time to disclose.

理想情况下是的。好吧。而且，你觉得，人们会，嗯，感到害怕，因为他们正在获取关于这个的信息，就像，也许这是，嗯，负面的东西。是的。不。

Ideally. Okay. And, like, you think, like, people will be, like, scared because they're they're getting information about this, like like, maybe it's, like, negative stuff. Yeah. No.

他们其实相当高兴，因为他们也感觉到了。如果你雇佣的是优秀的人，他们几乎总能感觉到。有趣。这时候他们就会开始发言，说，哦，我也有这种感觉。让我们一起解决它吧。

They're actually quite happy because they sensed it too. If you hire good people, they almost always sense it too. Interesting. That's when they start speaking up and be like, oh, I'm feeling this as well. Let's solve it together.

是的。我有一位，嗯，领导力教练，他叫David Clements。他教了我很多这类事情。他教我的其中一点是，公司里唯一能承担犯错后果的人最终是创始人，因为他们不太可能真的被解雇。我的意思是，这并不总是对的。

Yeah. I have a, like, a leadership coach, and his name is David Clements. He's taught me a lot of these things. One thing he's taught me is that the only person that can afford to be wrong in the company is ultimately the founder because they can't really get fired. I mean, it's not always true.

显然，创始人确实会被解雇。是的。但无论如何，就像，大多数人都是风险规避的

Obviously, founders do get fired. Yeah. But for all intents and purposes, like, most people are risk averse

是的。

Yes.

因为他们有被解雇的风险。创始人是唯一能说'我要做必须做的事并承担所有风险'的人。这意味着，如果你不是100%正确，说明你承担的风险不够，也就是说你没有尽到创始人的职责。很多关于创始人模式的定义都是：做只有创始人能做而别人做不了的事。所以在这种情况下，我认为风险基本上是创始人唯一能独特承担的事情，而其他人会难以承担。

Because they have the ability to get fired. The founder is the only person that can say, I'm gonna do the thing I need that needs to be done and take all the risk. And so what that extrapolates to is that if you're not if you're right a 100% of the time, you didn't take enough risks, which means you didn't do the job of a founder. And, like, a lot of founder mode definitions are, like, do the thing only the founder can do that no one else can do. And so in this case, I'm saying that risk is basically the only thing the founder can do uniquely that other people will struggle to take.

有意思。或者他们会想要得到你的许可。

Interesting. Or they'll want permission from you.

或者他们会想要许可。他们会降低风险。他们会承担一半风险。

Or they'll want permission. They'll tone it down. They'll take half risk.

是的。没错。

Yes. Yeah.

这些都是对冲。

All these are Hedge.

他们会进行对冲。

They'll hedge.

他们会进行对冲。完全正确。一个好的例子是去年年中，当时仍然没有人真正相信我们能实现收入计划。我基本上是说：嘿，我们要将销售团队规模翻倍，我认为我们不仅能实现今年的收入计划，还能实现明年的，并且再次翻倍。

They'll hedge. Exactly. So good examples of this is mid last year, no one really believed in us hitting our revenue plan still. And I basically said, hey. We're gonna double the sales team headcount, and I think we're not only gonna hit this year's head count, like, revenue plan, but also next year's and, like, double again.

但数据并不支持。是的。我的感觉是市场告诉我，对这个产品的需求比以往任何时候都大，而且这会体现在我们的渠道中，比如新增客户，但不会体现在我们历史的数据指标中。所以这全都是前瞻性指标。

But the the numbers were not there. Yep. What I was feeling is that the market was telling me there's more demand for this product than ever, and that it was gonna show up in our pipeline, like, net new, but not in our, like, metrics historically. So it's all, like, forward metrics.

你是怎么得到这种感觉的？是通过和客户交谈吗？比如，你是怎么

How did you get that feeling? Is it from talking to to your customers? Like, how did

你知道吗？

you know?

我开始直接管理销售团队。大概，嗯，有四个经理和大约20名销售代表。好吧。我开始直接上手工作。

I started managing the sales team directly. So it was, like, maybe, like, four managers and, like, 20 reps. Okay. I started working directly.

你们之前有销售经理吗？

Did you have a sales manager before?

有的。我们有一位销售主管。我们当时处于那位销售主管和首席营收官之间的过渡状态

We did. We had a sales leader. We were kind of in between that sales leader and a chief revenue officer

好的。

Okay.

我们正在寻找首席营收官。好吧。而学习如何招聘首席营收官的一部分就是我自己来做这份工作。嗯。比如，阅读相关书籍，管理团队，学习这项工作。

That we were looking for. Okay. And part of learning how to hire a chief revenue officer was doing the job myself. Mhmm. Like, reading the books, managing the staff, like, learning the job.

结果，通过与所有那些客户交谈，我意识到实际上在每个团队，我们都低估了需求，而限制因素仅仅是我们人手不足。

And then as a result talking to all those customers and realizing actually in every single team, we had under forecasted the demand, and it was actually just limited by the fact that we didn't have enough people.

