Why Marketers Need a Product Mindset | Michael Kaminsky (Recast, Harry’s, Away)

所以，你建议营销人员如何重新思考他们的衡量策略，以应对线上线下都能产生转化的全渠道环境，以及线上线下同步开展营销的现实？

So how do you advise marketers to rethink measurement strategies to keep up with not only an omnichannel world where conversions happen both online and in store, also like an omnichannel world where you're marketing online and in store.

但我发现，最有效的方法是企业会做几件事。

But the thing that I've seen that seems to work the best is where companies do a couple of things.

首先，他们接受不确定性是不可避免的。

So one, they accept that there's going to be uncertainty.

他们不可能做到完美。

They're not gonna have perfection.

他们只是坦然接受这一点。

They just sort of embrace that.

我觉得这其实是个非常好的起点，因为它让所有人都从一个健康的心态出发。

And I think that that is actually a really good first step because it gets everyone starting from a healthy place exactly.

接下来，他们会认真思考如何运营营销项目，让规划、预算、实验和衡量全都同步进行，融为一体。

And then the next thing that they do is they really sort of think about operating their marketing program in a way so that planning, budgeting, experimentation and measurement all live on the same plane and happen together.

你好，欢迎来到《增长对话》，我们将与行业领袖深入探讨增长与营销的世界。

Hi, and welcome to Growth Talks, where we dive deep into the world of growth and marketing with industry leaders.

我是克莉斯蒂娜·鲁比诺，Right Side Up 线下业务的总经理，也是本节目的主持人。

I'm Krystina Rubino, the general manager of offline of Right Side Up and your host.

今天我邀请到了迈克尔·卡明斯基，他是 Recast 的联合创始人兼联席首席执行官。Recast 是一家正在打造全球最严谨的媒体组合建模平台的初创公司。

Today, I'm joined by Michael Kaminsky, a co founder and co CEO of Recast, a startup that is building the most rigorous media mix modeling platform in the world.

迈克尔原本是学统计和计量经济学的，职业生涯一直致力于开发软件，帮助人们做出更明智的决策。

Michael was trained as a statistician and econometrician and has spent his career building software to help people make better decisions.

在 Recast，迈克尔正帮助品牌从整体出发，用数据驱动的方式评估线上、线下以及营销漏斗各阶段的表现。

At Recast, Michael's helping brands take a holistic, data driven approach to evaluating marketing performance across online, offline, and every stage of the funnel.

他曾经与 Harry、Mockingbird、Away 和 Rocket Money 等高增长企业合作过。

He's collaborated with high growth companies like Harry, Mockingbird, Away, and Rocket Money.

好了，接下来我很期待和他深入聊聊。

And with that, I'm excited to get into it with him.

迈克尔，欢迎来到节目。

Michael, welcome to the show.

谢谢你邀请我，克莉斯蒂娜。

Thanks for having me, Krystina.

我很高兴能来这里。

I'm excited to be here.

太好了。

Sweet.

那我们直接进入正题吧。

So let's get right into questions.

你这一路都特别关注数据。

You've been way into data throughout your career.

在创办Recast之前，你是在Harry's剃须刀公司内部工作的。

Right before you started Recast, you were in house at Harry's Razors.

你是从零开始搭建分析体系的吗？

Were you there to build the analytics infrastructure from the start?

还是说你接手了原有的流程，觉得可以做得不一样？

Were you refining prior processes that you inherited and were like, I wanna do this differently?

你在那里的经历是怎样的？

What was your journey like there?

是的，以上几点我都涉及了一些。

Yeah, so a little bit of all of the above.

我是团队里第一个专职做数据分析的人，但说实话，Harry's可能是我待过最注重数据分析的团队了。

I was the first dedicated analytics hire on the team, but I will say that Harry's is probably it's one of the most fully analytical teams that I have ever worked with.

公司创始人是一群前咨询顾问，团队里有大量经验丰富的前咨询背景的人。

It's founded by a bunch of ex consultants, very deep bench of ex consultants.

他们的市场团队里也有很多前咨询顾问，都是特别擅长数据分析的人，这对我的工作来说简直是理想环境——因为我希望打造一些工具，让其他人也能自己做分析，而且是以更聪明的方式，而那里的人对这种想法特别接受。

A lot of their marketing team was actually ex consultants, they were very deeply analytical people, which is a wonderful environment for me because I wanted to build tools that other people could use to do their own analysis and do it in a smart way, and I had an amazingly receptive audience there.

所以我刚开始做的很多工作，其实是想搞清楚：该怎么去衡量这件事？

So a lot of the work that I was doing, especially when I was getting started, was trying to learn like, how do you think about measuring this?

怎样才能更准确地理解业务，然后利用我在计量经济学和因果推断方面的背景，开发出更好的工具，帮他们更高效地做分析，确保我们提出的问题是正确的，并把数据交到真正需要做决策的人手里。

What is the right way that you to think about the business and then building the tools to help them do that in better and more efficient ways using my background in econometrics and causal inference to really make sure that we're thinking about these questions appropriately and then putting the data into the hands of the people that need to make the decisions.

这确实很有道理，尤其是因为等会儿我们会谈到一些非常棒的、全面的归因方法，比如最后点击归因。

Yeah, that makes a ton of sense, especially because, well, we'll get into really wonderful comprehensive attribution methodologies like last click a little bit later.

