Thank you for coming.

謝謝

>> Move us all, thank you.

這門課太棒了

Cool, so you guys this is awesome,

前面幾節課我一直關注著

I've been watching the lectures in this course,

講的真是太棒了！

isn't it absolutely amazing the content?

今天只能委屈大家聽我講了

And now you're stuck with me today.

我們正式開始

>> We'll, so how that goes.

在Paul演講的問答環節中

Unlike Paul, when he was talking in the Q and A,

同學們問他如果他現在上大學的話他會學什麼專業？

and you guys asked him what he would do if he was at

他回答說是物理

college today and

我大學時學的就是物理

he said physics, I actually indulged myself.

在劍橋讀物理

I went in, I went and did physics,

我認為物理很棒從中學到的知識

did physics at Cambridge and I think physics is

可以為其他領域所用

an amazing class to give you transferable skills that

今天大家來聽的不是這個問題

are really useful in other areas.

今天講的不是物理

But I guess that's not,

我自費學習了大學網絡營銷課程直接營銷課程

that's not why you're listening to me today.

我從上世紀90年代開始接觸SEO（搜​​索引擎優化）

Like, physics isn't the class, like, so I paid for

我創立了一家紙飛機網站並壟斷了這個市場

college doing online marketing.

在紙飛機這個小眾市場

Direct sales marketing.

有著壟斷地位

I started with SEO in the 1990s.

當你準備創業的時候

I created a paper airplane site.

要從長遠角度考慮一下市場有多大紙飛機網站市場不大

I had a monopoly.

但是就是這個經歷教會了我如何做SEO

In the small niche market of

當時用的搜索引擎是Altavista

paper airplanes globally which you know,

做SEO的方法是

when you want to start a startup also see how big

把白底白字的內容放在網頁下方不明顯的位置

the market could be in the long term it wasn't great.

如果該文本中出現20到30個紙飛機字樣

But what that taught me was how to do SEO.

在Altavista搜索排名中就能佔首位

And back in those days it was Altavista, and

上世紀90年代就是這樣做SEO的

the way to do SEO was to have white text on

很容易上手

a white background five pages below the fold.

我上大學時學的是物理專業

And you would rank top of Alta Vista if you just paper

當時我想紙飛機可以讓我變酷

airplanes 20 or 30 times in that text.

我當時是物理班上最書呆子的了

And that was how you won at SEO in the 1990s.

所以我就創建了雞尾酒網站

Like it was a really easy, easy skill to learn.

我就是在這個時候學習了編程

When I went to college, being a physicist,

那個網站後來發展成了英國最大的雞尾酒網站

I thought paper airplanes would make me cool, and

當Google嶄露頭角的時候上述經歷使​​得我做SEO游刃有餘

I was actually the most nerdy person in

在Google上做SEO時需要考慮網頁排名算法

the physics class, so I created a cocktail site.

需要考慮有多少網頁鏈接到你的網站上

Which was how I learned to program.

此時

And that grew to be the largest cocktail site in

如果雅虎黃頁鏈接了你的網站

the UK.

並且網頁底部也有白底白字的關鍵詞的話

And that really got me into SEO properly when

你的網站就可以在谷歌頁面排名中高居榜首

Google launched.

當谷歌有了關鍵字廣告

So with Google, you had to worry about page rank, and

此時我開始自己做生意

you had to worry about getting links

也就是先從谷歌買下廣告鏈接

back to your site.

再通過會員程序為eBay打廣告賺錢

Which basically, at that stage,

差不多能有20%的利潤

meant one link from the Yahoo directory.

這些經歷讓我率先進入了

Got you to the top listing in Google if you had white

目前大家所說的

text, on a white background, below the fold, as well.

增長或是增長黑客領域

When Google launched AdWords that's when I

在我看來就是網絡營銷而已

really started to learn how to do all my marketing.

利用各種渠道達到期望目標

And that was buying paid clicks from Google, and

這就是我學營銷的經歷

reselling them to eBay for

我本想成為一名物理學家結果成了營銷人士

a small margin of like 20% using our affiliate program.