所以风险在于我们会花很多钱。是的。我们会招聘一大堆人。而招聘是有风险的，因为解雇是最糟糕的事。

So the risk was we're gonna spend a lot of money. Yeah. We're gonna hire a ton of people. And hiring is risky because firing is the worst.

而你绝对不想经历那种裂痕（指解雇带来的伤害）。

And you you never wanna go through the rift.

你不想经历那种裂痕。那太打击士气了。是的。

And you don't wanna go through the rift. It's so demoralizing. Yes.

而且我无法用数据证明我是对的。

And I couldn't prove with data that I was right.

没错。你除了直觉之外没有任何依据。

Right. You had nothing to back it up except your gut.

是啊。还有很多反对者。

Yeah. And a lot of naysayers.

哦，是的。全是反对者。

Oh, yeah. Only naysayers.

好的。这是个好问题。

Okay. That's a good question.

我当时

I was

正想说，就像你说的，有时候当你凭直觉有这些感觉时，其他人也会察觉到，他们会说，谢天谢地你说出来了，因为我也有同感。但这次情况并非如此吗？

just gonna say, like, you said that sometimes when you have these gut feelings, other people will have sensed it, and they'll be like, oh, thank god you said it because I was feeling it too. In this case, was that not true?

说得很好。坦白说，高层中有很多反对者。是的。但在基层，有很多人非常赞同。你跟销售代表们聊聊，他们会告诉我他们将完成两到三倍的配额。

That is a great point. Candidly, there was a lot of naysayers at the top. Yeah. On the ground level, there was a lot of people that strongly agreed. You talk to the reps, they're telling me I'm gonna close two to three x of quota.

是的。好的。

Yeah. Okay.

跟经理们谈谈，他们会说，是的。卡希什，我还没告诉你，但我知道我会比你预期的多完成大概400万。

Talk to the managers, they say, yeah. I haven't told you this yet, Kashish, but I know I'm gonna hit, like, 4,000,000 more than you think I am.

哇。

Wow.

然后你翻到最底部，他们会提供所有这些定性反馈，这些不一定在数据中体现，但他们感到非常乐观。

And then so you go to the very bottom, and they have all this qualitative feedback that's not necessarily in the data where they feel very, optimistic.

哦，有意思。

Oh, interesting.

哇。

Wow.

这是一个非常好的创始人模式例子。

That's a very good founder mode example.

一个很棒、非常具体的例子。

A great, very specific one.

所以我还有另一个观点，我认为很多时候，平级之间的协调是最难的。很多创始人模式的人会说，最终有人向我汇报，我可以明确告诉他们该做什么。但你的联合创始人更像是你的平级。

So the other one I have is, I think oftentimes, like, coordinating amongst equals is the hardest. So a lot of founder mode folks are like, well, ultimately, someone reports to me, I can tell them exactly what to do. Your cofounders on the on the other hand are more your equals.

没错。

Right.

是的。所以在联合创始人之间做决定可能是压力最大但回报也最大的，因为如果你们非常一致，就能创造奇迹。对我们来说，今年早些时候，我们在开发一个人工智能产品。那是一个主要不使用LLM的强化学习产品。你会看到所有其他AI公司都专注于使用LLM，那是主要的逻辑单元。

Yeah. And so making decisions between your cofounders is probably like the most stressful but the most rewarding because as if you're very aligned with the cofounders, then you can just move mountains. So for us, this was earlier this year, we were working on AI product. It was a reinforcement learning product that predominantly did not use LLMs. So you see all these other AI companies that are focused on using LLMs and that's the main logical unit.

对我们来说，是强化学习，我们认为这是营销的正确方式。是的。因为CMO们真的很害怕LLM，他们对内容生成非常抵触

For us, it was reinforcement learning, which we thought was the correct way for marketing Yeah. Because CMOs were really scared of LLMs and they were like really averse to content generation at

那个时代。害怕大语言模型。

the time. Scared of LLMs.

去年整整两个季度都是这种情况。我会说今年大部分时间，首席营销官们都说内容生成还没准备好

All of h two last year, that was the case. I would say most of h one this year, CMOs said content gen is not ready for

我。好吧。

me. Okay.

于是我们想，让我们用强化学习来选择最佳内容，并围绕这个构建了这款AI产品。市场反响相当热烈，但他们的采用速度比较慢。此时，假设我们公司有四个主要产品，但只有大约50名工程师。所以我们的人手严重不足，却要覆盖非常大的业务范围。后来我和联合创始人商量，我说，嘿，各位。

And so we thought let's use reinforcement learning to choose the best content, and we built this AI product around it. The market was pretty excited, but they were, like, adopting slowly. And so at this point, let's say our company has, like, four main products and we have, like, 50 engineers. So we're, like, very understaffed for, like, a very large surface area. What happened is I talked to my cofounders and I said, hey, guys.