但我一直很好奇，是什么让人决定自己出来创业，觉得这个问题其实更普遍？

But I'm always curious as to what makes people strike out on their own and be like, actually, I think this is a broader problem.

我想创办一家公司来解决这个问题。

I want to create a company that helps solve it.

有没有那么一个瞬间，或者某个迹象，让你意识到这其实是个更广泛的问题，需要像Recast这样的公司？

Was there like a moment or an indicator when you were like, oh, this is a broader problem and there's a need for a company like Recast?

毕竟，市面上已经有不少媒体混合建模平台了。

Because there are obviously other media mix modeling platforms out there.

做归因分析的合作伙伴更是多得数不清。

There's a million attribution partners.

那你是什么时候决定，不行，我得做点不一样的？

What was the point where you were like, no, I wanna do something?

我觉得，和很多创业者一样，我当初踏上这条路，是因为以为这事儿会比实际简单得多。

Well, I think like many entrepreneurs, I only embarked on this journey because I thought it was going to be a lot easier than it actually turned out to be.

我严重低估了背后要付出的工作量。

So I severely underestimated the amount of work that was going to be involved.

但我觉得，在过去十到十五年里，有两个领域的发展特别值得关注。

But I think, over the last ten to fifteen years, there's been a real rise in sort of two different fields that I think are really interesting.

一个是因果推断。

One is causal inference.

这是我特别关注的一个领域，简单说就是认真思考如何用统计和数据去发现世界的真实规律。

So this is a thing that I care deeply about, but it's sort of thinking really hard about how we can use statistics and data to learn true things about the world.

随着学术界出现可重复性危机，因果推断这个领域的重要性越来越凸显——很多曾经广为流传、甚至登上新闻头条的著名研究，结果发现根本是假的。

And the salience of causal inference as a field has grown a lot as we've seen things like the replication crisis in academia, where it turns out that a bunch of sort of very famous studies and papers, some of which you have probably seen in like headlines, turned out to just be like false results, right?

那些结论根本就不是真的。

They weren't actually ever real.

这促使科学界很多人开始深入反思：我们到底哪里出错了？

And so it caused a lot of people in the scientific community to start thinking a lot harder about, okay, where did we go wrong?

怎样才能用更好的方法来改进，确保我们从数据中得出的结论是真实的？

And what are the better methods for improving that and for making sure that the things that we think we're learning from data are actually real.

关于这个话题，学界出了很多文献，也展开了大量讨论，我一直觉得特别有意思，也跟我个人高度相关。

And so there's been a lot of literature and a lot of discussion about that and I've always found that really fascinating and really personally interesting to me.

其次，技术上也出现了很多突破，不是生成式AI那种，虽然那也很有意思，而是某些统计建模方法的进步，让我们现在能构建和估算出比十五、二十年前复杂得多的模型。

And then secondly, there's been a lot of technological innovations, not generative AI, although that's very interesting as well, but around certain types of statistical modeling that have allowed for the building and estimation of much more complex models than what was possible fifteen, twenty years ago.

这两点结合起来，让我觉得现在正是从头开始重新审视媒体组合建模问题的绝佳时机——把最近才发展起来的这些新技术和新经验全都用上，看看能不能在这个领域取得真正的突破。

And those two things combined made it seem like, okay, now is a really interesting time to try to reapproach the problem of media mix modeling from the ground up, taking all of these lessons in new technologies that have been really that have been developed really recently and seeing if we can apply them to this problem and really have any breakthroughs.

再加上市场本身的需求很明显：对很多企业来说，营销是最大的支出项目，哪怕只是稍微提升一点效率，都能带来巨大的成本节约和利润增长，所以这个问题真的值得去攻克，因为它直接关系到很多企业的生死存亡。

And so it's really those things plus obvious market demand like for many many businesses, marketing is their largest line item, and so if you can make that even a little bit more efficient, it results in huge amounts of dollars of value and so it's just feels like a problem worth tackling just because it can be so make or break for so many businesses in terms of their profitability.

是的，完全说得通。

Yeah, no, that makes a ton of sense.

说到过去十五到二十年间，除了传统的归因方法之外出现了其他机会，我真的很惊讶，都2025年了，居然还有那么多公司依然依赖最后点击归因，尽管这种方法在衡量很多东西时都存在明显缺陷，尤其是对品牌初期影响力的根本性误判。

And to your point about there being other opportunities that have evolved in the last fifteen to twenty years besides traditional attribution methodologies, it's incredible to me that in the year of Our Lady Taylor Swift, 2025, many companies still rely on last click attribution despite its limitations in measuring many things accurately, not least of all top of funnel impact.

在你看来，这种归因方式最大的盲点是什么？为什么会这样？营销人员又该怎样重新思考归因模型，才能获得更全面的视角？

What are some of the biggest blind spots in your words that this approach creates and why and how should marketers rethink attribution to get a more holistic view of their efforts?

这个问题其实特别棘手。

Well, this is the thing that that I think is really tricky.

我先退一步说，我对最后点击归因持强烈批评态度，接下来我会详细讲讲我的理由。

So taking a step back, I'm a big critic of last touch attribution and I will get into some of those criticisms.

但我也得承认，对很多企业来说，用最后点击归因其实没问题，也合情合理。

But I do wanna say that for lots of businesses, last attribution is fine and it makes sense.

我女朋友开了一家翻译公司，她只在谷歌上做广告。

My girlfriend runs a translation agency, and she advertises on Google.

她就只在谷歌上投广告。

It's the only place that she advertises.

