李開復針對 Citron Research 8 月 31 日的公開信，打算怎麼回復？

01-11

Citron Research 發表於 8 月 31 日的公開信地址：http://www.citronresearch.com/wp-content/uploads/2012/08/reply-points-to-Kai-Fu-Lee-August-31.pdf [PDF]

下面是我對 Citron 8 月 31 日公開信回應的中英文版，首發於知乎，轉載請註明出處。

________________________________________

在安德魯·萊福特先生於 2012 年 8 月 31 日發表的「逐條」回應中，他完全避開了我的七點質疑，支支吾吾，閃爍其詞。我在下面的最新回應中，將再次清晰地闡述我的七點質疑，並指出萊福特先生 8 月 31 日公開信中暴露出的如下問題：

萊福特先生避而不談對他完全不懂搜狗產品線的指證，這應該被視作他默認了自己的無知。
萊福特先生對於他所推薦的公司的理解水平，就像是一個分析師說：「買入微軟股票，因為 Xbox + Office 的組合能夠擊敗 PlayStation」。
萊福特先生新提出了一個與實際市場份額相差了兩至三倍的數字！
萊福特先生承認他利用了他認為虛假的數據，用於佐證他的股票推薦，卻不向新讀者指出這些數據是「虛假」的。
萊福特先生「 6.97 億美元市值」的觀點被擊潰後，他解釋他的意思只是「 2 億美元太低了」。分析師能因為「嫌低」就隨意上調三倍估值嗎？
在一家公司的市場份額從 0% 漲到 10% 的數天時間裡，另一家公司的市場份額停留在 4.5% 上不變，萊福特先生於是得出結論稱，這兩家公司的市值應該有同樣的增幅。

________________________________________

2012 年 8 月 31 日，Citron 的安德魯·萊福特發表了對我兩封公開信中七點質疑的回應（http://www.citronresearch.com/wp-content/uploads/2012/08/reply-points-to-Kai-Fu-Lee-August-31.pdf）。下面是我對他的逐條反駁：

1. Citron 缺乏對中國搜索市場缺乏的基本理解。
2. Citron混淆了搜狗的幾個產品，把搜索和拼音合併成了一個產品。

a) 我再這兩個部分里的主要觀點是，萊福特先生對搜狗的產品及其競爭力連最基礎的理解也不具備。他在這方面認知不足的可笑體現是，他將兩個本質上毫無關聯的產品混為一談，憑空創造了一個新產品！我會將他在此問題上的沉默，視為他承認了他對自己推薦的這家公司的所有產品都缺乏認知。（這就像一個分析師推薦微軟股票的原因是「Xbox將會擊敗 PlayStation，因為微軟的 Office 套件集成在了每一款 Xbox 遊戲之中」！！！）
b) 關於他提出的幾個新論點：

i. 萊福特先生稱，搜狗的規模約為百度的10%。他指的是收入規模還是搜索份額？我們就兩個都算吧：百度 2012 年第二季度收入是 8.59億美元，搜狗是 3千萬美元。百度二季度搜索份額是86.32%，搜狗是4.54%。因此，搜狗的規模是百度的3.5%或5.3%，而不是10%。令人驚詫的是，萊福特先生作為一名證券分析師，對市場份額數字的偏差竟然高達2-3倍！
ii. 萊福特先生稱，奇虎「數年之後可能會在中國搜索市場獲得10%的份額」。事實是，奇虎現在就已經有10%的份額了。
iii. 萊福特先生稱，奇虎的搜索與一個完全貨幣化的搜索引擎，完全不可相比。他大錯特錯，連我所提到的Google AdWords這一點都沒懂。只需一個季度，Google AdWords 就可以在現有的奇虎搜索上部署完畢，這將能產生每年9800萬美元的收入。AOL、http://ask.com 和騰訊搜搜都這樣做過。這就像將橙子變為橙汁（利用現成可用的榨汁機），而不是變成蘋果。

