Measuring Marketing Effectiveness

你会不自觉地偏向那些容易追踪的渠道，而这些渠道通常都是漏斗最底部的数字营销渠道。这样一来，那些位于漏斗上游、追踪难度更高的渠道，你就没法分配预算投入进去了。

You're gonna be biased towards the channels that are easiest to track, which are often, you know, the most bottom of funnel digital channels, which means that you won't be able to spend into these other more upper funnel channels that are harder to track.

那你的处境就会变得非常被动。

And that's gonna leave you in a really bad position.

也许只靠Meta和脸书的投放，你能把业务做到一定规模。

Maybe you can scale the business up to some point only spending on Meta Facebook.

但总有一天这个渠道的潜力会被你挖掘殆尽，你得去开拓别的渠道才行。

But at some point, you're gonna tap out that channel you're and gonna need to go somewhere else.

要是你没有一套完善的整体评估方法，这件事做起来会非常困难。

And if you don't have a good holistic measurement methodology in place, it's gonna be really hard to do that.

欢迎收听《数尽其用》，这是ninja cat出品的播客，聚焦营销绩效管理、指标体系与营销效用。

Welcome to What Gets Measured, a ninja cat podcast about marketing performance management, metrics, and effectiveness.

毕竟，只有先完成量化，才能有效管理。

Because what gets measured gets managed.

我是主持人杰克·桑德斯。

I'm your host, Jake Sanders.

作为Recast的创始人，迈克尔·卡明斯基致力于运用先进的技术和方法，解决营销测量中最关键的问题，从而减少营销预算的浪费。

As the founder of Recast, Michael Kaminsky is committed to ridding the world of wasted marketing spend using advanced techniques and technologies to solve the most important problems in marketing measurement.

他曾经担任过高级分析师、数据科学家、分析总监，现在他把这套擅长的数据分析技能带到《Ninja Cat》播客，来聊聊如何衡量营销效果。

He's worked as a senior analyst, data scientist, a director of analytics, and now he brings his spreadsheet of skills to the Ninja Cat podcast to talk about measuring marketing effectiveness.

迈克尔，欢迎来到节目。

Michael, welcome to the show.

杰克，谢谢你邀请我。

Jacob, thank you for having me.

我真的很期待这次对话。

I am so excited to be here.

老兄，你没我这么兴奋，不过咱们节目最后来个兴奋度比拼，看看谁才是真正热血沸腾的那个。

My man, not as excited as I am, but we will have an excitement off at the end of the show to see who's seriously jazzed.

现在咱们直接切入正题。

Right now, we gotta get into it.

现在大家总在谈营销效果，但我先想请你解释一下，营销中的‘效率’和‘效果’到底有什么区别？很多人可能搞不清楚，或者你干脆直接说说看？

There's a lot of talk about effectiveness in marketing, but I wanted to start with you maybe giving me like explaining a bit of the difference between efficiency and effectiveness in marketing because I think a lot of people maybe don't get this or why don't you just go for it?

是的。

Yeah.

我觉得人们之所以搞不清楚，是因为大家其实都挺糊涂的。

Well, I think part of the reason why people are confused about it is because I think everyone is confused about it.

我经常听到‘效率’和‘效果’这两个词被随便用。

I hear these words, efficiency and effectiveness thrown around.

不同的人说它们的时候，意思还不一样。

They're used by different people to mean different things.

我之前咱们聊过，我理解的效率，其实就是用更少的资源做更多的事。

I took a little bit about the conversation that we had earlier about how we sort of have been talking about it, which is efficiency is really this idea about do more with less.

怎么把现有的预算用到极致？

How can you sort of optimize the current dollars that you have?

这当然是个非常重要的概念。

Obviously, a really important concept.

而效果呢，范围就更广一些。

And then effectiveness is sort of a little bit more broad.

没错吧？

Right?

我们要怎么评估营销投入的效果呢？不只是要把每一分钱都花透，还要考虑拓展新渠道、尝试新方向，对吧？

How do we think in general about measuring the effectiveness of our marketing spend, not just sort of cranking down, right, getting every little last drop, but also potentially expanding into new channels and doing new things.

所以还是那句话，跟不同的人聊，这两个词的意思可能都会有点差别。

And so, again, depending on exactly who you're talking to, these words might mean slightly different things.

不过今天我和雅各布对话的时候，我们会这么用这两个词：效率就是把现有资源用到极致，而效果的定义更广一些，指的是如何从整体上最大化营销的影响力。

But I think that for the purposes of today's conversation between Jacob and I, that idea of, like, efficiency being get every last drop and then effectiveness being a little bit more broad in terms of how do we think overall about maximizing marketing impact, that's probably how we're gonna use those terms today.

没错。

Totally.

而且我就在想，这两个词是对立的吗？

And I mean, it just is are they opposites?

比如说，要想有成效，就得投入很多成本，

You know, like, if you wanna be effective, you gotta pay a lot.

