2012年12月3日 星期一

Google Casts a Big Shadow on Smaller Web Sites

生活在谷歌巨人的陰影之下

Annie Tritt for The New York Times
Wize Commerce的首席執行官傑弗里·卡茨和員工在一起。他說,公司旗下的Nextag比較購物網站約60%的訪問量都來自谷歌。

從2月份起,傑弗里·G·凱茨(Jeffrey G. Katz)發現他公司旗下的比較購物網站Nextag來自谷歌搜索引擎的訪問量不斷下降,因此越來越擔心。
Nextag的工程師和外聘顧問趕緊展開了一次怪異的救火演練,想知道問題是否出在網站自身。或許是網站上某個無意的改動導致谷歌算法降低了Nextag的排名,這樣一來,當人們在谷歌中輸入“餐桌”或“割草機”等和購物相關的搜索名詞時,該網站的排名會比較靠後。
然而工程師發現,問題並非出在網站自身。與此同時,來自谷歌搜索引擎的訪問量繼續下降,降幅達50%。
怎麼辦?Nextag的對策就是,在過去五個月內,它將用於谷歌付費搜索廣告的開支增加了一倍。
此舉代價高昂,但卻是留住顧客的必要措施。凱茨說,因為Nextag的訪問量約有60%來自谷歌,包括自由搜索和付費搜索廣告,後者與搜索結果相 關,就出現在搜索結果旁邊。“我們必須這麼做,”凱茨說,“因為我們生活在谷歌的世界裡。”凱茨是Nextag的母公司Wize Commerce的首席執行官。
谷歌在網絡搜索和廣告業務中佔據主導地位,美國和歐洲的監管機構正在對它開展廣泛的調查。谷歌的崛起得益於其科技創新能力和敏銳的商業嗅覺;在美 國,谷歌佔據了搜索市場份額的67%,其搜索廣告收入也占整個市場的75%。公司規模大並不違法,但如果一個大公司利用其市場影響力來壓制競爭,這就違反 了反壟斷法。
因此,政府正在關注谷歌世界中各個網站的運營狀況,在這個蔓延的經濟生態系統中,它們的生存依賴於在谷歌搜索結果當中的排名。生活在巨人的陰影之下,這是一種怎樣的狀態呢?這些網站的體驗各不相同,頗為複雜,總的說來是混合著敬意和惶恐。
谷歌與網站、出版商和廣告商之間的關係經常都顯得不平衡,甚至是不公平。但是,谷歌也為它們提供並培育了一個充滿機會的平台。據谷歌估計,僅在美國,它掌控下的經濟生態系統便為180萬個企業、網站和非營利性組織創造了800億美元(約合元4988億元人民幣)的年收入。
政府對谷歌的調查十分徹底。自20世紀90年代末對微軟的調查以來,這是對大公司開展的最為詳盡的調查。
一些獲悉調查內情的人士不具名透露,美國聯邦貿易委員會(Federal Trade Commission,簡稱FTC)的工作人員已經建議有關方面準備對谷歌發起反壟斷訴訟。但是,此建議還需要FTC的委員們投票通過。即使通過,政府和谷歌也可以達成和解。
谷歌大張旗鼓地突破自己主導的搜索及搜索廣告產品領域,挺進電子商務和本地點評之類的領域,由此引起了反壟斷官員的注意。要判定谷歌是否違反了反壟 斷法,關鍵在於谷歌有沒有利用其搜索引擎來壓制競爭對手,為谷歌購物(Google Shopping)和谷歌+本地(Google Plus Local)等旗下產品提供優勢。
對於政策制定者來說,谷歌十分棘手。
“如何應對一個像谷歌這樣極具吸引力的壟斷公司,的確是反壟斷領域的一大難題,”哥倫比亞大學法學院(Columbia Law School)教授、前FTC高級顧問蒂姆·吳(Tim Wu)說。“反壟斷的目標是激勵企業通過持續創新而不是壓制競爭對手來保持主導地位。”
上個月(本文最初發表於2012年11月4日——編注),在亞利桑那州的谷歌時代精神(Google Zeitgeist)會議上,谷歌的聯合創始人及首席執行官拉里·佩奇(Larry Page)表示,考慮到谷歌的規模和影響力,他理解政府所進行的調查。“我們做的很多決定都實實在在地影響到了很多人,”他說。
問題的主要原因在於,谷歌一直在不停調整其搜索算法——該智能軟件決定着搜索結果的相關性、排名和呈現方式,這些搜索結果通常是導向其他網站的鏈接。
谷歌說,調整算法是為了改善服務。而且,長期以來,谷歌一直堅稱其算法淘汰了低質量的網站,呈現的都是最有用的結果,不論它們是否導向谷歌的產品。
“我們的首要和最高目標必須是,以儘可能快捷便利的方式為用戶提供他們想要的信息,”谷歌反網絡垃圾小組負責人馬特·卡茨(Matt Cutts)說。
然而,谷歌算法處在保密狀態,算法調整往往會讓其他網站手忙腳亂。
想想創立於2003年的非營利性組織Vote-USA.org吧。該網站為選民提供網上信息,以免他們到了投票站之後才沮喪地發現,選票上的名字幾乎有一半都不認識。網站免費提供聯邦、州和地方選舉的樣本選票,上面有候選人的照片、個人信息以及在相關問題上的立場。