3. Citron 在進行網站流量分析時，製造不對等的比較關係。

a) 萊福特先生又一次忽視了我的觀點：1) 用兩個網站的流量之比，完全不能推導出兩家公司的市值之比（雅虎的流量是亞馬遜的三倍，但市值卻只有後者的1/7）；2) 將兩個業務不同的公司的流量做對比，也完全沒有意義（MSN曾經在很長一段時間裡流量都比Google 大，但二者不可對比）；3) Citron對數據進行了扭曲，以得到他們期望的結果。
b) 萊福特先生稱，「門戶網站和導航網站都是流量分發器（gateways），因此他們屬於同一類別。」在中國，導航站的目的是將用戶帶到其他網站去（因此它更像推薦引擎，易於貨幣化，從而得到更高估值）；門戶網站的目的是將用戶留下來（因此耗費大量成本用於內容建設）。我會更願意擁有一個流量只有門戶網站一半的「導航站」，因為它更能貨幣化，運營成本也更低。這些基本要點，都被萊福特先生忽略了，他看起來更願意用Alexa 數據來判斷股價。

4. Citron在對比搜狐和奇虎的財務指標時，略去了關鍵數據。

a) 萊福特先生稱其「比較了搜狐和奇虎的財務指標，便於每個投資人得出自己的結論」。但他卻隱去了收入和利潤增長率不談！拋開如此關鍵的數據，一個投資人怎麼可能得出結論？如果萊福特先生想要拿財務數據進行對比，那麼他就必須遵照行業慣例，提供收入和利潤的增長率。他將最關鍵信息略去不提（以得出其錯誤結論）的這種財務分析方法令人震驚！
b) 萊福特先生稱，在對奇虎進行評估時，沒有考察增長率是OK的，因為這家公司的營收數字是假造的。既然如此，萊福特先生所使用的營收信息就也都是假造的了。為什麼萊福特先生會在他的分析中使用他認為是虛假的數據？（並且不向新讀者指明這一點！）

5. Citron不明白搜索是如何貨幣化的，錯誤地將奇虎搜索9800萬美元的營收預估化為零。

a) 這裡有三點：(A) 奇虎若使用Google的搜索和廣告，分析師預計2012年的收入為1億美元；(B) 奇虎若使用自己的搜索，不加廣告，那麼收入則是0；(C) 奇虎若使用自己的搜索，但按我建議的那樣加上Google的廣告，它在2013年的收入又會達到9800萬美元。
b) 萊福特先生的兩個反擊觀點都像是閃爍其辭：(1) 奇虎管理層聲稱搜索0收入，是因為只有上文提到的 (B) 選擇。(2) 萊福特先生承認分析師預估的1億美元是 (A) 可能。這兩點都沒有回答我對他報告提出的批評。
c) 我的批評很簡單：萊福特先生最初的報告提到了(A)和(B)，但沒有提到(C)。換言之，萊福特先生認為奇虎要花很多年才能建立起貨幣化的技術基礎，這就給收入帶來了負面影響。萊福特先生不懂的是，奇虎並不需要搭建自己的廣告系統，他們只要使用Google AdWords即可。閃爍其辭再多，也不會讓人們以為Citron懂了這些基本信息。

6. Citron 對視頻業務的分析貽笑大方，明顯地忽視了重大的商業因素。

a) 我的批評針對萊福特先生的這句話：「我們假設搜狐視頻的估值等於土豆（ 6.79 億美元）。」這實在令人咋舌，我給出了四個商業方面的原因來論證了他這樣對照毫無可信之處。在萊福特先生的「反駁」中，他現在再次重申了他認為「JP 摩根的 2 億美元估值實在過低」的觀點，我會將他這個重申觀點（從 6.79 億美元降到了「高於 2 億美元」）的行為，視作他承認了最初犯下的「毫無原則地胡亂複製粘貼數字」的過錯。

7. Citron 將兩家公司的市值胡亂加總做法，十分荒唐。

a) 我之前的觀點就很簡單：Citron將奇虎在四天時間裡的市值增量，簡單地加到搜狗的市值上去，這期間，奇虎的市場份額從0%升至10%，而搜狗的份額從4.5%變成了4.5%。這種加法實在驚人。萊福特先生的回復是，這多少與JP摩根的分析（？？？）有些關係。他顧左右而言他並不意外，因為沒有任何辦法能解釋他那種瘋狂的分析。