但要想追求效率，好像随便做点什么零成本的事也行得通。

If you wanna be efficient, you could just do whatever you want for free.

你明白我的意思吧？

You know what I mean?

我只是在想，如果你真想取得效果，是不是非得花钱不可？

I'm I'm just, like, thinking, like, if you really wanna be effective, do you have to spend money?

但效率会不会其实是一种达成效果的途径呢？

But maybe isn't efficiency a way to get to effectiveness?

我觉得效率确实是效果的一部分。

I I think efficiency is definitely a part of effectiveness.

对吧？

Right?

没错。

Right.

作为营销人员，我想我们最终都希望尽可能高效地做事。

You want you know, I think we probably all want to be as efficient as possible in the long run as marketers.

对吧？

Right?

刚才提到的「长远来看」特别关键，因为有时候要实现整体更高的效率和成果，短期里就得先牺牲一点效率，去积累经验，摸清到底什么方法可行。

The in the long run there is important because sometimes getting to an overall higher level of efficiency or effectiveness requires being a little less efficient in the short term to get learnings, figure out what works.

这个概念非常重要。

Very important concept.

没错。

Absolutely.

但长远来说，我们都希望能尽可能地兼顾效率和成果。

But in the long run, we all want to be as efficient and and as effectiveness as possible.

无论是某个营销渠道还是某条广告，它的效率都会影响你整体效果的提升，帮你最大化自身创造的全部价值。

And the efficiency of some given marketing channel or some ad is definitely gonna be a part of your overall effectiveness journey trying to maximize the total amount of value you're creating in the world.

哇，说得太对了。

Man, just be beautifully said.

不过我们确实很会衡量效率，

I mean, we know how to measure efficiency.

这方面大家的做法都是统一的。

It's we are doing the same thing.

我们产出了成果，顺利推出了产品，结果很不错。而且我们用更少的资源做到了，是用巧劲，不是蛮干。

Output, we're shipping, it's good, and we're doing it with less or we're doing it smarter, not harder.

可要是谈到衡量营销效果，这里面都包含哪些维度呢？

But when we're talking about measuring marketing effectiveness, what what goes into that?

要怎么才能准确拿到那些衡量指标？

How how how are you properly getting those metrics?

你都会从哪些地方着手梳理？

Where do you look?

那具体该怎么做呢？

What do you do?

你说说看。

Go.

你听我说。

Look.

这个问题问得特别好，已经有很多非常优秀的人在这件事上投入了大量时间研究。

This is a great question, and there's a ton of really smart people who spent a lot of time on this.

我个人的看法可能比较局限，因为我对M和M这个话题特别感兴趣。

My, like, personal viewpoint is probably a little bit more narrow because I'm really interested in this particular topic of M and M.

但效果评估其实涵盖很多方面，比如注意力指标，这几年特别火。

But effectiveness can include all kinds of different things, including things like attention metrics, right, which have been really hot in the last couple of years.

认真思考一下受众覆盖面。

Thinking really hard about audience reach.

你到底在跟谁说话？

Who are you talking to?

为什么？

Why?

他们处在漏斗的哪个阶段？

What stage of the funnel are they at?

对吧？

Right?

这些都很重要，得去衡量，然后想想：我们怎么才能打造一个真正高效的完整营销体系？

All of these things are important to measure and start to think about, okay, how are we gonna build a really effective holistic marketing engine?

而我花最多精力的地方，正是效率和效果的交集。

And then where I spend a lot of time focused really is on that very tight intersection of efficiency and effectiveness.

我们花一百块，到底能换来什么？

When we spend a $100, what does that actually get us?

我们能获得什么回报？

Us?

我们该怎么准确衡量这一点？

How do we measure that correctly?

又该怎么用这些数据来做决策？

How do we use that to make decisions?

怎样才能做出明智的投入？

How do we make smart bets?

在营销的效果与效率这个领域，我主要就聚焦在这块。

That's the corner of the world that I tend to operate in in the in the realm of marketing effectiveness and efficiency.

对。

Mhmm.

那我们就直接来聊聊这个吧，你看现在关注度指标是个刚冒出来、特别受追捧的新词。

And let so let's just get into that because, I mean, attention metrics, that's a new sort of darling term that's coming out.

然后不少人听完就会觉得，行吧，好像还挺厉害的。

And then is another one that you're like, okay, cool.

但这背后还得搞一大堆复杂的数学计算才行。

So there's like a bunch of advanced math that now needs to happen.

好多营销从业者一想到这个就一个头两个大。

And a lot of a lot of marketers are like, ah.

他们就躲在自己学戏剧、艺术的背景背后，慌慌张张的，

They're dodging behind their drama and arts backgrounds to be like, uh-oh.

这里面到底要用到多少数学啊？

How much math is there?

要是有人搞不懂营销组合模型，你会怎么解释呢？

If somebody's confused about marketing mix modeling, what do you do?

你有没有一套能短时间讲清楚的说法？

Do you have like an elevator pitch for it?

对。

Yeah.