在2004和2006年的選舉中,用戶創建了數萬張樣本選票。Vote-USA.org的經營者羅恩·卡洛(Ron Kahlow)說,到了2008年,網站的流量卻急劇下降,原因是“我們從谷歌世界的地圖上掉了下來”。
卡洛是一個搜索引擎優化公司的創始人,還享受了谷歌給非營利性組織提供的免費廣告優惠政策,他對谷歌世界裡的生存之道有所了解。他仔細研究了谷歌關於網站的指導手冊,對自己的網站做出了一些改變,然後給谷歌發了郵件。然而,他沒有得到任何回復。
“我失去了支撐網站運行的全部捐贈來源,”卡洛說。“當時的情況非常非常艱難。”
私人關係給事情帶來了轉機。卡洛的一個朋友認識計算機和通信行業協會(Computer & Communications Industry Association)的首席執行官埃德·布萊克(Ed Black),而谷歌也是該協會的會員。布萊克幫卡洛詢問了情況,谷歌的一位工程師便進行了調查。
卡洛後來得知,問題在於自家網站的州選舉頁面上也列有國家級候選人的信息——對於在單一數字位置搜尋選舉信息的人來說,這樣的安排十分合理。但是,谷歌算法會把網站上的重複內容看做網站試圖讓自身規模看上去比實際更大的取巧伎倆。
卡洛解決了這個問題,這樣一來,瀏覽州選舉頁面的用戶就必須打開另外一個網頁才能看到國家級候選人的選票和信息。很快,Vote-USA.org被谷歌撤出了黑名單,在這個選舉季期間,該網站的日訪問量一直高達33.3萬。
谷歌對自己和具體網站的關係不予置評。然而,一個獲悉谷歌和Vote-USA.org互動情況的人說,谷歌也在後者的網站上發現了其他的重複內容, 比如說,有關愛荷華州選舉的頁面上出現了密歇根候選人的信息。該知情人士還說,幾乎沒有任何其他網站導向Vote-USA.org,而鏈接情況也是谷歌算 法中最重要的衡量因素之一。
去年年初,全國各地小型地方新聞網站的運營者都發現,自家網站的讀者數量驟降。這是為什麼呢?原來,他們的網站已經從谷歌新聞(Google News)上消失了,而後者往往是它們最主要的訪問量來源。
這些網站的所有者給谷歌寫了郵件,搜遍了所有的網上論壇,但卻無濟於事。
“沒有任何解釋,也不知道去哪兒了解更多的信息,”北卡羅來納州凱里的地方新聞網站“凱里公民”(CaryCitizen)的主編和出版人哈爾·古德特里(Hal Goodtree)說。
古德特里說,谷歌新聞可以為網站帶來訪問量和可信度。“如果能出現在谷歌新聞的搜索結果中,你就是可信的,”他解釋道,“所以,當時的日子很不好過。”
加利福尼亞州伯克利的地方新聞網站Berkeleyside的聯合創始人蘭斯·克諾貝爾(Lance Knobel)在Twitter上抱怨了這件事。身為谷歌工程師的一個Berkeleyside讀者最終看到了這條信息,並且向克諾貝爾承諾,他將私下調查這個問題。
12個小時後,Berkeleyside重新出現在了谷歌新聞當中。谷歌表示,撤走該網站是個錯誤。
而在北卡羅來納州,幾個月的時間裡,古德特里一直都不知道到底發生了什麼。6月,他收到了來自谷歌的一封郵件,郵件里說,關於網站遭到撤除的原因,谷歌“無法提供具體細節”。
到了8月,谷歌又寫郵件說,“最近,我們再次審查了你們的網站,並決定把它重新放到谷歌新聞上。”谷歌沒有給出進一步的解釋。“凱里公民”的訪問量提高了24%。說到谷歌的時候,加利福尼亞的克諾貝爾表示:“這是一個完全不透明的公司。”
谷歌的卡茨表示,公司不可能就每個網站所有者的問題一一作答,因為問題太多;總共有2.4億個域名,人們每天使用谷歌搜索的次數超過了33億。
卡茨說,儘管如此,谷歌還是在努力改進。去年,它開始啟動和網站所有者的視頻對話,回答相關問題。谷歌也開始發表博文,解釋自己的算法調整,並且播放了一段時長8分鐘的內容,內容來自一次關於算法調整的絕密會議。
“我們十分努力,確保和網站管理員之間的溝通,”卡茨說。他還舉例說,在颶風“桑迪”導致Gawker媒體公司(Gawker Media)癱瘓之後,他就向該媒體保證,它仍會出現在谷歌搜索結果中。
地方新聞報道和寫作是一個勞動密集型的瑣碎領域,谷歌並不與“凱里公民”和Berkeleyside之類的網站在這一領域形成競爭關係。然而,Berkeleyside的克諾貝爾指出,在地方廣告業務方面,地方新聞網站確實以自己的方式成為了谷歌的競爭對手。
克諾貝爾說,在創建自己的當地企業名錄和點評服務提供者谷歌+本地的過程當中,谷歌搜索引擎可以給這個網絡巨頭帶來“巨大優勢”。他說,結果就是,谷歌會把尋找當地企業信息的用戶引向谷歌+本地,而不是Berkeleyside之類的小網站或者Yelp之類的大型網站。
佩奇表示,谷歌最重要的目標就是不斷完善自己的產品,這意味着公司需要增加收集和解析數據的服務項目。它的一些競爭對手可能要遭殃了。但是,他補充道,“我們的工作是為用戶服務。”
谷歌已經向監管部門着重強調了這一點,因為在反壟斷調查當中,消費者利益是一個十分重要的衡量標準。
本文最初發表於2012年11月4日。
翻譯:谷菁璐