有趣之處是，萊福特先生的「反駁」沒有回答我在第一篇文章中提出的任何一個問題。這我能理解，因為他的報告錯誤連篇，漏洞百出。諷刺的是，萊福特先生曾經說過：「我只說事實。我們所有的事情都是整個團隊完成的，我們不能確定的事情，我們就不會發表。」看來他對「事實」的定義和我們地球人不同。

李開復

DR. LEE RESPONDS TO CITRON"S LETTER OF AUG 31

________________________________________

In Mr. Andrew Left』s 「point-by-point」 response dated August 31, 2012, he completely evaded Dr. Lee』s original seven criticisms, and chose to just mince words. Dr. Lee』s new response below restated the seven clear criticisms, and also pointed out the following new observations from Mr. Left』s letter dated August 31:

Mr. Left』s unwillingness to engage on the direct accusation that he totally misunderstood Sogou』s product line should be taken as an admission of ignorance.
Mr. Left』s understanding of a company he recommends is at the level of an analyst who says : 「Buy Microsoft because the Xbox+Office product will beat PlayStation」.
Mr. Left made a new statement about market share that is off by a factor of 2-3X!
Mr. Left admits that he used what he considered fraudulent data in his recommendation, without revealing the 「fraud」 to new readers.
When Mr. Left"s "valuation of $697M" was discredited, he said he was just giving "an opinion that $200M is too low". Does this mean analyst can arbitrarily triple any valuation that he considers too low?
During a period where company』s market share went from 0% to 10%, while another stayed flat at 4.5%, Mr. Left concludes the two companies』 valuations should increase by exactly the same amount.

________________________________________

On August 31, 2012, Citron』s Andrew Left issued a response (http://www.citronresearch.com/wp-content/uploads/2012/08/reply-points-to-Kai-Fu-Lee-August-31.pdf) to my two letters and seven points. Below are my point-by-point rebuttal to him:

1. Citron lacked basic understanding of Chinese search market.
2. Citron misunderstood Sogou products and combined two products into one.

a. My main point in these two sections was that Mr. Left didn』t even possess a rudimentary understanding of Sogou』s products and their competitiveness. His lack of understanding was comically demonstrated in his fabrication of a new product by combining two basically unrelated products! I will take his silence on this point as his acknowledgement of his inept understanding of all the products offered by the company that he is recommending. (Imagine an analyst recommending Microsoft because 「Xbox will defeat PlayStation because Microsoft Office is integrated with every Xbox game」!!!)
b. Regarding the NEW points he brings up:

i. Mr. Left now claims Sogou is about 10% the size of Baidu. Does he mean revenue or search share? Let』s try both: Baidu』s 2012 Q2 revenue was $859M, and Sogou』s was $30M. Baidu』s Q2 search share was 86.32% and Sogou』s was 4.54%. So Sogou is 3.5% or 5.3% of Baidu, not 10%. It is appalling that Mr. Left as an analyst could be off on his market share numbers by a factor of 2-3!
ii. Mr. Left states that Qihoo 「might gain a 10% foothold in search in China, after several years」. The fact is that Qihoo has 10% now.
iii. Mr. Lefts states that Qihoo』s search compared to a fully monetized search is comparing 「apples」 to 「oranges」. He is completely wrong, and didn』t get the basic points I made about using Google AdWords. Google AdWords can be added to the existing Qihoo search in about one quarter, and it would produce $98M of revenue per year. This has been done by AOL, ask.com, and Tencent』s Soso. This is about converting oranges to orange juice (with a proven juice machine), not apples!

3. Citron compared apples to oranges in its web traffic analysis, and did so unfairly.

a. Again, Mr. Left ignores my points that: 1) comparing web traffic is not a meaningful way to compare two companies』 market value (Yahoo has three times the traffic of Amazon, but only one-seventh the value), 2) it is not meaningful to compare web traffic of two companies in different businesses (MSN had more traffic than Google for a long time, but are not comparable), and 3) Citron distorted the data to get the result it wanted.
b. Mr. Left states that 「portals and navigation sites were both gateways」 thus they』re in the same category.」 In China, Navigation sites』 purpose is to send the users to another site (possessing much value as it serves as a recommendation engine that could be monetized); portal sites』 purpose is to keep users there (costing a great deal to build all the content). I would prefer to own a 「navigation site」 that has one-half the traffic of a 「portal site」. It monetizes better, and costs less to run. These are simple points missed by Mr. Left, who appears to want to use Alexa to determine stock price.