那先说说营销组合建模这件事，首先它并不是普适所有企业的。

So marketing mix modeling, look, first of all, it's not for everyone.

所以对很多正在听这期内容的人来说，这个方法可能并不适合你。

And so, for a lot of the people who are listening to this, might not be right for you.

但我觉得对于那些年度付费营销预算超过五六百万美元的大型品牌而言，它不仅效果很好，还是营销工具库里非常重要的一部分。

I think works really well and is really important part of the tool set for bigger brands spending $5.10 plus million dollars a year on paid marketing.

但对很多品牌来说，这么做根本不划算

But for a lot of brands, it just doesn't make sense.

实在太复杂了

It's too complex.

花精力做这个根本拿不到对应的投资回报，所以你不该把时间耗在这上面

The return on investment of, like, doing an isn't really there, and so you shouldn't focus your time there.

所以我得先把这点说在前面

So I wanna lay that out front.

也就是说，这个工具我不会随便推荐给所有人。

Like, I wouldn't pitch this tool to everyone.

而对那些可能适用的人来说，你可以这么理解：这套方法体系非常完善，它能持续衡量你各个营销渠道的增量效果，而且完全不依赖数字追踪。

For the people who who it might be interesting to, one way to think about is it's just a really good structured way of continuously measuring incrementality of your different marketing channels without relying on digital tracking.

收听这期播客的大部分听众，应该都很熟悉归因分析或者各种数字追踪的方法。

So most probably a lot of people who are listening to this this podcast are familiar with attribution or digital tracking mess methods.

就是那种在网络上追踪用户行为的模式。

You follow people across the Internet.

用户会点击什么内容？

What do they click on?

用户在下单前会先浏览什么？

What do they view prior to purchase?

之后你再去分配转化功劳。

And then you assign credit.

这种方法在特定场景下勉强能用。

That can work okay in certain circumstances.

但对于那些既有线上又有线下渠道的复杂品牌来说，可能他们的分销渠道也是线上线下并存的，就很难依赖这种方法了。

But more complex brands that have online and offline channels, maybe they have online and offline distribution channels as well, can't really rely on that.

再加上 iOS 14.5 以及其他隐私政策的变化，跨平台追踪用户变得越来越难，你根本没法准确知道用户在购买前点过什么、看过什么。

And then, you know, with iOS 14.5 and other changes to the privacy landscape, it's difficult to track people across the Internet, so you don't really know what they viewed or clicked on prior to purchase.

对的。

Right.

所以这是一种自上而下的统计模型，你只需要分析你过往营销活动的历史数据模式。

Anyways, so is a top down statistical model where you're just gonna look at historical patterns in your marketing activity.

有时候，你在 Facebook 上多花点钱。

Some days, you know, you spend a little bit more on Facebook.

有时候，你在 Google 搜索广告上投入更多。

Some days, you spend a little bit more on Google search.

可能四月份你做过一场 CTV 广告活动。

Maybe in April, you ran a CTV campaign.

可能九月份你大力推过 YouTube，不管那具体是什么。

Maybe in September, you did a big YouTube push, whatever that is.

对吧？

Right?

这些年来，各种变化都发生过了。

There are all of these historical changes that have happened.

然后你得用统计方法，分析这些数据里的模式。

And then you're gonna use statistics to look at the patterns in that data.

四月份你做CTV广告的时候，收入比预期高出多少？得把其他所有营销活动的影响都考虑进去。

In April, when you did that CTV campaign, how much higher was revenue than you would've than you would've expected controlling for all of the other marketing activity.

九月份你大力推YouTube的时候，扣除季节性和其他因素后，额外带来了多少收入或转化？

September, when you did the YouTube push, how much additional revenue or additional conversions were generated controlling for seasonality and everything else.

所以一旦你开始梳理这些数据，就能用统计模型把所有信息整合起来，从而评估营销的真实增量效果，而不用依赖追踪用户跨互联网的行为——这对视频渠道来说特别难，像在亚马逊这种没有点击追踪的销售平台上更是如此。

And so once you start to pull the you can use an as a statistical tool to pull all of those pieces of information together to start to get a read on incrementality without relying on trying to track people across the Internet, which can be very difficult to do for video channels or if you have sales on a channel like Amazon where there's no touch based tracking.

没错。

Right.

营销组合模型其实就是做好预测、优化媒体投放和制定策略的核心。

I mean and so marketing mix modeling is like the heart of the work of getting towards forecasting, of getting towards better media buys, kind of planning strategy.

我的意思是，这些怎么才能体现效果呢？

I mean, how does that all go into effectiveness?

我的意思是，这很明显啊。

I mean, it's it's you it's obvious.

你只要算一算，就能得出结论。

You just do the math and you figure it out.

等等，我刚才是不是自己把问题答了？

I mean, did I just answer my question?

对啊，算一算，自然就明白了。

Well, I mean, do the math and figure it out.

这背后显然隐藏着不少复杂性。

It's obviously masking a fair amount of complexity.

