免费增值模式的正确做法：来自价值数十亿美元应用的启示——克里斯·哈尔斯，Life360

对吧？

Right?

它会变成这样：哦，只差十分之一个点，但显著性计算器却说这是六西格玛之类的。

It'll be like, oh, it's a tenth it's a tenth of a point, and the significance calculator says it's six sigma or whatever.

然后你就想，等等，我们好好想想。

And you're just like, well, let's think about it.

不知为何，我们在分析数据时实际上没遇到过这个问题。

We actually haven't had that problem in terms of looking at data for whatever reason.

我有一个特别在意的问题，我打算把这次采访发给他。

One of my pet peeves I'm I'm actually gonna send this interview to him.

我很高兴我们做了这次对话，因为这正好表达了我所有想发给我们团队的想法。

I'm glad we're doing this because it's kind of all my thoughts that I can send to our

发给团队，这样我就能按自己的方式来。

team and just can get my own way.

Life360团队的内部沟通。

Internal comms for the for Life360 team.

是的，没错。

Yes, exactly.

这实际上会是一些内部沟通。

This will actually be a bit of internal comms.

我讨厌我们根据用户基数的百分比来做测试。

I hate when we do tests based on percent of user base.

我非常讨厌这样。

I hate it.

哦，你是说只在用户基数的一个子集上做？

Oh, like on just on a subset of the user base, you mean?

不是，我们不是用n，而是用百分比。

No, that we, instead of n, we do percent.

我们用50%的人、10%的人、2%的人来做测试。

We're testing this with 50% of people, 10% of people, 2% of people.

我说，以我们这么大的用户基数，可以近似看作一亿。

And I say for our size base, can round us to 100,000,000.

然后人们会说，我们需要20%才能达到统计显著性。

And then people say, well, we need 20% to get statistical significance.

那可是两千万人啊。

Well, that's 20,000,000 people.

所以你是说，一家连两千人都没有的公司，就根本没法做测试吗？

So are you saying a company that doesn't have 2,000 people, they just literally can't test?

我仍然一次又一次地看到这种情况发生。

I still see this happen again and again and again and again.

我认为这实际上阻止了人们进行明智的测试。

And I think what it does do is actually stops people from doing smart tests.

因为如果你想要做一个假门测试，比如双知的假门测试，本质上就是假的。

Because if you want to do like a painted door test, painted door for dual knows, like, it's basically fake.

你点击它，它就坏了，于是你制造了一种优雅的失败。

You click it and it breaks, so you have made some graceful failure.

你完全可以只在100个人或10个人身上做这个测试。

You should be able to do that on a 100 people or 10 people.

然后，好吧，我们可能得不到足够的预测力，但我对那种门板测试中获得70%确定性的情况还挺能接受的。

And then, well, then we're not gonna get enough predictive power, but I'm kinda okay in a pane of door that you get, like, 70% certainty.

是的。

Yeah.

如果你的改动足够显著，你还是会获得足够的预测力的。

Well, you you get enough predictive power if it's a powerful enough change.

你明白我的意思吗？

You know what I mean?

一百个人，只要改动够大，你就能看出来。

A 100 people with a big change, you'll see it.

对。

Yeah.

没错。

Exactly.

或者，如果我们只是做门板测试，那我们就直接做真实测试。

Or enough just like, if we're just doing a pain in door, then we'll do the real test.

但这个观点认为，人们需要真正理解他们正在测试的内容和时机。

But this idea that we need to just like people need to really understand what they're testing and when.

所以，当然，如果我们只是想改一下标题和文字，没错，风险很低，可以直接推送给所有人，你大概能捕捉到0.1%的变化。

So sure, if we're trying to like change some header and some text, yeah, low risk, roll out to everyone, and you're kind of okay kind of picking up a 0.1% thing.

但如果你要尝试一些全新的东西，或者大范围地改动，比如给100个人换门漆。

But if you're trying to try something really new or build a lot, like do the paint the door to 100 people.

但别跟我说我们要只推1%，因为那样你会展示给将近一百万人。

But don't tell me if we're gonna do it 1%, because then you're showing it to almost a million people.

是的。

Yeah.

对。

Yeah.

百分比会大得多。

The percents get much bigger.

对吧？

Right?

但我认为人们只是默认1%很小，50%就很大了，实际上在做发布时，我们确实会用一、五、十、五十这样的比例，因为这涉及到更多的服务器流量，而且

But I think people just assume 1% small, 50 is big, and there are times you wanna do, like, you're doing a rollout, we do do like one, five, ten, fifty, because that's more server traffic and

哦，因为你正朝着100%前进，对吧？

Oh, because you're on your way to a 100, right?

是的，完全正确。

Yes, exactly.

如果我们是在逐步推进，我会称这为发布，而不是测试。

If we're on our way, but I call that is a rollout, not a test.

而且，如果只是像邮件标题这样的内容，当然可以。

And again, if it's and if it's like email headline copy, sure.

用百分比没关系，因为这真的无关紧要。

Like, do the percents because it really doesn't matter.

但你到底在测试什么？

But what are you testing?

为什么？

Why?

正确的研究方法是什么？

What's the right methodology?

要批判性地思考。

Like, think critically.

从第一性原理出发思考。

Think first principles.

不要只是照搬操作手册。

Don't just follow a playbook.

但我敢保证，我会在内部展示这个，然后回去说我们需要测试1%的样本，然后

But I guarantee I'll show this internally, and then I'll go back to say we need to test 1% and

得把它写进权利法案里。

Gotta put it in the Bill of Rights.

不要用百分比。

No percents.

不要用百分比。

No percents.

是的。

Yeah.

你经常做定性研究吗？

Do you do a lot of qualitative?

假设你对100个人进行了一次假门测试。

Let's say you run a test on a 100 people fake door test.

然后你会尝试联系这些人打电话吗？

Do you then get this try to get those folks on the phone?

还是说，这是你流程的一部分？

Or like, is that part of your process?

随着我们的发展，有些人让我们让各个团队更自主地行事。

Some people did as we've grown, we've kinda let more individual teams do their thing.

我与用户的交流可能比大多数人建议的要少。

I probably spoke to users less than most would suggest.

我不是在夸耀这一点。

I'm not saying it in a good way.

我本该多做些这样的事。

I probably should have done more of that.

我觉得自己天生就擅长设身处地为别人着想。

I think I have naturally been gifted at sort of like putting myself in someone else's shoes.

我会真的闭上眼睛，试着想象：如果我是用户，我会怎么反应？

I literally close my eyes, and I try to think like if I'm I do a little, like, what am I what how am I gonna react to this thing?

我会怎么做？

What am I do?

不是我，而是客户。

Like, not me, but the the customer.

里克·鲁宾的方法。

The Rick Rubin method.

对吧？

Right?

你就只是闭上眼睛。

You just, like, close your eyes.