对她来说，最后点击归因简直完美适用。

For her last attribution, works Like it's perfect.

而且你的广告组合主要是付费搜索。

And there are lots Your mix is paid search.

你的广告组合就是付费搜索。

The mix is paid search.

所以衡量起来特别简单。

So it's it's pretty easy to measure.

但哪怕再往上走一两个层级，如果你只在一二个数字渠道投广告，那你也还是个小到中型的企业。

But even even at, you know, one or two levels above that, if you're only advertising in one or two digital channels, you have a small, medium sized business.

在这种情况下，最后点击归因完全够用，你根本不需要别的工具，而且你能明显感觉到——这些公司花在Facebook上的钱，银行账户立马就涨了，因果关系非常直接。问题在于，当这些公司发展壮大，从中小型企业升级成中大型企业时，过去在小规模阶段特别好用的这套方法，就突然不管用了。

Last touch attribution works, you don't need anything other than that and you can feel it, you know, these businesses, they spend money on Facebook, they see their bank account going up, like the connection there is actually pretty tight, the problem comes when, you know, those businesses graduate then to being like a higher mid market, mid size large businesses and all of a sudden this tool that they that worked really well when they were smaller, no longer works well.

所以这一点很重要，我们要记住，这些建议并不是放之四海而皆准的，你得清楚自己处在哪个阶段，才能选择最合适的衡量方式。

And so that's a thing that I think it's important to keep in mind, like this is not one size fits all advice, and you really have to think about where on that journey you are as to what type of measurement approach is going to make the most sense.

说到这里，对于那些已经进入下一阶段、却还在用最后点击归因的品牌，我想我们这些听这个播客的人，都清楚它的局限性。但归根结底，最后点击归因天然偏向那些最靠近购买环节的数字渠道。

So that being said, for these brands that have sort of gotten to that next stage, still on last touch attribution, I think, you know, we're all at least the people who are listening to this podcast are gonna be very familiar with these limitations, but touch based attribution, excuse me, by definition, touch based attribution is biased towards the most bottom of funnel digital channels.

也就是那些能追踪到点击、离成交最近的渠道。

The ones that have clicks that are observable and the ones that are closest to the point of purchase.

从名字就能看出来，最后点击归因本来就是偏向这种渠道的。

It's almost by definition in the name of last touch attribution that that is what you're going to be biased towards.

对。

Yeah.

再说一遍，如果你的广告渠道只有付费搜索和Facebook或Meta，也许问题不大，但一旦你开始拓展到中段和前端的广告，这种衡量方式就注定会有偏差。

And again, if your media mix is only paid search and Facebook or Meta, maybe that's fine, but as the media mix expands and you're doing more mid funnel, top of funnel advertising, that measurement method is just going to be inherently biased.

没错。

Yeah.

我觉得大多数公司出问题的地方在于，市场团队其实明白这一点，但不幸的是，他们早就让高管和财务团队习惯了依赖最后点击的报告，把它当成唯一真相。

And I think the place where most people, most organizations get themselves into trouble is that the marketing team sort of understands this, but unfortunately, they've trained their executive and finance team to like look at the touch base, the last touch reporting tool and use that as a source of truth.

结果他们突然得回头纠正一堆东西，大家都懵了，这真是个很糟糕的处境。

And now all of a sudden they have to like walk a bunch of things back and everyone is getting confused and that is really a bad place to be.

是啊，我觉得吧，说到底，人们怎么想就是现实，对吧？

Yeah, it's to me, I mean, listen, perception is reality, right?

比如为什么Meta在媒体投放中这么普遍？这背后肯定有原因，对吧？

Like if people get there's a reason that meta is so prevalent in terms of its prevalence in the media mix, right?

我认为，它其实是最容易快速启动的平台。

It's the easiest platform to spin up, I would argue, frankly.

说实话，在优化方面，我觉得它甚至比谷歌还容易上手。

I actually think it's easier to spin up than Google in some ways in terms of optimization.

当然，这事儿我也能从正反两面说。

I can argue it either way.

我发现，不管对客户、潜在客户、团队成员、合作伙伴还是供应商，只要我足够坦诚，我的成功就来自这儿，对吧？

I've found that the more honest I am with clients, prospects, team members, partners, vendors, whatever, I've found that that is where my success comes from, right?

坦诚地跟人沟通，直接说：其实你现在还没准备好做线下的广告。

Having those honest conversations and telling people like, hey, actually, you're not ready for offline media yet.

你还不适合做播客。

You're not ready for podcasts.

你还不适合做YouTube代言。

You're not ready for YouTube endorsement.

你还不适合做流媒体音频。

You're not ready for streaming audio.

为什么？

Why?

这些是你在媒体组合中会遇到的挑战。

Oh, well, these are the challenges that you're gonna run into in your media mix.

你有没有优化过落地页的转化率？

Also, have you ever like conversion rate optimize your landing page?

没有啊，行，明白了。

No, no, cool.

别在这一块儿上折腾了。

Yeah, don't do it on this channel.

但我觉得，营销人员往往喜欢把事情搞复杂，特别是当你之前在规模更大的初创公司待过的时候。

But I think there is like, there's a tendency for marketers to want to add complexity to the mix, especially when you might've had experience at a more scaled startup.