没错，确实如此。

Think That's true.

核心想法是，如果做得好，营销组合模型能成为一个集中点，把所有相关信息都整合起来。

Idea here is that when done well, a marketing mix model can be one central location that pulls together all of the information that you have.

正如你提到的，它还能用于预测、规划，最重要的是优化预算分配。

And then exactly as you mentioned, can be used for forecasting, planning, and most importantly, budgeting optimization.

那会怎么样呢？

What would happen?

如果我们从这个渠道抽走资金，转投到另一个渠道，会发生什么？

What do we believe would happen if we take money out of this one channel and reallocate it to this other one?

考虑到我们预期的季节性波动，一年中什么时候花钱才能最大化效果？

When should we spend in the year in order to maximize effectiveness given our expected seasonality?

也许我们夏天多花点，冬天少花点。

Maybe we wanna spend more in the summer and less in the winter.

这种模型能帮你系统地思考这些问题。

And this sort of model can really help you think through that in a structured way.

我猜很多公司之前都是靠一个乱七八糟的谷歌表格，一堆列，一堆假设，然后琢磨‘如果我们这么干，那样干会怎样’？

I think prior to setting up an a lot of companies, they have this like crazy Google Sheet, you know, with all of these different columns and then tons of assumptions built in, and they're like, okay, well what if we do this and that?

这种做法内部逻辑不一致，容易导致规划效果不佳。

And it's not very internally consistent and it can lead to suboptimal planning.

随着公司或市场团队逐渐成熟，他们需要一个更系统、更全面的工具，来支持这种高层次的规划与优化，而不是继续在那个混乱的谷歌表格里不断调整各个渠道的销售数据——等渠道一多，那玩意儿根本管不过来。

And so as companies mature or their marketing organizations mature, they want something that's a little bit more structured that has a more holistic view of the world and will allow them to do that higher level planning and optimization, as opposed to just trying to like tweak a bunch of sales in this crazy Google spreadsheet that can be really difficult to manage once you're in a whole bunch of different channels.

没错。

Absolutely.

让大家都达成共识，非常关键。

Getting people on the same page, critical.

你刚才提到 iOS 5？

So you've mentioned iOS five.

是 14.5。

14.5.

哦对，是这个数。

Oh, right.

对。

Yes.

没错。

That's right.

你明白我意思吧？

You know what I meant?

那些在苹果店外通宵排队的人，都觉得他说得不对。

Everyone out sleeping outside of the Apple store is like, he's wrong.

哎，你们还是好好过自己的日子吧。

Well, get a life, friends.

不是的。

No.

我开玩笑的。

I'm just kidding.

你们该干嘛就干嘛吧。

Please do your thing.

该买iPhone就去买吧。

Buy all the iPhones.

接下来要来的，是那个所谓的Cookie末日——第一方数据、第三方数据，全都要迎面而来。

That the coming cookie pocalypse, you know, tracking first party data, third party data, all of these things are coming our way.

然后你还有各种AI，再就是GPT这些工具。

Then you got your AIs, and then you got your GPTs.

你觉得在衡量营销效果这个领域，未来会面临哪些挑战？

What kind of future concerns do you see coming our way in the world of measuring marketing effectiveness?

我的意思是，到底是什么，到底是什么

I mean, what what's what's what's

卡明斯基的观点是什么？

what's the Kaminsky stance?

我觉得很多人都知道，现在因为跨设备追踪的能力受限，或者说是追踪用户跨设备行为的能力大大下降了，不管是基于Cookie还是其他标识符。

I mean, this is a thing that I think a lot of people are familiar with, which is that with the limited visibility and the sort of lack of ability to track people across devices, or let's say reduced ability to track people across devices, and use the different individual identifiers, whether those are sort of cookie based or other things.

过去几年情况已经变得难多了，而且未来几年只会更难。

It's it's gotten a lot harder over the last couple of years, and it seems like it's only gonna get harder over the next couple of years.

这种情况你已经看到了。

You have that.

所以这是第一个问题。

So that's one problem.

另一个问题就是营销渠道太分散了。

And the other problem is just marketing channel fragmentation.

没错吧？

Right?

现在不光有Facebook、Snapchat、TikTok这类封闭生态平台，还有不少新兴渠道正在崛起，比如联网电视、Netflix，还有其他各种视频平台。

So you've got walled gardens, like Facebook and Snapchat and TikTok, and then you've got, you know, other channels that are starting to open up, whether that's connected TV or just Netflix or other video providers.

是这样吧？

Right?

现在营销人员要管理的渠道五花八门，想要把用户在所有这些平台上的行为路径串联起来，已经变得非常困难，而且未来这个难度只会越来越大。

There's all of these different channels that marketers now need to manage and stitching, a user's journey across all of those different platforms has gotten really hard and, again, is only gonna get worse and harder.

所以在这两大难题面前，营销人员必须好好想清楚，到底要用什么样的衡量方法，才能真正评估营销效果。

And so between those two things, marketers need to think really hard about what measurement methods they're they're using to try to get at this concept of effectiveness.