4. Citron omitted critical data in its financial comparisons of Sohu and Qihoo.

a. Mr. Left states that 「The "financial comparisons" between Sohu and Qihoo are for every investor to draw their own conclusions」. But he left out revenue and earnings growth! How can an investor draw conclusions without the most critical data? If Mr. Left wants to compare the financial performance, then he must follow industry practice and provide revenue and earnings growth. His approach of doing financial analysis omitting the most critical information (in order to reach his wrong conclusion) is outrageous!
b. Mr. Left states that it is OK to leave out the revenue and earnings growth figures in assessing Qihoo because the company revenue numbers are fraudulent. In that case, Mr. Left』s usage of the revenue information directly is even more fraudulent. Why did Mr. Left use what he considers to be fraudulent data in his analysis? (and not pointing it out to the first-time readers!)

5. Citron did not understand search monetization and zeroed $98M of revenue.

a. There are three points: (A) When Qihoo used Google search and ads, analyst predicted $100M revenue in 2013, (B) When Qihoo used its own search, and no ads, it would receive $0, (C) If Qihoo used its own search but Google』s ads as I suggested, it would get back the $98M in 2013.
b. Both of Mr. Left』s defensive points are just mincing words: (1) Qihoo』s management stating zero search revenue was about (B) only. (2) Mr. Left』s acknowledgement of analyst $100M estimates are about (A) only. They do NOT answer my criticism of his report.
c. My criticism was simply that: Mr. Left』s original paper mentioned (A) and (B) but not (C). In other words, Mr. Left assumed it would take years for Qihoo to build the monetizing infrastructure, thus negatively affecting revenue. Mr. Left did not understand that Qihoo did not need to build an ad infrastructure, and can just use Google AdWords. No amount of word-mincing will trick people to think Citron understood this basic point.

6. Citron』s video analysis was ludicrous, and missed obvious business issues.

a. My criticism was on Mr. Left』s statement: 「we"ve assumed a valuation of SOHU"s video matching Tudou ($679M).」 This was outrageous, and I gave four business reasons to discredit his analogy. In Mr. Left』s 「rebuttal」, he now restates his position to be 「JP Morgan』s $200M is too low」. I will accept his restatement ($679M down to 「more than $200M」) as admission that he was originally guilty of 「undisciplined and wild copy-pasting of numbers.」

7. Citron』s adding market cap from one company to the next was wild and ridiculous.

a. My original point was very simple: Citron took Qihoo』s market cap gains in four days and simply adding it to Sogou』s market cap, during a period when Qihoo』s market share went from 0% to 10%, while Sogou』s went from 4.5% to 4.5%. This is shocking. Mr. Left』s reply was that this was somehow related to JP Morgan』s analysis (???). His evasiveness is not surprising, as there can be no way to justify that kind of wild analysis.

It is interesting that Mr. Left』s 「rebuttal」 didn』t answer any of the issues I brought up in my original paper. This is understandable because his report had fundamental flaws and errors. It is ironic that Mr. Left has said in the past: 「I only tell the truth. We run everything by a team and if we are unsure of it, we don』t publish it」. It certainly seems that his definition of 「truth」 is not from this planet.

Kai-Fu Lee

其實我覺得說一句話就行了：

Mr. Left, you"ll always be non-right!

一個漏洞和錯誤的尋找者，通常有兩種人，一種針對某一領域極其專業，專業到超越某一領域的絕大多數人，例如索羅斯對於金融，例如某些黑客對於互聯網，也包括某些律師對於法律，各行都有該行業的頂尖尋漏人才。另一種人，拾荒者，無所謂是皮具服裝業還是電器電子產品，亦或是傢具裝潢，或者是花草陶瓷，總之，在拾荒界奮戰多年，總能減到別人不小心或慪氣傷心或小偷行竊後扔掉的仍然值錢的東西，好比撿到一個阿瑪尼錢包，然後高舉起來叫喊道：我找到了一個你的錯誤，這就是事實，同時我宣布，因為在垃圾場撿到了阿瑪尼，足以證明阿瑪尼錢包就是垃圾。這就是某些人的不可思議的外星系理論。