比如你从一家中型、成熟、C轮以上的公司，跳到一家早期阶段的企业，会不会有个临界点——如果你还没到那个阶段，我们根本没必要谈？

Maybe you're going from something that's mid market, something that's scaled, something that's Series C and you're going to an earlier stage company, is there an inflection point at which you think, wow, if you haven't hit that yet, we shouldn't even be talking.

你在跟潜在客户聊天的时候，会遇到这种情况吗？

Does that come up in your conversations with prospective clients?

经常碰到。

It comes up a ton.

真的经常碰到。

It comes up a ton.

我觉得这里面涉及很多因素。

And I think broadly, there's a lot of factors here.

如果单说测量这块——这正是我的业务领域，营销本身就很复杂，这点谁都躲不开，大家都想找一劳永逸的解决方案。

I think if we're focusing on the measurement side, which is my business, Marketing, I think the tough thing for a lot of people is that just marketing is really hard and there's no getting around that and everyone wants a magic bullet.

所以每个人都觉得，要是能有这个工具、那个技术，我的所有梦想就都能实现了。

And so everyone's like, oh, if only I have this technology or this thing, then I will be able to do, you know, all of my hopes and dreams will come true.

只要一看到‘人工智能’或者‘高级统计’这类词，人们就觉得终于有救了，这些问题都能解决了。

And if it says AI or like advanced statistics, then it's like people like, yes, finally, it's gonna solve all of those problems.

所以这里面有很多情况，我也能理解他们的心情。

And so there's a lot, you know, there's just and I feel for them, right?

因为营销人员面对的问题真的太难了。

Because marketers face an impossibly hard problem.

这个环境本身就有太多因素，让人特别难做。

It's just a really, really hard environment to work in for a lot of different reasons.

所以我懂这种想法和期待，但现实是——我经常跟人聊到这一点：增加这种复杂性，并不是万能钥匙，也不是灵丹妙药，反而会带来一堆新问题。

And so I understand that inclination and that desire, But the truth of it is, and this is again, the conversation I have a lot is that adding this additional complexity, it's not a silver bullet, it's not a panacea, it brings a lot of problems.

当你在测量方法或者营销渠道上不断往上加复杂度的时候，其实情况都差不多。

And so as you are moving up the complexity stack in terms of measurement method or even different marketing channels, right?

这里面有很多相似的地方。

There's a lot of parallels here.

你真得好好想想，这么做值不值得，因为往测量体系里堆更多复杂工具，只会让问题越来越多。

You really need to make sure that it's actually gonna be worth it because adding additional complexity to your measurement stack brings a lot of problems.

会让管理层提出更多问题来问你。

It brings more questions from executives.

还会引来更多审视的眼光，毕竟你在这上面投了更多钱。

It brings more scrutiny because you're investing more into it.

核心问题是，你必须先做好应对这一切的准备。

And really it's mostly about you have to be ready for that.

所以说，那些一年只在少数几个数字渠道花个几百万美元的品牌，其实像媒介组合模型这类工具对你来说没什么太大用处。

And so brands that are spending, you know, I don't know, a couple million dollars a year on a couple of digital channels, it's like something like a media mix model is not gonna be that helpful for you.

模型倒是能跑，也能得出结果，但你的业务规模还没到那个程度，根本产生不了多少价值，不值得为了这点收益投入这么多成本，分散那么多精力，还要承担随之而来的各种麻烦。

Like we can run the model, you'll get results, but you're just not at the scale where it's gonna create that much value that makes the juice worth the squeeze for the cost and the distraction and all of the other things that are gonna come along with it.

所有企业都要权衡取舍，而对大部分这类企业来说，把时间和精力花在重新调整定价、优化产品组合，或者打磨宣传话术上，收益会好得多。

Businesses face trade offs and I think for most of those businesses, their time and their energy is gonna be better spent thinking about pricing differently or packaging, you know, bundling their products differently or reworking their messaging.

对那些还没发展到大规模的小企业来说，把时间花在这些事情上才是更划算的投资，因为只有这样才能帮你把业务规模推上一个新台阶。

All of those things are for those smaller businesses that aren't at, you know, huge scale yet, that's just a better place to invest your time, because that's what's gonna unlock the ability to get to the next level of scale.

对，我完全同意这个说法。

Yeah, I couldn't agree more.

我觉得那时候你简直是在提前借未来的麻烦来用。

I think there's a It's like you're almost looking to borrow the problems of the future at that point.

如果你已经找到了产品市场契合点，网站做得好，产品出色，转化漏斗也顺畅，那当然再好不过。

Like you're hopefully if you have product market fit, you have a great site, you have a great offering, your funnel is good.

所有这些条件都满足，你的总潜在市场够大，各项业务指标也都健康，迟早你会发展到面临这些问题的阶段。

All things are true and your TAM is big enough and all of your business metrics are sound like you're going to scale to a point where you will hit these challenges.

但在还没到值得在媒体投放上大笔投入的时候，我的建议是：多投媒体，尽可能把每一分钱都花在能直接带来转化的预算上，因为任何归因、媒体组合建模或效果衡量方案，都会增加额外成本。

But until it makes sense from a media investment perspective, my advice to people is usually like spend more on media, literally put as much money into working dollars as possible because any kind of attribution, media mix modeling, measurement, etcetera solution, it's an added cost.

而且这些成本，往往并不主要来自付给供应商的费用。

And again, and the costs often aren't majority in like dollars paid to a vendor.

更多是内部沟通的代价——解释、说明，还有各种协调。

It's often just like internal communication costs and explanations and

这成本不只是付给合作伙伴的明面上的开销。

It's all not like just the hard cost spent to a partner.