而且如果你同时布局了这么多不同渠道，几乎不可能完整捕捉到一位客户转化路径上所有相关的触达节点。

And if you are operating in all of these different channels, I would say it is very unlikely that you're actually going to be able to see all of the different relic relevant touch points on a customer's potential journey.

总会有些节点被漏掉，可能是因为用户用了别的设备，或是在别人家里登陆之类的各种情况。

Just you're gonna miss some because they're on some device or they're at someone's house or whatever.

你说对吧？

Right?

有些触点存在于某个封闭生态里，这个生态的追踪方式和其他封闭生态还都不一样。

They're in some walled garden that has a different way of tracking than the other walled garden.

要是你所有的预算决策都基于触点归因模型，可你明明知道自己已经漏掉了一大堆触点，那肯定会出问题，没错吧？

And so if you're basing all of your budgeting decisions off of like a touch based attribution model, and you know that you're missing a bunch of touches, well, you're gonna have problems, right?

你会不自觉地把预算倾斜到那些容易追踪的渠道上，这些渠道往往都是靠近转化漏斗底部的数字渠道，到头来你就没法给那些漏斗顶部、追踪难度大的渠道分配预算了。

And you're gonna be biased towards the channels that are easiest to track, which are often, you know, the most bottom of funnel digital channels, which means that you won't be able to spend into these other more upper funnel channels that are harder to track.

那你的处境就会非常被动。

And that's gonna leave you in a really bad position.

可能你只投放Meta和Facebook的广告，也能把业务做到一定规模。

Maybe you can scale the business up to some point only spending on Meta Facebook.

但这个渠道的流量迟早会见顶，到时候你就必须开拓新渠道了。

But at some point, you're gonna tap out that channel and you're gonna need to go somewhere else.

如果没有一套完善的全局衡量方案，开拓新渠道这件事会做得非常艰难。

And if you don't have a good holistic measurement methodology in place, it's gonna be really hard to do that.

这种情况我在合作过的很多品牌身上反复见过，他们最后都卡在这个问题上了。

And I've seen this over and over and over again with a lot of the brands that I work with where they get stuck.

财务部门和高管团队只知道末次互动归因模型这一种方法。

The finance team and the executive team only knows about last touch attribution.

想要做其他调整根本行不通，因为这些团队会说，其他渠道用末次归因法算出来的数据都很差，但这并不代表这些渠道本身就不行。

And then when it comes to doing anything else, they can't because the finance and the executive team is like, well, everything else looks terrible on last touch basis, but that doesn't mean it's actually a bad channel.

只能说这套评估方法本身有问题。

It just means the measurement method is actually bad.

所以这个问题会非常棘手。

And so that can be a real problem.

所以我希望大家在规划未来的时候，可以好好想想：我们要怎么摆脱现在这种全靠末次互动归因模型的局面，这个模型本身就有缺陷，现在问题只会越来越严重。

So as people are thinking about the future, I want them to be thinking about, okay, how do we get out of this world where we're just doing everything on last touch attribution, which has always been flawed and is only getting worse.

对。

Yes.

我现在感觉就像会场后排有个人站起来，慢悠悠地开始鼓掌。

Just I'm I'm feeling like somebody in the back of the room standing up slowly clapping.

你懂的吧？

You know?

但人家连为什么鼓掌都不清楚，因为这个世界已经够复杂了。

But also not knowing why they're clapping or what is to be applauded because there is so much complication already in this world.

所以你就会想，为啥啊？

So you're like, why?

对小公司来说，我是说那些大企业，你们有数据团队。

For smaller folks out there, I mean, enterprise people, you have a data team.

说不定还托了老天爷的福，有专人负责。

You have people that are in there maybe perchance by the blessings of the higher powers.

你们有数据治理。

You have data governance.

有一套自己的数据处理方式。

You have some kind of way that you approach data.

但其他人呢？是不是就只能先拉个表格，凑合着搞？

But for the other people, are they just like kind of like you just got to get a spreadsheet going and kind of cook it up?

而且外面那些资源匮乏、自己打拼的人呢？

And like, I mean, out there scrappy, no resources.

那营销组合建模，或者类似的东西，算是他们该用的方法吗？

Is marketing mix modeling or an approach to something that resembles that?

我的意思是，对他们来说，理应如此。

I mean, of course, it should be for them.

但营销组合建模，是你规模大到靠直觉和经验搞不定的时候才需要的。

But marketing mix modeling is when you reach a scale that you can't do it with your web brain.

你意思是这个意思吗？

Is that kinda what you're saying?

差不多就是这样。

That's that's probably right.

但我认为，小公司、自由职业者，或者任何这类人，其实有个优势——事情对你来说相对简单。

But I think, again, you know, the small companies or solo operators or whoever it is, like, you have an advantage in that things are relatively simple for you.

所以你用个表格就能做很多事。

And so you can do a lot with just a spreadsheet.