那位著名的制作人根本不演奏任何乐器，几乎什么都不做。

The famous producer who just doesn't play any instruments, doesn't really do anything.

他只是感受一下，然后就知道该怎么做。

He just, like, kinda feels it and, like, knows what to do.

我合作过的每一位出色的消费者领域创业者，都具备强烈的情感共情能力，同时也很擅长理性分析。

Every great consumer founder I've worked with has always had a really both, like, emotionally empathetic, like, ability, but then also, like, they're very analytical about that.

你明白我的意思吗？

You know what I mean?

他们能够机械地把自己代入到许多不同人的思维模式中。

They can, like, put themselves mechanically into the mindset of many different people.

当我从事消费者领域工作时，有时发现有太多声音，单个对话有时会具有误导性。

And, like, when I was working consumer, I sometimes found there's so many voices that, like, individual conversations can sometimes be misleading.

因为行为差异很大，你个人的特质几乎不可能和你所服务的客户高度相似。

Because, like, there's a lot of variance in behaviors, and, like, it's very unlikely that you personally are all that close in characteristic to the person that you're selling to.

有时我觉得这种联系可以如此强烈，所以我常说：创办一家B2B公司吧，这样你就被迫时刻和自己待在一起。

And sometimes I think that connection can be so I say, start a b to b company, then you're forced to, like, be in the room with you all the time.

这实际上是个优势，因为稍微容易一些，但当我做消费类产品时，我也和你处在同样的位置。

It's actually been a benefit because it's a little bit easier, but I was in the same place as you when I was in consumer.

我觉得这仍然很好。

I think it's still good.

我觉得，而且我也不是很确定你们会不会做调查之类的事情。

I think, like and also, like, I don't know if y'all do surveys and stuff like that.

你们说的是，要超越单纯的二元实验中的统计显著性。

You're talking about, like, extending beyond just, like, pure, like, a b test sig statistical significance on a binomial experiment.

给那100个遇到故障门的用户发个调查问卷，看看会不会有回复。

Just sending a survey to those 100 people that hit the broken door and being, you know, see if get any responses.

我觉得这些方法都不错，但你必须清楚为什么这么做，不能只是盲目跟风，要非常审慎，不要武断地认为这就是唯一正确的方式。

I think they're all good and you just need to know why, and you don't do it just because, and you're really thoughtful about it, and you don't overly kind of like saying this is the way.

除了我们的固有权利，别跟这个较劲。

Other than our villa rights, you know, don't argue with that one.

开个玩笑，但说实话，这些正是我们真实的面貌，而我们对此有非常清醒的认知。

But jokes aside, maybe up to date, these are the things that just who we are and but we're being very self aware of it.

这些是宗教信仰的产物，因为我们相信它们是我们身份的核心，这些正是我们所测试的内容，也是我们思考问题的方式。

These are the artifacts of religious faith because we believe they're core to who we are, and these are the things we test, and this is how we think about it.

对于这些免费用户，你们如何考虑将他们变现？

With those free users, how do you think about monetizing them?

那么，将他们变现的路径是怎样的？

And and what has been the path of monetizing them?

我知道最近你们一直在开发一个内部广告网络，但你们一直以来都有展示广告吗？

I know most recently, you have been working on an in house ad network, but have you shown ads this whole time?

或者像

Or like

广告是新推出的。

Ads are new.

我们在主页上放了一些非常难看的横幅广告，每次看到它们，我都会产生真实的生理不适反应。

We have some really ugly banners in the home screen that make me wanna I genuinely have a physical reaction every time I see them.

好吧，克里斯，你只要订阅了，这些广告就会消失。

Well, you just gotta subscribe, Chris, then they go away.

这是

It's

但广告是用户体验的一部分。

But ads are part of the user experience.

免费用户的体验是，我从我的个人项目应用中移除了广告，因为它们的体验实在太差了，所以我改用硬性付费墙，因为这确实是用户体验的一部分。

The free user experience is I removed ads from my side project app because the user experience of them were so bad, and I went to a hard paywall because it really is part of the user experience.

是的。

Yeah.

所以回到数据的模糊性问题，团队做了对照组实验，证明广告根本没有造成任何负面影响。

So how going back to this the fuzziness of data, the team do the holdout group and show that the ads didn't hurt anything at all.

我认为这些数据实际上是有效的，但如果你重复做20次，你就没把握了。

I think that data was actually valid, but do that 20 times, you don't know.

但我们决定收益值得尝试，因此我们正试图启动这个新业务，并希望随着时间推移，能将广告产品逐步优化成用户真正喜欢、甚至根本察觉不到是广告的形式。

But we decided the upside was worth it, and we are trying to start this new business, and that over time we can evolve the ad product to something that customers actually love or don't even realize it's an ad.

所以我们现在的横幅广告，简直糟糕透顶。

So the banners we have, they're horrible.

我讨厌它们。

I hate them.

我已向团队提出挑战。

I've challenged the team.

比如，我们能在一年内做到吗？

Like, can we find in, like, a year?

我们能不能别再用这么难看的横幅广告？

Can we not have ugly banners?

但每天都要大力推动，表明你关心客户，希望这些广告消失。

But really pushing every day, like show that you care about the customer, you want these gone.

我们做过一些非常酷的广告，比如当你抵达机场时，会看到：‘嘿，David，欢迎来到旧金山机场，需要Uber吗？’

We've done some really cool ones, like when you land at an airport now, like, hey David, welcome to SFO, do you need Uber?

一键搞定。

One tap.

你们是直接和Uber谈的这笔合作吗？

Did you do that deal directly, like with Uber?

我们确实做了，是的。

We did, yeah.

我们还有另一个即将发布的，非常类似，它真正对客户有好处，人们可能知道这是广告，但不会打断他们的工作流程。

We have another one we're announcing, is very similar where it's not it's actually good for the customer, people probably know it's an ad, but it's not interrupting the workflow.

如果你真的想与广告互动，它会带来价值。

And if you actually want to interact with the ad, it's adding value.

我觉得做得最好的公司是那些与订阅无关的，比如Credit Karma和信用卡广告。

Think the company that did that the best, nothing to the subscription, but Credit Karma and credit card ads.

嗯。

Mhmm.

对。

Yeah.

你甚至都察觉不到那是广告。

You don't even know it's an ad.

对吧？

Right?

你只是在使用他们的网站。

You're just using their website.

嗯，我觉得你可能确实如此。

Well, I think maybe you do.

但对于不了解Credit Karma的人来说，本质上是你提供了所有非常私密的敏感信息，比如你的社会安全号。

But like for people who don't know Credit Karma, it's basically you give all your very personal sensitive information, your social.

他们会为你匹配信用卡优惠。

They match you with credit card offers.

你的信用评分是X，你住在这里，你需要积分、需要旅行，等等。

And your credit score is X, you live here, this is you want points, you want travel, whatever.