说白了，你等于主动给自己找了一堆和CFO、甚至可能还要和CEO反复讨论的麻烦。

It's literally like you're signing yourself up for conversations with your CFO, probably with your CEO, depending on the scale of the company.

说到品牌，尤其是做跨媒体的营销活动时，

And I think that like with brands, especially when it comes to running campaigns across multimedia.

追踪完整的客户旅程，恐怕比以往任何时候都更复杂了。

Tracking the full customer journey has probably never been more complex.

每年我们都这么说，但每年确实也越来越复杂，对吧？

We say that every year, but it also becomes more and more complex every year, right?

如果你只做社交媒体和付费搜索，那还好说。

Like if you're running meta and paid search, that's great.

但如果你同时还投电视、传统广播、直邮，再加上播客，那 marketers 怎么能把线上线下的数据打通，真正衡量出投资回报率呢？

But if you're also running like television, traditional broadcast radio and direct mail, and then you're throwing podcasts in the mix, like how can marketers bridge the gap between offline and online data to measure true ROI?

在你们公司里，最常见的那些坑是什么？就是你一看就觉得，这事儿真没必要这么干。

And what are some of the common missteps that you see across your business to where you're like, could you not do that?

要是你当初直接这么做，生活会轻松多了。

Your life would be so much easier if you just did X.

我先抛砖引玉吧。

I'll seed with you.

我最喜欢的是同时在新加入的媒体渠道上全面铺开推广。

My favorite is scaling multiple channels at the exact same time, fresh to the media mix.

干脆直接上最混乱、最头疼的场景。

Let's just go for peak chaos, peak pain.

这是其中一个吗？

Is that one of them?

这确实是个很好的例子。

That is a good example.

我本来就想说，这正是那种让情况变得特别复杂的做法。

I was gonna say, that's a really good example of a thing that adds a lot of complexity.

我觉得，说实话，在这个圈子里，很多公司都犯了一个很常见的错误——虽然作为一个数据科学家，这话听起来可能有点疯狂。

I think, you know, I, in this world, I think that a lot of companies go wrong by honest, and like, this is gonna be sound maybe crazy from someone who is like a data scientist professionally.

我一直在跟数据打交道，但总是对数据期望太高。

I've always been in the data, but expecting too much of data.

有些问题，数据根本回答不了，你得真正接受这一点。

There are questions that you cannot answer with data, and you sort of really need to deeply accept that.

我们做营销，我在Recast负责营销团队，经常思考到底该怎么衡量效果。

And like, you know, we run marketing, you know, I sort of run the marketing team at Recast and we think a lot about what are the right ways to measure.

有很多东西我们根本测不了。

And there's just a lot of things that we can't measure.

我们知道哪些东西就是测不了。

We know what we aren't gonna be able to measure.

也许能多拿点数据，但成本太高了，最后还是得扔掉。

Maybe we could get more signal, but it would be a lot of it's just like, we're just gonna throw that stuff out.

我们只看大方向，是不是有起色？

We're gonna look at directional signals, does it seem like it's working?

然后根据这个去优化，等规模再大一点，可能才有别的测量方式合适。

And then we're gonna use that to optimize until we get to that next scale where maybe some other measurement solution makes sense.

我觉得很多品牌要是能这么想，效果会更好。

And I think a lot of brands, they would do better by adopting that thinking.

人们总说要数据驱动，意思就是每个决定都得靠数据说话。

People are like, oh, we need to be data driven, which means every single decision we make needs to be based on data.

我觉得这种想法完全不对。

And I actually think that that's totally wrong.

很多决策你完全可以、也应该基于战略、对市场的直觉，以及其他不一定完全依赖数据的因素来做。

A lot of decisions you can and should make should be based on strategy and your feel for the market and other things that aren't necessarily based solely on data.

如果你营造一种文化，认为每件事都必须有数据支撑，那最后你只会去优化那些容易衡量的东西，而不是真正对业务最有帮助的。

And if you build this culture where like everything has to be based on data, then you end up optimizing towards the things that are easy to measure instead of the things that are actually best for your business.

所以我希望更多公司能真正接受这一点：我们这里会有一些不确定性，但我们会大胆下注，因为我们相信它，而且我们能大致判断它有没有成功。

And so that is actually like a thing that I wish more companies really embraced is like, we're gonna have some uncertainty here, we're gonna take this big bet because we believe in it, we think that we're gonna be able to like roughly measure like, did it work?

这事儿比面包盒大吗？

Is it bigger than a breadbox?

有时候，这已经是你能做到的最好程度了，完全没问题。

And that's sometimes the best that you can do and that's totally reasonable.

所以我建议更多公司认真思考：哪些决策我们非得靠数据，哪些其实靠直觉就够了。

And so I would push more companies to really be thoughtful about what decisions do we absolutely need data for and what decisions can we just make with gut feel anyway.

所以我觉得，营销团队更需要诚实地面对自己，也诚实地告诉管理层：听着，这个我们确实测不了，但我们打算用一小部分预算做个有依据的赌注，这完全合理。

And so that is a thing that I think more marketing teams need to be like honest with themselves and with their executive team about about being like, look, we're not gonna be able to measure this, We're gonna make a calculated bet with some small part of the budget and that's gonna be totally reasonable.

是的，完全同意。

Yeah, I yes.

绝对支持，一百万次同意。

A million percent yes.