只要你在做事的时候多留心就行。

And you just need to be thoughtful about what you're doing.

还有个不错的方法，比如如果要开通一个新的推广渠道，那就先做个前后对比。

And, you know, a good way to do it is like, okay, if we're gonna launch a new channel, like, let's do a before and after comparison.

看看开通这个渠道之后，销量有没有涨？

When we launched this new channel, did sales go up or not?

你可以翻一翻自己的整体数据，多问这类问题，做这类分析就够了。

Like, you can sort of look at your overall metrics, ask those sorts of questions, do that sort of analysis.

这种方法可能比较粗糙，也不完美，但完全没关系。

It'll be rough and, you know, it won't be perfect, but that's totally fine.

在你目前的业务阶段，你只需要这些大致的方向指引，用一张电子表格就能把这件事做得很好。

For the stage that you're at as a business, you want these, like, directional signals, and you can do that pretty well in a spreadsheet.

然后你就可以开始构思一些基础的实验，这些实验搭建起来其实没那么难。

And you can start to think about basic experiments that aren't really that hard to set up.

比如你可以做地域转化测试：连续几周不在国内一半地区的这个渠道投放，看看会发生什么变化。

You can do geo lift tests where we'll hold out half the country from this channel for a couple of weeks and see what happens.

这些实验根本不需要你有统计学博士学位就能操作和分析。

These are all you don't need a PhD in statistics to run and analyze that sort of experiment.

用Excel就能搞定。

You can do it in Excel.

不会完美，但这没关系。

It will not be perfect, and that's okay.

但你要养成这种思维方式：关注真正的增量效果，别只依赖点击归因。

But getting into this mode of thinking about one incrementality, right, which is the thing that actually matters, don't rely only on touch based attribution.

其次，想想怎么设计聪明的实验。

And then two, thinking about how can we do smart experiments?

同样，也不用追求完美。

And again, don't have to be perfect.

粗糙一点、快速上手也没关系。

They can be rough and ready.

咱们把这个渠道的预算提高一个月，看看会发生什么。

Let's boost up spend in this channel for a month and see what happens.

销量有没有达到我们预期的增长幅度？

Did sales go up as much as we expect or not?

这种经验总结真的能给一家企业带来颠覆性的改变。

That sort of learning can be, you know, game changing for a business.

而且我们今天聊的这些内容，其实很多都是面向中小型运营者的。

And since you're, again, we're talking a lot to like sort of the smaller operators here.

是这样的。

Right.

如果你是小企业，做这件事会容易很多，毕竟你只布局了少数几个营销渠道。

Since you're a small business, it's easier to do that because you're only in a couple of channels.

如果只用到三个渠道，那要做一场还不错的实验、总结出经验，需要提前做的假设就少很多。

If you're only in three channels, there's less assumptions you have to make in order to run an okay experiment and get some lessons.

所以我会说，一定要去试试这件事。

And so I would say definitely go do that.

好好利用你规模小的优势，大胆去做这些尝试。

Take advantage of your size to go take these swings.

再说一遍，就算不完美也没关系。

Again, if it's not perfect, that's fine.

先有个大致方向，然后全力推进，赶紧去试下一件事。

Get a directional read, double down, go run to the next thing.

你可以这样快速高效地操作，很多大竞争对手反而做不到。

Like, you can do that sort of stuff and be really fast and efficient in a way that a lot of the your bigger competitors maybe can't be.

确实做不到。

Cannot.

我突然想到一点，就是营销。

And one thing that's hitting me is marketing.

现在的营销太复杂了。

It's so complex nowadays.

这根本不是件容易的事。

This is not this is not easy.

你这么一说，让我觉得我们好像又要回到过去那种时代——只能靠能追踪到的东西来判断方向，比如销售数据，这些你从营销活动末端能直接看到的硬性指标。

And the way you're talking makes it seem like we're headed back to an era where you could only ascertain directions and, you know, kind of trajectories from the things you could track, which were sales, which were like these hard metrics that you're experiencing on the other end internally from your marketing campaigns.

但不知为啥，过去二十来年，甚至更多一点，我们一直被灌输一种想法：嘿，你在谷歌上就能轻松追踪到数据。

But for some weird reason, for the last, like, I don't know, twenty years, you know, maybe a little bit over, you know, fifteen years, we've been given this thought that, hey, you can track it easy here on Google.

我们有办法。

We got it.

在Instagram上也能轻松追踪。

Hey, we can track it easy on Instagram.

你根本不用操心这个。

Oh, you don't have to worry about that.

我们有数据分析。

We got analytics.

想要这些吗？

Want all that?

我们都有。

We got it.

大家都一头扎了进去，没错。

And we all, like, jumped in the boat, you know, and yeah.

亚马逊也一样。

Amazon.

我还有个数据面板可以用。

I got a dashboard.

所以大家就觉得，行吧那没问题了。

And you're like, well, then word.