这是最适合你的卡片或现金返还。

This is the best card or cash back.

他们只是为你匹配卡片。

They just match you with cards.

所以这完全是线索生成。

So it's a 100% lead gen.

他们为你匹配的每一样东西都是真实的广告，但对客户来说，你实际上是被匹配到了最适合你的产品和服务。

Everything they match you with is an actual ad, but to the customer, you are actually getting matched with the product and service that is good for you.

所以我觉得这和垃圾信息轰炸完全不同。

And so I think that's very different than spamming someone and all that.

我要回到《权利法案》了。

I'm going back to Bill of Rights.

我们说过，我们不会——这实际上写在我们的《权利法案》里。

What we've said, we will not That's actually in our Bill of Rights.

我忘了那个观点了。

I forgot all that point.

这和广告有关。

Was tied with advertising.

我们不会妨碍用户完成他们要做的事，通过

We will not get in the way of the job to be done with an

当然。

Sure.

知道全屏插屏广告，需要三十秒才能关闭。

Know full screen interstitial, thirty seconds to close.

是的。

Yeah.

那么哪些广告让你特别反感呢？

What are the ads that upset you so much then?

不是机场里的优步广告。

It's not not the Uber at the airport ones.

那些看起来

Those seem

不是。

No.

那个很好。

That one's great.

我真心觉得，我可以直视着对方的眼睛，坦然地说，这确实为你提供了价值。

I genuinely think I can look someone in the eye with a straight face, and I actually think that's giving you value.

因为如果你以前想要那种一键操作，现在根本不会对你造成任何困扰。

Because if you used to want that one tap, it's not hurting you at all.

所以我们现在只有这些横幅广告。

So we just have these banners.

它们只是说明横幅广告只对品牌曝光有效。

They're just it's just more that like banners are only good for brand impressions.

它们其实并不能推动实际转化。

They don't really drive performance.

所以

So

而你可能无法控制通过广告交换平台或其他方式引入哪些品牌吧？

And you don't get to control probably using an ad exchange or something like what brands get pulled in?

我们确实在这么做，但我们有自己的广告单元，会强制用户做一些操作来让它看起来更自然。

We are doing that, but we have our own ad unit that does force people do a little bit to make it look there.

那也是我们当时设定的若干条任意界限之一。

That was one of those just somewhat arbitrary lines we did as well.

顺便说一下，这些都是一些只有在拥有像Slide三十六这样的大量用户时才能做到的事情。

By the way, these these are all, like, things you only get to do when you have as many users to Slide three sixty.

对吧？

Right?

如果你只是一个普通的应用，你根本没有这样的议价能力来做这些。

Like, if you're a random app, you don't have the leverage to do this.

是的。

Yeah.

没错。

Exactly.

你根本做不到，不行。

You you would not be able to no.

你没有这个选择。

You don't have that choice.

所以你看，你从来没有只是简单地安装了Google AdMob然后让Google直接投放广告吗？

So you see, you never, like, just installed Google AdMob and just let Google serve?

我想我们确实这么做了，但我们自己设置了广告单元，并围绕它做了一些测试。

I think we did, but we put our own ad unit, and we did some testing around it.

然后我们正在逐步推出。

And then it's we're easing it.

在某个时候，我们其实可以允许更多这样的功能，但性能实际上被限制了。

And we and and at some point, we could actually allow more of that, but the performance is actually capped.

把产品重新上线，然后试着思考什么才是真正对用户有益的。

Like, put the product back on and, like, try to think about what's actually good for users.

对我们来说，终极目标是类似Credit Karma的驾驶保险产品。

The the holy grail for us, like, the credit karma equivalent is is drive insurance products for driving.

因为我们了解你的驾驶习惯。

Because we know how you drive.

但谁愿意把自己的数据交给保险公司呢？

And people are who wants to give your data to insurance company?

首先，大多数人并不像科技圈媒体渲染的那样在意数据共享。

First off, most people don't care about sharing data as much as that's like this very kind of like tech bubble press group.

就像大多数消费者认为亚马逊和谷歌已经在记录他们所说的每一句话了。

Like most consumers think that Amazon and Google are like recording everything you say and doing already.

他们说不喜欢这样，但这已经无能为力了。

And they say they don't like it, but like, it's lost cause.

是的。

Yeah.

对他们来说，这已经无能为力了。

It's a lost cause to them.

而且我认为他们其实从未真正关心过。

And not that they, I think, they ever really cared.

据我观察，他们实际上反而享受抱怨这件事。

They actually, I think, enjoy complaining about it, from my perception.

我认为，人们声称关心的事情，他们并没有撒谎，但人们真正会采取行动的是什么？

I think that just well, things people say they care about, they're not lying, but what do people actually take action on?

我们希望对用户坦诚，但如果我们直接这么说，人们可能会觉得这本质上就是一件坏事，而从不深入思考。

We want to be straight with our users, but if we were like, I think people think this is inherently bad thing without thinking about it.

但想象一下，嘿，雅各布，你是前5%的优秀司机。

But imagine if like, hey Jacob, you are in the top 5% of drivers.

你绝对应该把你的保险数据提供给保险公司。

You absolutely should give your insurance data to insurance companies.

这样你会省钱。

You'll be saving money.

而在这里，我们可以帮助你。

And here, we can help you.

如果我们说，大卫，你属于后10%的司机。

And if we're like, David, you're in the bottom 10%.

你绝对不应该提供你的

You should never ever give your

也许该停止开车了。

Maybe stop driving.

是的。

Yeah.

或者，嘿，你有酒后驾驶记录。

Or, Hey, had a DUI.

我们会为你匹配一位酒驾专家。

We're going to match with a DUI specialist.

比如，我不知道。

Like, I don't know.

这确实是个现实问题，那些难以投保的人。

That's actually a thing, the hard to insure people.

这其实是一种用途，我们正在帮你处理数据，虽然这会是一个广告，但想想这有多好。

That's a use like, we're actually helping you with your data and it's gonna be an ad, but think how good that is.

这并不是我们在耍花招或类似的事情，真正来说，如果你是优秀驾驶员，为什么要为那些糟糕的司机买单呢？

It's not us being sneaky or like whatever, like truly, if you're a top driver, why should you be paying to subsidize all the bad ones?

当这种情况发生时，你不会觉得它像广告，但它确实就是。

That will not feel like an ad when that happens, but it will be.

这会像一个潜在客户获取体验。

It'll be like it'll be a lead gen experience.

是的

Yeah.

和信用宝非常相似。

Very similar to Credit Karma.

对吧？

Right?

比如，不同之处在于，你不是获取金融数据，而是收集位置和出行信息。

Like, except for in instead of the finance data, you're taking in location and transport them.

你可能不信，但信用宝现在也正朝这个方向发展，因为他们可以做到。

Believe it or not, Credit Karma is now moving into that same lineup because they can see.