因为当一个组织过于依赖数据时，另一个问题就会出现：你开始扭曲数据，让它迎合你想讲的故事，而不是真正倾听数据本身。

Because the other problem that you run into when you get too data centric as an organization is that you start introducing this tendency to make data fit the story you want to tell as opposed to listening to the data.

还有一点，就是把信号误当成噪音。

And then the other thing is mistaking signal for noise.

或者把噪音当成信号——其实这两种情况都可能发生。

Or mistaking noise for signal like that is I mean, actually, yeah, you can go either direction.

我们都见过这种事，但一到这儿我就忍不住要较真了。

We've all seen it happen, but it's like, that's where I just get into them.

你这是在给自己制造一个自我实现的预言。

Oh, you're just, you're setting yourself up for like a self fulfilling prophecy.

这根本不是理想的状态，尤其是对我而言，一个挺有意思的问题是：那些从纯线上直销模式发展起来的品牌。

And that's just not, it's not an ideal situation, especially like, one of the interesting challenges to me is the brands who have expanded from being purely DTC, like pure play online.

他们卖的是商品或服务，而且只通过直接面向消费者的渠道销售。

They're selling a good or a service and they're selling it only direct to consumer.

后来开始涉足数字平台、电商平台，还有线下实体店，比如Casper、Warby Parker、Harry's这些品牌。

Then getting into both like digital, like marketplace retail, physical retail, like your Casper's, your Warby Parker's, your Harry's.

一旦进入实体领域，传统的归因模型就开始失效了。

When you expand into the physical realm, especially traditional attribution models start to break down.

我一直认为，这是营销人员必须重新思考的转折点——不仅要重新规划营销策略，还得重新设计衡量方式。

And that also I've always viewed as an inflection point for marketers to think, okay, I've got to rethink my marketing strategy, obviously, but I also have to rethink my measurement strategy.

那你会建议营销人员如何调整他们的衡量方法，以应对线上线下都能产生转化的全渠道环境，以及线上线下都在做营销的现实？

So how do you advise marketers to rethink measurement strategies to keep up with not only an omnichannel world where conversions happen both online and in store, but also like an omnichannel world where you're marketing online and in store?

确实如此。

Yeah.

这是个很好的问题。

It's a great question.

我花了很多时间思考这个问题。

And I've spent a lot of time thinking about this.

而且再说，这里也没有什么简单的答案。

And then again, this is another place where I don't think there's an easy answer.

但我看到最有效的方法是，公司会做几件事。

But the thing that I've seen that seems to work the best is where companies do a couple of things.

首先，他们接受不确定性是必然存在的。

So one, they accept that there's going to be uncertainty.

他们知道不可能做到完美，而是坦然接受这一点。

They're not gonna have perfection, they just sort of embrace that.

我觉得这其实是个非常好的起点，因为它让所有人都从一个健康的心态出发。

And I think that that is actually a really good first step because it gets everyone starting from a healthy place exactly.

接下来，他们会重新思考如何运营营销项目，让规划、预算、实验和评估全都同步进行、融为一体。

And then the next thing that they do is they really sort of think about operating their marketing program in a way so that planning, budgeting, experimentation and measurement all live on the same plane and happen together.

那么，具体会怎么操作呢？

And so what are some examples of the way that this might work?

好的。

Okay.

我们有个假设，用的是科学方法——认为这个电视广告活动会对品牌产生长期影响，虽然我们可能没法准确衡量，但至少应该能带来一倍的投资回报率，带动线下销售。

We have a hypothesis, right, thinking scientific method here, that running this TV campaign, you know, will have long term effects on our brand, which we don't think we'll be able to measure, but should at least drive a one x ROI and in store purchases.

所以我们要在两个月内大力推一把。

So we're gonna do a big push for two months.

我们会做个前后对比。

We're gonna do a before and after comparison.

就是看看广告上线前和上线后的情况，只要发现收入至少增长了x倍，我们就觉得效果不错，达到了预期。

So like look at, you know, before we launch the campaign and after the campaign, and as long as we see that revenue goes up by at least x, then we're gonna feel like, okay, we got to a good spot.

然后他们会想：我们要做广告，也要做效果评估，这是我们的衡量方式。我们知道有些影响没法量化，但没关系，我们会设定一个合理的阈值，把这种不确定性也考虑进去，接着就执行、推进。结果可能达标，也可能不达标，但这些都会成为我们三个月后做决策的依据。

And then so they think about, okay, we're gonna run the campaign, we're also gonna do the measurement, here's how we're gonna evaluate it, we know that there's gonna be effects that we can't measure, we're gonna be okay with that, we're gonna set the threshold that takes into account that strategy and then we're gonna operate it and they'll do it and they'll run and maybe it'll pass or maybe it won't and they'll use that to decide what they're gonna do three months from then.

我觉得这种思路特别聪明——把假设、实际营销动作、实事求是的评估方式全都整合在一起，不夸大可能性，又环环相扣，感觉特别通透。

That I think is like a really smart way to think about operating these campaigns where everything comes together, where there's the hypothesis, there's the marketing that you're actually doing, there's some way of evaluating that isn't sort of over promising on what's possible and that comes together and that feels really nice.

接下来，你还可以用同样的方式在其他渠道重复这个流程，别的渠道可能更容易做结构化测试，但核心逻辑是一样的。

And then you could think about repeating that with different channels, other channels are easier to test in more structured ways, but they bring that all together.