我们现在做的事情，今年可能会翻倍。明年很可能还会翻倍，但我认为还有一个更大的机会，虽然听起来有点吓人，但我觉得我们现在正在攻关的AI机会并不是最大的。我认为还有一个更大的机会没人攻克。我们必须自己开辟道路，开拓自己的市场。但如果我们做到了，我认为我们将建立一个更持久的企业。

Like, what we're doing, we're gonna double this year. We're probably gonna double next year, but, like, I think there's a much bigger opportunity, and it's kinda scary to hear, but I don't think the AI opportunity that we're working on right now is the largest opportunity. I think there's an even bigger one that no one has cracked. We have to define our own way and, like, pave our own market. But if we do that, I think, like, we're gonna build a significantly longer lasting business.

是的。我是这样阐述的：实际上，以我们现在的处境，即使我们三个停止工作，今年仍然会翻倍。

Yeah. And the way I framed this was, actually, like, given where we are now, if us three stopped doing our job, we're still gonna double this year.

哇。

Wow.

所以这应该让你们既安心又兴奋，去追求更大的目标，去争取最大的机会。这是个全新的认识，因为通常我们太关注细节，不相信没有强力参与就能达成目标。但现在团队成熟了，需求也在了。所以这是一种非常清新的感觉，但分享这个作为杠杆：我们能做什么来彻底重新定义营销？如果现在要组建一个营销团队，利用当今的技术

And so that should give you relief and excitement on going for the moon, on, like, actually going for the largest possible opportunity. And this was it's brand new information because usually, like, we were so in the details that we didn't believe that we could actually hit our numbers without really strong involvement. But now the team is there and the demand is there. So that was like a very refreshing feeling, but sharing that as the lever for what can we work on that would actually completely redefine marketing. If you were to build a marketing team right now, given the technologies we have today Yeah.

你会如何毫无偏见地组建那个营销团队？我们开始向每一个客户提出这个问题

How would you build that marketing team without any biases? We started asking every single one of our customers this

是的。

Yeah.

就像世界上一些最优秀的首席营销官一样，他们都说，是的，我们今天做的这些事情都不会做。真的吗？我们写了这么多简报，基本上就是在黑暗中摸索什么有效，不断试错，我们会做出更多基于数据的决策，让大型语言模型（LLMs）分析所有营销活动和所有数据。嗯。然后和我们一起做决策。

And like the some of the best CMOs in the world, and they all said, yeah, like, we would not be doing any of the stuff that we're doing today. Really? Like, we're writing all these briefs, we are like basically just like stabbing at the dark for what works and doing trial and error, we would be making way more data informed decisions and let LLMs like look through all the campaigns, all the data Mhmm. And make decisions with us.

哦，哇。

Oh, wow.

于是我们基本上就开始创造这个产品，我们开始和他们坐下来谈，嘿，我可以免费给你这个产品，但请和我一起合作，和我一起开发它。

So then we started basically inventing that, like we started sitting with them saying, hey. I'll give you this product for free, but, like, work with me and develop this with me.

是的。你的联合创始人和你意见一致吗？或者你需要做多少说服工作？

Yeah. And your cofounders were aligned with you, or how much convincing did you have to do?

老实说，我们花了大概两到三个月的时间互相说服。

It took us on, honestly, like, two to three months of convincing each other.

有意思。

Interesting.

因为我不知道具体该做什么。是的，我们谁都不知道。只是觉得我们可以做得更多，而且现在市场上正发生技术转变，没有其他人会做，因为他们没有我们这样的营销人员资源。所以我觉得，如果当时有一个明确的提案，那就会简单得多，好吧。

Because I didn't have, like, what to do. Yeah. None of us had what to do. It was just like we could do a lot more, and there is a technical shift in the market right now that no one else is gonna do because they don't have the access to marketers that we do. And so it was I think if, like, if there was a clear proposal, then it'd be like, okay.

让我们做一个是或否的决定。是的。但这更像是，嘿，我们做点什么吧。我们一起想办法决定该做什么。

Let's make a yes, no decision. Yeah. This was much more like, hey. Let's do something. Let's figure out together what to do.

是的。我们大概花了三个月时间。

Yeah. And it took us probably three months.

你有两位联合创始人？是的。好的。

You have two cofounders? Yes. Okay.

所以总共我们三个人。

So three of us total.

是的。哇。嗯，那真的很有趣。我很喜欢你的例子。太有意思了。

Yeah. Wow. Well, that was really interesting. I loved your examples. It's so interesting.

是的。

Yeah.

这种具体性很棒。

The specificity is great.

是的。基于这些访谈，我们将获得非常好的信息。所以谢谢你

Yes. We are gonna have such good information based on these interviews. So thank you for

参与进来。是的。谢谢邀请我。

coming on. Yeah. Thank you for having me.

非常感谢。

Thank you very much.

是的。谢谢。也很高兴认识你。

Yeah. Thanks. So nice to meet you too.

我也很喜欢。

I love too.

顺便说一句，对High Touch超级期待。

Super excited for High Touch, by way.

听起来是一家很棒的公司。

Sounds like an amazing company.

我期待看到你们持续取得成功。

I look forward to watching your continued success.

我很感激。是的。谢谢。

I appreciate that. Yeah. Thank you.

谢谢。好的。再见。

Thank you. Alright. Bye.