所以，如果你使用信用宝应用，现在也可以用它来追踪驾驶行为。

They so if you use a Credit Karma app, you can do that with the driving now.

还有一些保险资料也是这样，是的。

And for some insurance, profile stuff too, yeah.

他们正在做完全相同的事，这对客户来说并不坏。

They're doing the exact same thing, and it's not bad for customers.

这对客户有好处。

It's good for customers.

所以我们最终希望达到这个目标，而目前只有大约八分之一的家庭是付费用户。

And so we eventually wanna get there, and only about one in eight of our families are paid.

如果我们能实现同等的转化效果，或者更准确地说，如果我们能将免费用户的变现效率提升到八分之一，我们的收入实际上就能翻倍，然后就能形成良性循环：有更多资金用于市场营销，有更多资金投入研发，从而与竞争对手拉开更大差距。

So if we get equivalent performance, or sorry, if we get one eighth performance of monetizing free users, we've essentially doubled our revenue, and then we get the virtuous flywheel of more to spend on marketing, more to spend on R and D, and then just put more daylight between us and our competitors.

你说这是最近才开始的吗？

And you said this was fairly recent.

你们是什么时候开始尝试通过广告来创造收入的？

When did you all start playing with the ads as a revenue generation thing?

大约一年前。

About a year ago.

好的。

Okay.

哇。

Wow.

所以是在公开上市之后之类的。

So like post, after publicly listing and and all of that.

在此之前，唯一的收入来源都是订阅，比如一个应用？

The only revenue before that was all subscriptions, like an app?

基本上是的。

Essentially.

还有一些其他收入，或多或少吧。

It's some others, more or less.

我想是相当可观的收入。

Significant revenue, I guess.

但有趣的是，你会说有十七年吗？

But so it's interesting is because like, would you say seventeen years?

所以十六年都没有广告？

So sixteen years, no ads?

然后是的。

And then Yeah.

我们确实有一些其他东西。

We did have some stuff.

我们曾试图进入保险领域，和全美达合作，但那只是产品的一小部分，里面有一些难看的广告，我们还没能达到理想的状态。

We were trying to move an insurance space with Allstate, but it was sort of, so we did have a sliver of the product that these ugly ads, and we didn't really we still haven't got it to where we wanted.

但就传统广告而言，才刚刚超过一年。

But in terms of, like, the kind of traditional ads, it's only been about bit over a year.

作为一款面向家庭、孩子和青少年使用的应用，你还需要考虑另一个维度。

And you you have another vector that you have to think about as a family focused app that kids and teens are gonna be in this app.

你如何确保整个体验，包括广告，都适合家庭使用？

How do you think about keeping that whole experience, including the ads, family friendly?

我们不会向特定年龄段的人推送广告，并且会剔除大量广告类别。

We don't serve ads to people on a certain age, and we eliminate huge categories.

尽管仍有一些广告漏网，但我们有一支专门团队在监控。

We still had a few that have crept in, but we we have a team that looks at it.

我们在公司内部的Slack频道开设了一个公开渠道，供员工举报广告，目前情况还不错。

We have a public Slack channel or public within the company Slack channel for reporting ads, and it's been okay.

这真的促使团队下定决心去清除那些愚蠢的横幅广告。

It really challenged the team to, like, get rid of the dumb banners.

是的。

Yeah.

而且我们过去也通过出售数据来实现盈利。

And a lot of what we also want to do is we did use to monetize by selling data.

据我们所知，从未发生过任何滥用数据的情况，但这种做法还是引起了媒体和监管机构的关注。

We never had a single instance of misuse as far as we can tell, but that kind of got the target on it from press and regulators.

所以我们停止了出售原始数据。

So we stopped selling raw data.

但事实上，这一切只是为了帮助人们精准投放广告。

But basically, all that was doing though, was helping people target ads.

我个人喜欢精准广告，这是一件好事。

I personally like targeted ads, and it's a good thing.

我常开玩笑说，唯一让我讨厌广告的时候，就是我老婆用电脑时，我却收到瑜伽裤的广告。

I kinda joke that the only time I don't like them is like, if my wife's using the computer, I get a yoga pants ad.

这并不是来自位置信息。

That's that's not from location.

我认为这通常是通过IP地址和NAT技术，而不是其他方式，是的。

I think that's usually just IP address and NAT stuff, than it is, yeah.

精准广告是好的。

Target ads are good.

比如，我更喜欢相关的广告，实际上如果我搜索某个产品但还没准备购买，我知道接下来一周会不断看到它的十个竞争对手的广告，这还挺好的。

Like, I'd rather relevant ads, and I actually kinda like if I search for a certain product, and I'm not ready to buy, that I know for the next week, I'm gonna get their 10 competitors.

所以我们想做基于位置的广告，希望推动基于位置的广告发展。

And so we wanted to do location based we wanted to empower location based advertising.

因此，我们会与第三方签订合同，让他们获得我们提供的去标识化数据，这与匿名化数据不同。

And so we would do these contracts with third parties where they would get our they would get de identified, which is different than anonymized data feeds.

比如，我们会过滤掉GPS点，尽量排除敏感地点，但人们会说，理论上，有人可能黑入数据或违反合同，

So it's like rod like, the GPS points we'd filter out, like we try to filter out sensitive locations, but people say, well, in theory, someone could, I guess, hack the data or breach your contract and

你是说从第三方那里？

From the third party, you mean?

是的

Yeah.

然后弄清楚你是谁，做坏事。

And then find out who you are and do bad things.

理论上确实如此，但让第三方拥有你的数据，比如谷歌，是非常普遍的。

And in theory, that was true, but it's very normal to let third parties have your data like Google.

你们就是个例子。

You you guys as an example.

比如，有百分比。

Like, have percents.

这是一个很好的例子，说明更重要的是感觉和精神，而不是字面上共享了什么。

It's a good example of how it's more the feeling of it and the spirit of it than it is the literal, like, what's being shared.

我知道，Foursquare的SDK过去也是这样盈利的。

I knew, like, Foursquare's SDK used to monetize that way as well.

有很多公司都会以这种方式盈利。

There's a lot of companies that would monetize that way.

这个决定只是因为觉得不值得为了那点收入去承受公关压力、解释工作等等吗？还是说有其他原因要停止这样做？

Was the decision just like it wasn't worth the PR and, like, the explaining and all that stuff for the revenue, or what was the decision to stop doing it?

所以我们能够稍微调整方向，从更长远的角度思考，这其实又回到了广告业务上——我们的用户根本不在乎，但我们不想一直靠这种负面舆论、监管机构和FTC的审查来混日子。

So we were able to pivot a bit and then think longer term, which actually comes back to the ads piece where our users never cared, but we didn't wanna keep getting away with a bad press and regulators and FTC and all that.