利用互聯網進行營銷

And that was what really kicked me into overdrive

你們認為什麼對於增長最重要？ 59 00:02:24,350 --> 00:02:26,​​940 你們聽過很多講座了這個話題被討論很多次了 60 00:02:26,​​940 --> 00:02:28,970 你們認為什麼對於增長最重要？

into doing what everyone nowadays talks about is

告訴我個答案

growth, or growth hacking, or growth marketing.

好產品

And in my mind it's just internet marketing.

好產品對的

Using whatever channel you

好產品會帶來什麼呢？

can to get whatever output you want.

顧客

And that's how I paid for college and how I

顧客需要顧客做什麼呢？

ended up going from being a physicist to a marketer and

傳播口碑坐著回答就行

transitioning to the dark side of the force.

有人說用戶留存對的

So what do you think matters most for growth.

用戶留存對於增長最重要

You've had loads of lectures and

目前Facebook有非常優秀的增長團隊

people have said it over and over, so

能成為其中一員我非常榮幸

what do you guys think matters most for growth?

重要的是我們有非常棒的產品

Someone give me an answer.

Facebook的增長團隊發揮著關鍵作用

>> Great product.

因為我們推銷的產品

>> Great product.

全世界人民都想用

I agree, great product.

這點非常贊

What does great product lead to?

一旦讓人們接觸到Facebook 他們就會愛上Facebook

>> Customers.

我給很多創業公司提供過諮詢服務

>> Customers.

我最得意的是與Airbnb的合作我還跟Coursera合作過

And what do you need those customers to do?

我還跟其他一些不如這幾家公司有名的一些公司合作過

>> Spread the word.

有一件事被反複印證

Someone said stay on your site.

那就是如果你看一下這條曲線

Someone said that.

用戶活躍度隨用戶獲取時間的變化

That's it. Retention.

如果你能得到

Retention is the single most important thing for growth.

幾乎平行於x軸的用戶留存曲線

Now we have an awesome growth team at Facebook, and

那麼你的公司就是成功的

I'm super proud to work on it.

你的產品適合一定的市場

But the truth of

用戶數飛漲的大部分公司

the matter is we have a fantastic product.

會使用增長黑客病毒式傳播和其他一些手段

But getting to work on growth at Facebook is

我盡量不罵他們但有時會忍不住

a massive privilege, because we're promoting something

他們的用戶留存曲線沿X軸下滑

that everyone in the world really wants to use.

最終與X軸相交

Which is absolutely incredible.

當我把這個圖展示給別人看時

If we can get people on, and

大部分人說：不錯啊

get them ramped up, they stick on Facebook.

當你開創Facebook的增長小組時

So many times I go advise multiple startups,

Facebook每天有100萬新用戶

my favorite was working with Airbnb, but

或是已經有了5000萬用戶

I've worked with Coursera, I've worked with

有很多人訪問Facebook

other ones that haven't done as well as those guys.

你有大量的數據進行統計

But the one thing that is true over, and over, and

我們把同樣的方法用在了B2B業務增長上

over again is if you look at this curve.

吸引用戶註冊自助廣告服務

The sentiment, the active,

我們用這種方法

versus number of days from acquisition.

來分析我們在該市場中的增長

If you end up with a retention curve that

當時

ascentotes to a line parallel to the x axis.

當我加入Facebook時該產品推出只有3天

You have a viable business, and

推出該產品90天后

you have product market fit for some subset of market.

我們就能用這個方法來計算

But most of the companies that you see fly up,

一位廣告商一年創造的價值

we talk about growth hacking and virality and

我們預測出的數值

all of this other stuff.

是真實值的97​​%

I'm gonna try so hard not to swear, but it'll happen.

我認為觀察用戶留存曲線非常重要

Their attention curve slopes down towards the X-axis, and

我們是這麼操作的

in the end intercepts the X-axis.

大家看這裡

Now when I show this chart to

紅色線表示

most people they say that's all well and good.

使用產品一定時間的用戶數

You had a million people a day in terms of

有些人或是說沒有人

growth when you started the growth team at Facebook, or

有些人一直使用你的產品

you were at 50 million users.