所以这不是说，有个分析团队在那边盯着市场团队，指手画脚。

So it's not like, oh, there's an analytics team that's over there that's sort of like looking over the marketing team's shoulder.

我们更倾向于一起推进，让业务、数据测量和媒体投放同步优化，所有事情都在一起协同进行。

It's more like we're gonna do this all together so that we're continuously optimizing the business and the measurement and the media, it's all happening jointly.

我们每期节目中间都会玩个小游戏。

We always do a little game in the middle of our episodes.

这完全谈不上什么科学。

It's not scientific at all.

我们管它叫A/B测试，但其实这可能是任何营销人做过最不科学的A/B测试了。

We call it the AB test but it's like the least scientific AB test that hopefully any marketer will ever run.

我会问你一些工作相关的问题，还有一些轻松有趣的问题。

But I'm gonna ask you some work related questions and some fun questions.

然后很‘科学地’，你只能选一个。

And very scientifically, just have to pick one.

好。

Okay.

你可以解释你的选择，也可以不用解释。

You can substantiate your answer or not.

我先给你来个简单的吧。

Let me give you I'll give you a layup first.

Mac还是PC？

Mac or PC?

Mac。

Mac.

挺有意思的，我本来以为搞数据的人会选PC，但我自己也很早就换用Mac了。

It's funny because I always assume p like, data people are gonna say PC, but I I also switched to Mac a long time ago.

我用Mac已经十年了。

I've been on a Mac for like ten years.

我现在已经完全离不开它了。

So I'm I'm very I'm deep in it now.

我懂。

I get it.

塔可还是汉堡？

Tacos or burgers?

这个可真难选。

Oh, this one is so hard.

汉堡。

Burgers.

对。

Yeah.

我超爱一个美味的奶酪汉堡。

It's love a good cheeseburger.

这真的是我的安慰食物，说实话。

It's really it's my comfort food, honestly.

要是我心情不好，就会选奶酪汉堡配巧克力奶昔。

A cheeseburger and a chocolate milkshake is like if I'm having a bad day, that is that's what I go for.

我知道，我特别能理解那种‘吃点好的安慰自己’的感觉，不管是什么方向。

I, you know, I'm a go I love an eat your feelings moment, like in either direction.

其实，我们今晚就吃汉堡。

And actually, we're having burgers for dinner tonight.

所以我现在对它们更期待了。

So I just got extra excited for them.

挺有意思的，因为这类问题，我总觉得没什么对错之分。

It's funny because with these, I'm always like, there's no right or wrong answer.

这个问题确实也没有什么对错之分。

And with that question, there really is no right or wrong answer.

但有时候我们问起来，我反而觉得，这个答案其实有个更对的。

But like sometimes when we ask them, I'm like, no, the right answer to that one.

比如，选 Slack 还是邮件？

Like, you know, Slack or email?

我都讨厌死了，根本不知道选哪个。

I hate them both so much.

我实在拿不定主意。

I don't know which one to choose.

哦，这答案也挺好。

Oh, that's a good answer too.

是的

Yeah.

我还是选邮件吧。

I'll go with email, I guess.

哦，行吧。

Oh, let's okay.

今天Slack挂了，所以我也谈不上有多讨厌它。

Slack's down today, so it's not having especially negative feelings.

我刚才过来的时候，正好看到这个问题，心想：今天不就是我们录播的日子吗？

I literally I literally was looking at that question as I was coming up, and I was like, this is the day that we are recording.

Slack今天全球宕机了。

Slack is down across the world.

简直乱成一团。

And like, it's chaos.

我之前跟人聊天，还说这情形就跟以前大家在办公室，Wi-Fi突然断了一样。

I was talking with somebody earlier and we were like, this is like when you everybody used to be at a physical office like the Wi Fi would go down.

哦，是啊。

Oh yeah.

结果啥正事都没干成。

And like bro nothing got done.

所以我现在觉得，今天我其实没怎么高效。

And so I'm like, I don't know like I've been hugely productive today.

但反过来讲，我又觉得怪怪的。

But then on the flip side of it, like, I just feels like weird.

就是有种跟所有事都脱节的感觉。

Like it's an, I don't know, it just feels like I'm disconnected from everything.

所以 yeah，这大概说明我太依赖 Slack 了，真不太健康。

So yeah, that's probably a note of my reliance on Slack, which is really unhealthy.

出去自驾游，还是宅家里？

Road trip or staycation?

宅家里，没得选。

Staycation, easy.

我也超喜欢宅家度假。

I love a staycation too.

我特别喜欢

I love

宅家度假。

a staycation.

哦，我现在简直太享受了。

Oh, I'm loving this.

我心想，咱俩居然有这么多共同点，我之前完全不知道。

I'm like, we have so much in common that I didn't know about.

付费推广还是自然流量？

Paid media or organic?

当然是自然流量啊。

I mean, organic, obviously.

如果你能不花一分钱在付费推广上就把生意做起来，那你就占了大便宜了。

If you can build a business without ever without ever spending $1 on paid media, you're in a much better position.

那咱们的想法就不一样了，我特别喜欢可复现的实验，付费媒体恰恰就是可以反复测试、还能规模化的玩法。

See, that's that's where we diverge because I I love a repeatable experiment and paid media is nothing if not repeatable and scalable.

你确实能人为影响自然流量，但这个过程的时间节奏根本不由你说了算。

I can't you can impact organic, but the time frame is not up to you.

对，你说的没错。

That's that is true.