实际上，我们合作的那些公司都没问题，但就是一直面临持续的审视，而我们并不需要这种压力。

Like they were actually, all the ones we worked with were fine, but it was just kind of constant scrutiny that we didn't need.

我们的观点是，我们可以转向做更大的事情，以不会引发这种敌意的方式进行。

And our view is that we could move into doing bigger things in ways that would not spark that ire.

所以我们和一家做聚合数据的公司达成了合作，而聚合数据和去标识化数据是完全不同的概念。

So we have one partnership with a company that does aggregated data, and aggregated is very different than even de identified.

我用的词有点专业了，但比如……

I'm getting a little esoteric in the words, but like

不，这对这个受众来说很好。

No, it's good for this audience.

他们很懂。

They're good.

他们很好。

They're good.

是的，但那些包含姓名的原始数据，一直以来都是绝对禁止的。

Yeah, But to kinda go down those as complete raw data with names, which has always been a no no.

去标识化是指原始数据，但没有姓名或电子邮件。

There's de identified, which is raw data, but no names or emails.

匿名化是指数据被混淆，理论上你完全无法追溯到任何个人信息。

There's anonymize, which is scrambled in a way where in theory, you truly can't find anything period, period, period.

而聚合则是更进一步，我们将你和大量其他数据合并在一起。

And there's aggregated, which is even further, which is we're kind of lumping you together with a whole bunch of stuff.

因此，我们现在与一家名为Placer.ai的公司合作，通过他们销售我们的聚合数据以获取洞察。

So we now have a partnership where we do sell our aggregated data for insights through a company called Placer dot ai.

我想我们每年从他们那里获得大约两千万美元。

I think we get like 20 something million a year from them.

基本上，如果你是房地产行业的，我来举一个使用案例。

And it's basically if you are a re I'm gonna give one use case.

你是个零售商。

You're a retailer.

我想知道人们是从哪里来的。

I want to know where people came from.

他们来到了我的商店。

They came to my store.

他们开了哪条路？

What roads did they drive on?

他们住在哪里？

Where do they live?

我们使用聚合数据。

We do it aggregated.

所以你是按人口普查区块计费的。

So you bill it is by census block.

所以72%的人来自这个人口普查区块，其他的人来自这个区块。

So 72% of people come from this census block, and whatever come from this one.

但根本没有原始的GPS数据之类的，我们还是被媒体抨击。

But there's no raw GPS line, all that, and we still had the press take swipes at us.

这实际上让我对苹果的运营方式感到沮丧。

This is actually something that frustrates me about the way Apple operates.

他们的隐私政策非常脱离现实。

Their privacy stuff is very ivory tower.

就好像在我们的广告追踪、SKAdNetwork之类的东西中，绝不可能有任何数据泄露一样，这种设计太过理想化；而现实中，从海量数据中找到可能泄露的零星信息，其价值远不如Facebook为寻找这些可能的泄露点而构建整套基础设施。

It's like there can't be any possible way for one bit of data to leak in our, like, ad tracking, you know, SKAdNetwork and things like that are so ivory tower design versus, like, the practicality of a needle in the haystack is not worth a Facebook building a whole infrastructure around finding a needle here and a needle there that might happen to leak.

是的。

Yeah.

与此同时，苹果如此限制性的做法，实际上是在激励人们绕过其系统，去侵犯其他平台的隐私机制。

Meanwhile, actually, incentivizing more privacy, violating systems outside of Apple, right, by being so limiting.

如果我们真的担心数据问题，我觉得我们正逐渐走向奇点和集体意识。

If we're truly worried about data, like, I I think we are kinda moving to the singularity and the hive mind and all that.

你真正该害怕的是，这些科技巨头掌握了所有数据，并拥有完全的控制权。

You should be really scared that these giants are the ones that have all of it with complete control.

而且，你难道不希望看到一些较小的公司组成反抗联盟，共同使用和分享它吗？

And, like, don't you want, like, kind of the rebel alliance of smaller companies kinda all using it and sharing it?

所以这里存在着非常奇怪且武断的区分。

So there's a there's a very weird and arbitrary distinctions.

比如，如果所有数据都留在你的体系内，那就完全没问题。

Like, somehow, if it's all within your bucket, it's completely fine.

但只要你一连接到别人，就被视为极其危险和高风险。

But the second you do a hook to somebody else, you're somehow being super dangerous and risky.

公平地说，另一方确实有一定道理：在每一个新兴行业，都会出现一个无法无天的西部时代。

To be fair to the other side, there I think there is some truth that in, like, every new industry, there's a wild west emerges.

所以我记得第一次，大约在2011年或2012年，我见到了迪士尼的二号人物。

And so I actually remember the first time, like, I met with a number two guy at Disney in like 2011 or '12.

那时我们都还很年轻。

And we were just young at the time.

我们当时非常非常小。

We were very, very small.

我们开发了一个非常酷的工具，直接去了迪士尼。

We built this really cool tool where we actually went right to Disney.

你看，这是一个人。

Like, look, here's a person.

看好了。

Watch.

看，这是他住的地方。

Look, here's where they live.

我们一边走路一边观察排队情况，但那时根本没有任何监管。

And like we walk and watch the lines, but it was before there's any regulation at all.

我们甚至没想过后续的影响，比如，这个人去了家里，这就是他住的地方，但当时这一切都还很新。

And we weren't even thinking about the downstream implications like, well, look, here's the guy going to the house and like, this is where he live, but it was just new.

所以我认为一些监管是有道理的，但通常它发生得太早，反而常常有利于大公司。

So I do think that some regulation makes sense, but oftentimes, it happens too early, and it just really often favors the giants.

而且我们现在其实正处于一个对我们有利的阶段，因为我们能请得起所有的律师，是的。

And and we're at the point actually kinda helps us because we can we can have all the lawyers and yeah.

处理吧。

Deal with it.

对吧？

Right?

600个人，很多律师。

600 people, lots of lawyers.

你能搞定的。

You can figure it out.

对吧？

Right?

应对它。

Navigate it.

是的。

Yeah.

确实如此。

It's true.

这些真的非常有趣。

It's it's these are, like, really interesting.

因为我觉得这里面有很多我不知道的东西。

Because I think there's a lot of I don't know.

你提到卖数据这件事，这挺有意思的。

I it's funny you mentioned the selling the data or the selling the data story.

这非常

It's very

因为

because

因为我一直在思考我们面临的这个问题。

because I thought about this problem for us.

我之前没意识到，但我早些时候写过，这属于《权利法案》的内容。

And, I didn't know I've written it in Bill of Rights, but this is something I wrote down early.

我们根本不会做那种事。

It's like, we're never gonna do that.

并不是因为我在这方面有多道德，而是因为我明白这很可能会成为媒体关注的问题，我只是不想制造这种诱惑。

Not because, like, I'm all that moral about it, but because I knew it'd probably be a press issue and, like, I just didn't wanna make it a temptation.