所有的用戶至少有一天使用了你的產品

You had a lot of people joining the site so

如果你的產品推出一年了

you had a ton of data to do this.

使用了366天的用戶數為零

We used this same methodology for

說得通嗎？好的

our B to B growth.

然後你這麼來看用戶數

Getting people to sign up as self-service advertisers,

用戶至少使用了一天你的產品

we used this analysis to understand how much

月度活躍用戶佔多大百分比呢？

growth we were gonna have in that market as well.

顯然第一個月是100%

And in that place,

因為第一個月只要是使用過一天的都算活躍的

when I joined Facebook, the product was three days old.

然後看第31天

And within 90 days of the product launching,

註冊31天后

we were able to use this technique to be able to

月度活躍用戶佔多大百分比呢？看第32天

figure out what the one year value of an advertiser was,

看第33天看第34天

and we predicted it for the first year to 97% of

這樣你就可以--

what the number turned out to be.

假設只有10000名用戶

So I think it's very important to look at

就能得到你的產品的

your attention curve, and this is how we did it.

用戶留存曲線

If you see here.

你就可以判斷其走向

This red line is number of users,

圖的右邊會有一些波動

who have been on your product for

我用的不是真實數據

a certain number of days.

圖的右邊會有一些波動

So, a bunch of people, that should say zero,

不過你能看得出來

a bunch of people will have been on your product,

線是不是平的

all of your users will have been on it at least one day.

如果不是平的就不要談什麼增長策略了

But, if your product's been around a year, or whatever,

不要談什麼病毒式增長不要雇增長黑客

you will have zero users who've been on it 366 days.

專心把產品搞好

Make sense?

因為歸根結底如Sam在本課開始提到的

The curves sensible?

創意產品團隊執行

Cool.

如果沒有好產品

So what you then do is look for all of your users,

執行增長策略也沒什麼意義

who have been on your product one day.

因為增長不了

What percentage of them are monthly active.

我在Facebook內部看到的

100% for the first 30 days obviously,

我在合作過的創業公司中看到的

because monthly active they all signed up on one day.

最主要的問題是沒有好產品但是他們認為有

But then you look at 31.

下一個顯然的問題

Every single user, on their 31st day after registration,

人們反復問的一個問題是

what percentage of them were monthly active, 32nd day,

好的用戶留存率什麼樣？

33rd day, 34th day?

你可能說：我們公司只有5%的留存率

And that allows you,

（Facebook的留存率要好一些）

with only something like 10,000 customers or

所以我們是失敗的

whatever, to get a real idea of what this curve is

當人們問我好的用戶留存率什麼樣時我會生氣

gonna look like for your product.

因為可以自己算啊

And you're gonna be able to tell does it asymptote.

我喜歡這個故事可能有點像“閒篇子”

And it'll get noisy out towards the right-hand side.

我在這裡給大家講講

Like I'm not using real data.

這個故事的後半部分比較有用

It'll get noisy out towards the right hand side, but

這幅三位一體核試驗圖片發佈於

you'll be able to get a handle on,

1950年的Life雜誌上

does this curve flatten out or does it not.

有個叫Geoffrey Taylor的人有同學知道Geoffrey Taylor嗎？

If it doesn't flatten out, don't go and

很棒有人知道

do growth tactics, don't go and

Geoffrey Taylor是一名英國物理學家

do virality, don't hire a growth hacker.

他後來獲得了諾貝爾獎

Focus on getting product market fit.

他能根據這幅圖片

Because in the end,

求出原子彈的威力

as Sam said in the beginning of this course,

這是美國的一顆原子彈

idea, product, team, execution.

俄羅斯發布了相似的圖片

If you don't have a great product,

不過是用因次推論做出來的

there's no point executing well on growing it.

這種技巧

Because it won't grow.

是我在英國讀物理時

Number one problem I've seen inside Facebook for

學到的最好的方法

new products, number one problem I've seen for

其原理就是

startups I ever devised has been, they don't actually

看一下問題中涉及的量綱

have product market fit when they think they do.

如果想計算出能量

So the next obvious question that people ask over and

能量等於牛頓·米

over and over again is okay, so

牛頓等於千克·米/秒^2

what does good retention look like?