于是我们就把所有东西都放到这个能量化追踪的平台里。

We're just gonna put everything into the place where we know we can measure.

那你知道你现在拿到的这些流量质量到底怎么样吗？

Now, do you know the quality of the traffic that you're getting?

其实我们开始录节目之前就聊过这件事了。

Just like we were talking before we hit record.

就说付费广告吧，对，流量确实能涨不少，但质量根本不行。

Sometimes paid ads, yeah, you're gonna get a bunch of traffic, but it's not high quality.

有些低成本的引流方式，反而能拿到质量更高的潜在客户线索。但要想发现这点，你得和销售部门紧密配合，把后端的工作都做扎实才行。

Maybe something lower budget, you could get higher quality leads, but you'd only know that if you really worked with the sales department, if you really worked on the back end.

我觉得挺有意思，我们之前把一堆技术都堆在一起，现在又回头说：等等，不对劲。

And I think it's interesting because we've been we just rolled up a bunch of technology and smoked it, and now we're headed back to this place where we're saying, hold on.

你刚才说的那句话，真的让我心头一震。

You know, I I you you just said something that really hit me.

你意思是，要把钱投在那些能保证出数据的地方。

It's like you're going to put money towards the places where you're guaranteed a metric.

我不在乎那是不是一个真正有质量的指标。

I don't care if it's a quality metric.

只要是个指标就行。

It's a metric.

我拿这个去跟股东、高管或者客户汇报的时候，看起来还挺好看的。

And I can bring this to the shareholders or the c suite or my clients and say, it looks pretty good.

你懂我的意思吧。

You know?

你觉得我们是不是正慢慢回到那些技术还没兴起的年代？

I mean, do you feel like we're maybe headed towards a place where we were before all this technology?

就是，你懂我的意思吧？

Like, you know what I mean?

对。

Yes.

当然了。

Absolutely.

没错。

Yeah.

绝对如此。

Absolutely.

你想想，像百事和可口可乐这样的公司，其实并不太依赖电商销售。

I mean, you know, if you think about it, companies like Pepsi and Coke, right, they don't sell a lot on ecommerce.

他们可能根本没怎么认真考虑过最后点击归因的问题，毕竟谁会去 coke.com 买可乐呢？

They have probably never really thought very hard about last touch attribution just because who buys a Coke at coke.com?

我反正从来没去过 coke.com。

Like, I've I've never been to coke.com.

我不确定可乐官网是不是他们的网站。

I don't know if coke.com is even their website.

反正没人这么干。

Like, one does that.

对吧？

Right?

所以这根本就不是个事儿。

So that's not a thing.

我们熟知的那些大品牌，根本就没法去翻Google Analytics的后台，看什么最后点击来源、每次转化成本这些数据。

And so a lot of the biggest brands that we know and are familiar with, right, have been have never had the ability to, like, look at the Google Analytics dashboard and be like, where's my last touch at what's my last touch cost per act Ever.

可乐公司从来就不这么想问题。

Like Coke does not think that way ever.

对他们来说，这根本算不上什么改变。

And so for them, it's not really like a change.

他们还是会继续按老办法做。

They're just gonna continue to do that.

我觉得，正如你提到的，过去十五年，电子商务兴起起来了。

And I think, you know, exactly as you mentioned, the last fifteen years, we saw the rise of e commerce.

我们看到很多优秀品牌，靠着投放谷歌和脸书广告一步步做起来的。

We saw a lot of like great, brands built on the back of running Google and Facebook ads and building brands that way.

这确实挺厉害的。

Very cool.

但我认为这其实是个例外。

But I do think it's an aberration.

我觉得未来是全渠道的，对吧？

I think the future is omni channel, Right?

这里的全渠道，指的是销售渠道的全面覆盖。

And it's omnichannel in terms of distribution.

你会在自己的官网卖货。

You're gonna sell on your.com website.

你也会在亚马逊上卖货。

You're gonna sell on amazon.com.

还会在沃尔玛的网站上卖货。

You're gonna sell on walmart.com.

未来不管出现什么其他的销售平台，你都能在上面售卖。

You're gonna sell on whatever other vendors that are gonna come in the future.

同时也会铺货到线下门店，把所有渠道都打通。

You're gonna sell in store, like you're gonna have that full mix.

我觉得在这种全渠道的环境下，还想追踪每一个用户，记录他们购买前的每一个营销触点，根本是行不通的。

And so the idea that we're gonna be able to track everyone, see every, you know, marketing touch point before they purchase just doesn't make any sense in that world, I don't think.

所以我确实觉得，过去这十五年里，很多正在听这期播客的营销从业者，他们只经历过这个时期，只懂这套逻辑。

And so I do think that, like, there was this fifteen year period and a lot of marketers who are probably listening to this podcast, this is all they know.

你说对吧？

Right?

没错，你从小接触的就是这套玩法。

So it's all you know.

但如果放眼整个营销和广告行业的发展历史，你会发现这十五年其实只是一段特殊的反常时期。

But really, if you look at the whole history of marketing and advertising, it really is an aberration.