比如，如果我们开始出售所有RevenueCat的数据，我明天就能轻松赚到双倍收入，但我不想让我的客户为此操心，我们绝对不会这么做，你知道的。

Like, we could still I could probably find double our revenue tomorrow if I, like, we started selling all of the RevenueCat stuff, but I don't want my customers to have to think about that, and we're just never gonna it, you know?

好的。

Okay.

没有绝对的对与错。

There's no right, right, wrong answer.

我们一直强调的是，我们会始终保持透明，任何我们推出的数据产品，我都会自己使用。

What we have said is we will always be transparent, and that any data product we have, we'll use it myself.

而且我从未弹出过任何提示，实际上我还主动开启了跨应用跟踪。

And I have, I never popped up, and I actually opt in to cross app tracking.

比如被广告定向推送。

Like being targeted with ads.

我总是会点'允许'。

I'm always like, allow.

拜托了。

Please.

我并不是什么大人物，但因为我有数百万用户的用户基础，关注我的人确实比普通人多得多。

And I'm not like a big deal or anything, but like I have certain more people more people after me than the average person just from like the user base of millions.

我经常收到各种奇怪、疯狂的死亡威胁，我真的有。

I get all sorts of weird, crazy death threats and things, and I genuinely do.

但我并不担心我的数据被泄露，因为那个疯子根本不会去买数据。

But I'm not worried about my data being out there because the the crazy dude is not gonna be buying data.

我只是不觉得。

I I just don't

这需要很大的主动性。

It takes a lot of agency.

发一条推文很容易。

It's easy to send a tweet.

但真正去追查某个人，可要难得多。

It's lot harder to, like, you know, actually hunt somebody down.

哪个没有数百万美元的个人能想办法购买原始数据？而且我根本不知道该怎么操作。

What individual without millions of dollars could find a way to buy raw data, and then I just don't even know how you'd go about that.

我首先得成立一家假公司来购买它，然后我还得实施欺诈。

I'd have to set up a fake company first off to buy it, and then then I'd have to commit fraud.

然后我还得成为一名数据工程师，而我并不是，或者找一些

Then I'd to become a data engineer, which I am not, or find some

我发现自己经常在内心进行这样的对话，虽然不直接相关于这个问题，但比如像我不确定。

I find myself having these conversations often internally too, not related to this exact question, but around like I don't know.

也许是我们正在签署的一份合同，律师说我们不能签这个东西。

Maybe it's a contract we're signing, and it's like, the lawyers say we can't sign this thing.

然后你会想，我们能不能只是把所有可能导致问题的条件概率列出来？

And it's like, can we just like line up the conditional probabilities, like, would have to all fire for this to like go wrong?

总之，我支持数据开放。

Anyway, I'm pro data.

我认为我们应该被允许并开放去做更多。

I think we should be allowed and open to do more.

如果我们没有如此宽松地使用数据，不是指个人数据，而是像版权数据这样的东西。

If we didn't have as as liberal and not necessarily personal data, but, like, let's say, like, copyrighted data, for instance.

如果我们所处的社会不是信息如此开放，我们就不可能在过去的几年里创造出或整合出如今这样的大语言模型世界，当然，这其中还涉及一些法律问题。

And, like, we would not have the world of LLMs that we've invented or been able to put together in the last few years if we didn't have a society where, like, information was pretty and then there's, obviously, there's questions about the exact legalities.

但我认为，我们可以都同意，大语言模型和人工智能的诞生总体上是一件好事。

But, like, I think we could all argue it's probably a net good thing that this thing was created being, like, LLMs and AI.

对吧？

And right?

如果我们当初扼杀了它，可能根本不会发明出这么好的东西——我忘了是在哪儿听到这个的。

If we had strangled it, we might not have ever ever invent there's this really good I forget where I heard of this.

有人把Llama 3跑在了Windows 3.1系统上，这意味着我们早在九十年代就本可以拥有它了。

Somebody got Llama three running on like a Windows 3.1 machine, which implies like we could have had this since the nineties.

你明白我的意思吗？

Do you know what I mean?

理论上是这样。

In theory.

但理论上，如果你能运行模型的前向传播，你就可以训练它。

But in theory, I mean, if you can run the model forward pass, you could you could train it.

你只需要比当时可能存在的更多计算机。

You just need more computers than probably existed at the time.

你或许可以只是不断猜测权重，直到找到正确的那个。

You could have maybe like just guessed weights until you found the right one.

但不管怎样，是的，我不确定。

But but anyway, yeah, I don't know.

这是一些有趣的伦理问题。

It's they're interesting ethical questions.

而且我认为我们的用户也在思考这个问题，像RevenueCat这样的公司也一样，我们一直专注于订阅模式。

And and I I think something our users are thinking about, and like RevenueCat too, like, we're quite aware of the we've always been focused on subscriptions.

我仍然认为，如果有人愿意为你的产品付费，这可能是最好的方式，听起来你也这么认为。

I still think that's and it sounds like you too, that's probably best if somebody will pay for your product.

也就是说，这是赚钱的最佳方式。

Like, that's the best way to make money.

这是因为非常契合。

It's because it's very aligned.

我不知道。

I don't know.

只是不想在这方面施加压力。

Just to not not exactly pushing on that.

在说‘最好’的时候要小心，你到底想做什么？

I'd be careful about saying best and, like, what are you trying to do?

对你来说什么最好？

What's best for you?

因为我不认为有一种方法适合所有情况。

Because I I don't think there's a one size fits all.

甚至一些更注重增长黑客的公司，我认为一个非常有效的订阅策略是，不是完全免费增值，而是给用户一个试用体验，这本质上就是一种隐性的高级服务。

And even some of the more growth hacky companies, like, I think a very valid subscription strategy too is, like, not exactly freemium, but it's basically give people a teaser, and there's been your implicit premium.

鸟类，它们销售——我的意思是，这并不是一种放之四海而皆准的做法。

Birds, they sell I mean, it's not a one size fits all thing.

所以对我们来说，这非常合理，因为我们有一个非常自然的用户参与平台，并且可以顺利进行追加销售。

So I for us, it made a lot of sense because we had a very, very, very natural engagement platform with the upsell.

但如果我们没有呢？

But what if we didn't?

而且

And

对。

Right.

是的。

Yeah.

我想这取决于是否足够多的人愿意付费。

I guess it's conditioned on that there's enough people who will pay.

对吧？

Right?

你的应用能够存在、盈利并持续运转，比采用某种特定的盈利方式更重要。

It's better that your app exists and is profitable and is is sustainable than it is that you use like one specific monetization method over another.

对吧？

Right?

是的。

Yeah.

没错。

Exactly.

你到底想做什么呢？

There could be what are you trying to do?

你的模式有什么具体独特的特点？

What are the what are the specific unique attributes about your model?