所以能量就等於千克·米^2/秒^2

Sure I can have five percent retention but

所以你就要算出

I'm guessing that Facebook had better than that, so

其中各個值具體是多少

that's not gonna be a successful business.

此處的質量是球體的體積

And I get really pissed off when

因此可以把米的立方代進去

people ask me that question.

那麼就變成了能量等於米^5/秒^2

Because I think you can figure it out.

能據此計算出

And I love this story, and this is like my one

原子彈的威力

gratuitous story that I love that I'm throwing in here,

計算出俄羅斯和美國的原子彈的威力比

so the rest of it may not be as gratuitous.

從而解密了

But this is a picture that was published in

當時世界的頂級機密

Life magazine in 1950 of

這個是很難的問題

one of the Trinity nuclear bomb tests.

計算出Facebook的用戶留存並不困難

And there's a guy,

全世界有多少網民？

Geoffrey Taylor, who's heard of Geoffrey Taylor.

說個大體數

Awesome.

有人說24億有人說23億差不多就是這麼個數

Yes.

Facebook在中國是被禁的那麼還有多少網民?

Someone's heard of him.

約10億

Geoffrey Taylor was a British physicist,

約20億？好的20億網民

who ended up actually winning the Nobel Prize.

Facebook在上次收益報告中稱

And he was able to figure out from looking at

約有13億活躍用戶

this picture, what the power of the atomic bomb was,

可以把兩個數相除

the U.S. atomic bomb.

結果是不准確的

And Russian's were publishing similar pictures,

當然結果不准確

just using dimensional reasoning.

但是可以讓你大體了解

And dimensional reasoning is I think,

Facebook的用戶留存率

one of the best skills that I learned during my

如果把全世界網民都算上的話

time studying physics back in the U.K..

得出的結果要比剛才的數大一些

And what dimensional reasoning is, is you look at

同理如果看Whats App的話

the dimensions that are involved in a problem.

他們公司宣稱有6億活躍用戶

So, you wanna figure out energy?

有多少人有智能手機呢？你可以估算出來

Newtons, meters, newtons are kilograms, meters are second

這個數據比較好找

to the minus 2, so you've got kilograms,

從而可以大體知道有多少用戶

meters squared, seconds to the minus 2,

Amazon很火

and then you try and figure out how you can get each of

幾乎所有美國人都註冊了

those numbers from what data you have.

你可以知道在美國有多少網民

So, the mass.

從他們公佈的數據

Is the volume of this sphere.

可以推出亞馬遜有多少用戶存留

So, that's a meter cubed you throw in there.

公司類型不同

So you've got meters the 5 over seconds the minus 2,

用於判斷公司是否成功的留存率標準也不同

and he was able to use that to just figure out

如果是電子商務公司

what the power of this atomic bomb was,

月度活躍用戶有20%或30%的話

what the ratios of the power between the Russian and

公司經營得就很好

US atomic bomb, and essentially reveal one of

如果是社交網絡公司

the top secrets that existed in the world at that time.

第一批註冊用戶

That's a hard problem.

留存率不足80%的話

Figuring out what

該公司就不會火起來

Facebook's retention rate is not a hard problem.

因此用戶留存率多少算成功

How many people are there on the internet?

得取決於評判標準

Give or take someone throw something out.

你需要有一套方法

>> 2.4 billion, 2.3 billion.

找出可以進行比較的模板

Something like that right?

你可以以此為標尺

Okay, Facebook's banned in China.

看自己處於什麼位置才算成功了

So.

因此用戶留存率對於增長最重要

What now?

用戶留存來源於好創意

>> About a billion.

將創意轉化為好產品並且產品適合市場

>> About 2 billion?

看一個產品的留存率是否好

2 billion. So

使用同期群分析的話

2 billion people on the internet.

就要看註冊用戶

Facebook in their last earnings call

是否長期使用

said something around 1.3 billion in

一個好方法就是

terms of the number of active users.

看產品的前100個用戶

You can divide those numbers by each other.

前1000個用戶前10000個用戶

And yeah, that won't give you the right answer.

他們能被長期留住嗎？