所以我觉得，人们会开始重新关注那些早就存在的其他方法。

And so I think people are gonna start turning back to these other methodologies that have been around for a long time.

实验思维，其实从直邮广告最早期就开始了。

Experimentation, right, goes back to the earliest days of direct mail advertising.

人们一直在思考，该怎么设计一个实验？

People have been thinking about how do we run an experiment?

怎么才能准确衡量它的实际效果？

How do we measure the true lift of this?

对。

Right.

所以这些东西，其实都在回归。

So like, this stuff is all coming back.

实验法，还有营销组合模型，都已经有五十多年历史了。

Experimentation, marketing mix modeling also been around for fifty plus years.

这些才是营销人员应该掌握的全面技能。

These are the tools that marketers need to have in a well rounded tool set.

我觉得，数字追踪始终会很重要，对吧？

I think, again, digital tracking will always be important, right?

尤其是那些纯电商或纯数字产品品牌。

Especially for some brands that are pure e commerce or pure digital products.

这确实很重要。

Like that's an important thing.

它永远不会消失，但你得把其他这些工具也纳入工具箱，从不同渠道综合起来，才能真正看清什么对你的业务最有效。

It's never gonna go away, but you need to bring in these other tools into the toolbox to help get that full well rounded triangulated picture from these different sources in order to really know what's most effective for your business.

说得太到位了。

Quote bomb.

讲点实际案例吧，你刚才聊了这么多关于店铺的事，还有那些自上而下的统计模型，给个具体故事听听。

Share some stories because you've been talking a lot about the shop, what goes into those things, top down statistical blivity blops.

听好了。

Look.

来，给我讲个精彩的故事。

Just give me a good story.

你能讲讲吗？有没有哪个客户真正重视营销效果，然后做出了特别棒的改变？

Can you share a story about a client that, like, really took the beast of marketing effectiveness to heart and made a beautiful change?

反过来，你有没有遇到过特别惨的营销效果案例？

And conversely, do you have a marketing effectiveness horror story?

说实话，我遇到的惨案可能比成功案例还多。

I mean, I probably have more horror stories than anything else.

毕竟我跟太多人聊过，时间一长，这种事见得太多了。

Just because that's with a lot of time talking to a lot of people.

典型的例子？

Typical?

你看，我们什么情况都见过，很多其实都在大家的预料之中。

Look, I mean, we've seen all kinds of things and a lot of them, I think, you know, are things that people would expect.

但我确实合作过一个全渠道的客户。

But we've got, you know, I've worked with a client that's omnichannel.

他们既在线上卖，也在实体店卖，之前一直靠‘最后点击’这种归因模型来运营。

They sell online, they sell in stores, and they had been running the business off of effectively a last touch attribution type model.

尽管他们的线上业务在整体业务中占比其实很小，毕竟还要算上实体店和亚马逊这些其他渠道。

Even though the .com part of their business was like a really small part of the overall business once you accounted for in store and Amazon and these other channels.

有意思。

Interesting.

结果发现，在最后点击归因模型里，他们的数字渠道表现简直太好了。

And it turned out that actually some of their, like, in in that last attribution, their digital channels looked amazing.

对吧？

Right?

这些渠道明显是他们最有效的。

Those are easily their best channels.

Facebook 表现特别好，因为它确实很擅长把人引到官网商城。

Facebook looks great because Facebook is really good at, you know, sending people to a .com store.

但事实上，一旦考虑到对实体店销售的影响——那些是没法追踪的——其他线下渠道对整体业务的贡献反而更大。

But it turned out that, like, other offline channels were actually better for the overall business once you accounted for their impact on in store sales, which aren't trackable.

比如广播，人们开车去店里的时候听广播，虽然不太能带来官网销量，但对带动实体店销售特别有效。

So things like radio, right, which people listen to in their car on the way to the store don't drive .com sales very well, but they do drive in store sales very well.

这其实很好理解。

There's like an obvious reason for that.

电视渠道也是同样的道理，对吧？

Same thing with, some of the TV channels, right?

这些渠道主要提升品牌认知。

Which again are more awareness building.

它们能影响消费者在店里选哪个品牌，但不太能直接带动他们去官网购买。

They help someone make a decision about the brand they're gonna purchase in the store, but less likely to drive someone directly to the .com site.

这个品牌因此彻底调整了营销策略，聚焦在真正推动业务增长的渠道上。

And so this brand was able to like really change their marketing mix based on what was actually driving the business forward.

这对他们来说更有利，毕竟现在每个品牌都在做Facebook广告。

And it's better for them because, you know, everyone, every brand is now on Facebook.

那里的千次展示成本已经高得离谱了。

Those CPMs have gotten really expensive.

相比之下，投广播和某些电视渠道、特别是某些联网电视广告的品牌已经少多了。

A lot fewer brands are advertising on radio and certain, you know, at least certain TV channels or certain versions of connected TV.