因为我觉得，如果我们真的能把这个保险项目做起来——我对此很兴奋——也许我们可以彻底放弃订阅模式，把所有东西都免费提供，只要我们能成为下一个全美保险或州立农场，把一切免费送出去。

Because I think at some point, if our if we get like this insurance thing working that I'm excited about, like, maybe we get rid of the subscription model altogether and we give everything away for free because if we can be the next, like, Allstate or State Farm, give it all away for free.

是的。

Yeah.

在别人之前先主动压缩自己的利润空间。

Erode your own margins before somebody else does.

对吧？

Right?

是的。

Yeah.

或者只是巩固我们的影响力，这其实是我们讨论过的事情，但属于公司不同发展阶段的问题。

Or just like just cement our reach, and that is actually something we've talked about, but it's sort of different stages of the company.

你会调整你的策略，而我们早期无法 monetize 数据或做保险，是因为那时还没有足够的用户基础。

You'll shift your strategy, and we couldn't have monetized the data or done insurance early on because you didn't have the user base first.

只有当你拥有庞大的用户基础时，这才行得通，而且你必须自筹资金。

It only works when you have a big user base, and you have to fund yourself.

所以，是的，这一步步发展起来了。

And so, yeah, this evolved.

而且我们已经走了十八年，它仍在持续演变。

And we've we're we're we're eighteen years in, and it's still evolving.

所以，我本来想就这一点问一下。

So I was gonna ask just off of that.

就像，是的。

Like, yeah.

我的意思是，这些其实都是只有当你在这个行业待得足够久之后，才有资格去思考的问题。

I mean, these are kind of, like, problems you only get the privilege to think about when you've been in it for this long.

我不得不想，真正能坚持十七年的CEO，哪怕在那些成功的案例中也寥寥无几，而听上去你依然充满活力和创意。

And I have to think, like, there's not too many it's hard to find CEOs even in, like, breakout successes and stuff that goes seventeen years, and it sounds like you you still got energy and ideas and stuff.

像你这样长期投入这个项目，感觉怎么样？

Like, what has it been like committing to this project for as long as it has?

是什么让你一直保持动力，继续憧憬着要成为一家保险公司呢？

And, like, what keeps you, like, still energized and excited to dream about becoming an insurance company, for example?

在早期，我只是那种咬紧牙关、绝不放弃、非常执着的人。

In the early days, really just someone who kinda lockjaw, don't quit, and very persistent.

钱对我来说是个很大的因素。

Money was a big factor for me.

说实话，我小时候并没有太多钱。

Honestly, didn't grow up with a lot of it.

他们并不穷，但关键是稳定，把这种稳定固定下来。

They weren't weren't broke, but, like, stability and locking that in.

而且，是的，随着我的财富水平提高，把资金从公司抽离后，维持下去变得很难。

And, yeah, as my wealth level has increased and put money out of the company, it's it's hard to keep it going.

说实话，很少有人，不管你喜不喜欢埃隆·马斯克或史蒂夫·乔布斯，对他们来说，钱从来不是目的，而是源于一种深层的驱动力——也许是想赢，也许是想让世界变得更好，又或者只是单纯想创造东西。

Mean, very few people, like, whether you love Elon Musk or Steve Jobs or whatever, like, they it was never about the money for them, or it was just about this deep, deep, deep drive to maybe it's to win, maybe it's to make the world a better place, maybe it's just to build stuff.

我其实没有那种驱动力。

I I actually don't have that.

所以这并不是我一生唯一的选择，我只是偶然进入了这个行业。

So this is not my forever and only thing, and I kinda fell into it.

但我对我们的团队和投资者有着深厚的感情。

But I have a lot of loyalty to our team, to our investors.

所以我一直说，只要我觉得自己还有必要在，我就会留下来。

And so I've I've always said until I really feel like I'm not needed, I'll I'll stay.

最近我确实在努力减少参与日常运营，转而专注于我更热衷的事情，少做那些让我疲惫不堪的工作。

And I have been trying actually get less operational these days, and try to focus on the stuff that I'm more passionate about, and do less of the things that kinda wear me out.

是的

Yeah.

但即使你不离开，作为CEO，你也必须一直这么做。

But even if you're not leaving, like that's the thing you need to, as a CEO, be doing all the time.

对吧？

Right?

你必须不断进行分层和杠杆化，反复这样做。

You like always need to be layering and or leveraging and repeatedly.

因为这样你才能腾出时间去产生那些疯狂又愚蠢的想法，花时间在这些事情上。

Because then it gives you the time to like have the crazy stupid ideas and things, and time to spend time on that stuff.

而且随着公司规模变大，这会变得越来越难。

And it gets harder and harder as you get bigger.

你看，我记得很长一段时间里，Facebook一直在收购很多种子轮和A轮的初创公司，这些公司自己没能独立崛起。

Look at like, I remember for a long time, Facebook was buying up a lot of like seed in series A startups that didn't quite break out on their own.

他们的想法是，让这些极具创业精神、主动性强的人在Facebook内部启动新的产品线。

The idea was that you're gonna get these very entrepreneurial high agency people to start new product lines at Facebook.

有哪个在工作中成功了吗？

Did any break out in work?

比如

Like

Instagram，我想我会说它是其中一个。

Instagram, maybe, would be the one I would I'd say.

他们收购了Instagram。

Well, bought Instagram.

那是11亿美元的收购，不过没错。

That was a 1,100,000,000.0 acquisition, but it yeah.

哦，是的。

Oh, yeah.

是的。

Yeah.

他们支付的不仅仅是团队的价值。

They paid for more than just the team.

是的。

Yeah.

只是给他们预算去开发新产品。

And just letting them giving them a budget to build new products.

我认为实际上没有哪个产品能独立成功。

And I I don't think any actually made it in their own right.

我认为他们把一些非常优秀的人才输送到Facebook的其他部门，但随着公司变大，要做出新东西真的非常非常困难。

I think they got some really good talent into other parts of Facebook, but it's very, very hard for companies to do new things as they get bigger.

这几乎是自然法则。

It's almost a law of nature.

你放慢速度确实令人沮丧，但换个角度看，我们之所以能有今天，也正是得益于这些自然法则。

And it's frustrating how much you slow down, but it's also hey, the the only reason we're here is because of these laws of nature.

因为从某种奇怪的理论角度来说，大公司不是应该吞下每一个机会吗？

Because in in a weird theoretical sense, shouldn't the big guys just suck up every piece of opportunity?

他们拥有大量的资金、动力和用户，但世界根本不是这样运转的。

They have lots of money and momentum and users, but it's it's just not how the world works.

你平时怎么在内部对抗这些自然法则呢？

What do you do to, like, kinda fight those laws of nature internally?

比如权利法案之类的东西。

Like, bills of rights, like, things like that.