确实是这样。

That is true.

不过对的，说真的，谁会不喜欢……等等，之前那些……就是那些……

But but yeah, mean, hell, like, who who wouldn't want like a nice little remember when the, like, the other

你听说过的那些体量特别大的企业。

businesses that are like huge that you've heard of.

我就会问，你们在付费媒体上砸了多少钱？

And I'm like, how much do you spend on paid media?

他们就会说，从企业创办到现在，已经砸了足足三百万美元。

And they're like, over the entire lifetime of the business, they spent like $3,000,000.

我觉得这太惊人了。

And I'm like, that is amazing.

你有几亿客户，却从来没怎么花过付费推广的钱，这简直是个奇迹。

Like, you have hundreds of millions of customers and you've never spent money on paid like that's a miracle.

这太了不起了，简直是商业上的奇迹。

It's an incredible it's a business miracle.

当我跟别人说，我们是一家营销咨询公司，也是一个平台的时候。

When I tell people, like we are a marketing consultancy, we're a platform, right?

人们总是问，你们是怎么跟企业达成合作的？

And people are always like, well, you know, how'd you agree with business?

实际上，95%都是靠口碑传播。

It's literally like 95% word-of-mouth.

大家一听就愣住了，说：什么？

And people are just like, wait, what?

你们可是全国最大的增长营销咨询公司之一啊。

Like you're one of the largest growth marketing consultancies in country.

我们确实做得不错。

I'm like, yeah, we do good work.

这简直太不可思议了。

Like, it is wild.

真的特别不可思议。

Like it's absolutely wild.

所以啊，如果你的产品已经和市场匹配了，那确实是好事。

So it's yeah, hey, if you have product market fit, it's good thing.

如果你的产品已经和市场匹配了。

If you have product market fit.

没错。

Yeah.

数字断联，还是始终在线？

Digital detox or always connected?

始终在线。

Always connected.

我也是这样。

I know me too.

我女朋友老是叫我把手机放一边。

My girlfriend is always yelling at me to like, leave my phone somewhere.

我才不听呢。

I'm I will not do it.

我们出去的时候，我坚决不把手机留在家里。

I refuse to leave my phone at home while we go out.

我真的做不到，实在做不到。

I just cannot be I cannot

等等，把手机放下？

Wait, leave your phone.

什么？

What?

不行。

No.

什么？

What?

现在还搞1880年那一套？

Like it's eighteen eighty?

不可能。

Like, no.

不行。

No.

绝对不行。

No.

没门。

No.

想都别想。

No.

我们不会把手机留在家里的。

We're not leaving our phone at home.

我整个过程中一直都在想，这到底在哪啊

I would literally the entire time be like, like, where

这正是我做不到的原因。

That that is why I can't do it.

因为我一想到就简直要崩溃了。

It's because I'm like, oh my god.

手机、钱包、钥匙。

Cell phone wallet keys.

这才是正确的方式。

It is the way.

说实话，我觉得餐厅就该是这种让人抓狂的样子。

Like, yeah, I believe I believe in a restaurant being like, my god, crap.

我真的做到了。

Like, I did it.

好吧。

Okay.

线上营销还是线下营销？

Online or offline marketing?

天哪。

Oh, man.

现在这两者之间的差别已经小到几乎没了。

There's like so little barrier between them today.

确实如此。

It's true.

我们只做线上营销。

I we do only online marketing.

所以我选线上营销，因为这正是我做的。

So I'll say online marketing because that is what Fair.

我就是做这个的。

I do.

有意思的是，我的背景是多渠道、全渠道的数字传播策划。

It's funny because my background is multichannel, omnichannel, digital, like communications planning.

当我说‘哥们儿，要是我明天要创业，至少一年内我绝对不会做线下营销’的时候，大家总是很惊讶。

And people are always surprised when I'm like, Yeah, dude, if I was starting a business tomorrow, I wouldn't do offline marketing for at least a year.

他们就问：‘真的吗？’

Like, they're just like, really?

我会说，是的，哪怕你对哪些播客效果最好了如指掌，

I'm like, yeah, even with everything you know, even with knowing exactly which podcast perform.

但说实话，谷歌广告的规模实在太成熟了，

But I'm like, yeah, dude, like, AdWords is so scaled for a It's

太棒了。

so good.

谷歌广告它真的特别

AdWords is It's so

简直太厉害了。

damn good.

十年前我刚接触营销的时候，最早投入的其实是付费搜索。

I've my deepest tea prior to getting more into offline marketing about ten years ago was paid search.

我会把第一笔、第二笔和第三笔钱都投在这上面。

Like, that would be my first, second and third dollar.

明白了。

Okay.

先聊点轻松的，然后再回正题。

A couple of fun ones and then we'll get back to business.

你喜欢早起还是熬夜？

Early bird or night owl?

我是早起型的。

I'm an early bird.

不错。

Nice.

音乐会还是看电影？

Concert or movie night?

哦。

Oh.

是的

Yeah.

五年前我肯定选演唱会，但现在我年纪大了。

It definitely like five years ago would have absolutely been concert, but I'm old now.

所以现在我 definitely 选看电影。

So now it's definitely movie night.

好的。

Okay.

最后一个，你更希望预知未来还是读心？

Last, would you rather be able to see the future or read minds?

预知未来。

See the future.

读心术听起来太糟糕了。

Reading minds seems like it would be a terrible experience.

哦，确实。

Oh, yeah.