我有时候就喜欢直接把东西全炸了。

Do you just, I like to just go and blow things up sometimes.

我规模小一个数量级。

I'm a order of magnitude smaller.

我不是那种天生优雅的领导者。

I am not a natural kind of graceful leader.

我经常就是横冲直撞，我知道理论上有个更好的方法。

I kinda just smash stuff in a lot, and I and I know there in theory is a better way.

我觉得最好的领导方式是清晰明确，做那个鼓舞人心的人，保持冷静，制定宏伟计划，然后执行，让每个人都感觉良好。

Like, I think the best way to lead is clear, like, be the inspirational person, be kind of really collected, and, like, have the master plan execute, make people feel good.

显然，这才是最好的方式。

Clearly, that's the best.

我认为这种人确实存在，一些首席执行官就是这样的人。

And I think it does exist, and some CEOs, I think, are like like that.

我敬佩像里德·哈斯廷斯这样的人，他们是很好的、沉稳的、深思熟虑的、不会过于 erratic 的人。

I look up to people like Reed Hastings and all that, like, good good human being composed, thoughtful, not overly erratic.

不过，我认为次优的方式就是纯粹靠蛮力，直接把事情搞定。

I think the second best, though, is just sheer brute force and just, like, get shit done.

埃隆可能是这方面的极端例子。

Elon's probably the extreme example of that.

但你这是在做价值判断。

You're making a value judgment there, though.

你为什么觉得这不是最好的方式？

Why do you think it's not the best?

我的意思是，用这种心态根本无法建成那么多东西。

I mean, so many things just can't get built with that mentality.

因为我觉得这种方式会耗尽执行者本人，也会耗尽周围的人。

Because I think it wears out the person doing it, and it wears out the people around them.

所以我想，等我们回来的时候，聊聊第三好的方式。

And so I think let me I'll go to the third best when we come back.

第三好的方式，我认为正是大公司失败的地方，那就是官僚主义，他们还以为自己在做对事。

The third best is I think where big companies fail, which is just bureaucracy, where they think they're doing it.

他们在做调查。

They're doing the surveys.

他们确保每个人都发表了意见，但结果是，只要一个人说不，就需要十二个人说 yes，然后一切就停滞了。

They're, like, making sure everyone's heard, but then you get this thing where any one person say no, but you need 12 people to say yes, and then you just grind to a halt.

我一直都是第二类的领导者，现在正努力提升自己，向更高的层次迈进。

So I have been very much a leader in the number two category, and I'm trying to get better at moving up the chain.

这其实是一种抽象的框架。

And and this is a sort of an abstract framework.

我并不是说每个人非得属于某一个类别，但愿意去争取斗争，正是我还能坚持在这里的主要原因。

I'm not necessarily like, no one's in one bucket or the other, but willingness to pick fights is a lot of the reason I'm still here.

你一定也挺享受这个过程的。

You must enjoy it a little bit.

不是在内部。

Not internally.

我不喜欢。

I don't.

我喜欢

I love

我想，至少是结果吧。

the results, I guess, at least.

对吧？

Right?

我觉得人们真的误解了我这一点。

I think people really misinterpret that about me.

我不喜欢冲突。

I I don't like conflict.

我唯一喜欢冲突的时候，是我觉得有人真的该被归入坏人类别。

The only time I like conflict is when I feel like someone's like, I can truly put you in a bad person bucket.

所以，当有人起诉我们，比如专利流氓，我就只是觉得

So, like, when people sue us, like, patent trolls, I just like

当你站在正义一方时。

When you're righteous.

是的。

Yeah.

那时候真的特别痛快。

Then it then it's really fun.

对。

Yeah.

我可以释放我的正义之怒，不管那意味着什么。

I I can can let out my righteous anger, whatever that means.

但不，我实际上非常非常讨厌冲突，我得想出各种办法来避开它。

But, no, I I actually deeply, deeply dislike conflict, and I have to have all sorts of strategies to get around.

这几乎是本能的，我看起来很强硬，所以我不躲闪，但这消耗我的精力，而不是给我能量。

It's it's reflexively, I'm I'm tough, so it's like I'm I'm not cowering, but it doesn't it takes my energy versus gives it.

我非常讨厌解雇员工，我的策略是，如果有人表现得不像高管，我会告诉董事会这样不行。

And I I hate firing people with a passion, and my my strategy on that is if someone's not working like an exec, I will tell the board it's not working.

我现在跟你说这些，感觉就像把自己置于一个不舒服的境地。

I'll feel like I'm telling you now to put myself in an uncomfortable position.

所以当你一个月后问我，问题解决了吗？还是没解决？

So when you ask me in a month, is it fixed or is it not?

不管怎样，我都得进行一场糟糕的对话。

I'm gonna have a shitty conversation either way.

我要么解决了问题，要么就会让董事会觉得我根本不敢做艰难的决定，显得我很糟糕。

I've either either fixed the problem or I'm gonna look really or I'm gonna look really bad to our board that I'm not making tough decisions.

所以我认为，你可以通过故意把自己逼入两难境地来管理自己的心理状态。

So I I think you can manage your own psychology by putting yourself between a rock and a hard place intentionally.

关于‘流程导致的死亡’，我认为这在很多层面都适用。

The point about like sort of process death by process, right, I think applies at a lot of levels.

我觉得我们在讨论实验的时候也稍微触及过这一点。

I think, like, we kinda touched on it with the experimentation conversation as well.

有时候你会过分地忽视直觉、显而易见的东西和常识，反而觉得我们必须用正确的方式去做，因为 stakes 太高了，或者类似的理由。

It's like, I think sometimes you can, to it to to a fault, ignore the the intuitive and the obvious and sort of the common sense in lieu of, like, oh, we have to do it the right way because there's too much at stake or, like, whatever.

结果发现，对于像 Life360、RevenueCat 这类复利型业务，或者那些拥有品牌和可重复模式的公司来说，你几乎不可能因为一个错误的举动就彻底搞砸，虽然作为 CEO 你也许能做到。

And it turns out that, like, especially for compounding businesses like Life360 or, like, RevenueCat or, like, ones that have brands and sort of, like, repeatable motions, it's very unlikely you're gonna do one wrong action that's going to, like you know, maybe you could in a position as a CEO.

但即使你真的做出了离谱的决定，董事会也会介入，公司第二天就能重回正轨——我觉得人们很容易陷入这种误区。

But even then, if you did something so off the rails, like, the board would take care of it and the company would be back on, you know, the next day, I think people fall into a trap of that.

就像是，嘿。

It's like, hey.

我总觉得我很容易搞砸，所以最好完全避免任何风险。

It's like, it's so easy for me to mess up that, like, I should just avoid that at all costs.

对吧？

Right?

是的。

Yeah.

我们还没完全解决这个问题，因为我们确实还有一些事情要做。

We haven't solved that one fully because we we do have a few things.