Google Casts a Big Shadow on Smaller Web Sites

Annie Tritt for The New York Times
Jeffrey G. Katz, the chief executive of Wize Commerce, seen with employees. He says that about 60 percent of the traffic for the company’s Nextag comparison-shopping site comes from Google.

STARTING in February, Jeffrey G. Katz grew increasingly anxious as he watched the steady decline of online traffic to his company’s comparison-shopping Web site, Nextag, from Google’s search engine.
In a geeky fire drill, engineers and outside consultants at Nextag scrambled to see if the problem was its own fault. Maybe some inadvertent change had prompted Google’s algorithm to demote Nextag when a person typed in shopping-related search terms like “kitchen table” or “lawn mower.”
But no, the engineers determined. And traffic from Google’s search engine continued to decline, by half.
Nextag’s response? It doubled its spending on Google paid search advertising in the last five months.
The move was costly but necessary to retain shoppers, Mr. Katz says, because an estimated 60 percent of Nextag’s traffic comes from Google, both from free search and paid search ads, which are ads that are related to search results and appear next to them. “We had to do it,” says Mr. Katz, chief executive of Wize Commerce, owner of Nextag. “We’re living in Google’s world.”
Regulators in the United States and Europe are conducting sweeping inquiries of Google, the dominant Internet search and advertising company. Google rose by technological innovation and business acumen; in the United States, it has 67 percent of the search market and collects 75 percent of search ad dollars. Being big is no crime, but if a powerful company uses market muscle to stifle competition, that is an antitrust violation.
So the government is focusing on life in Google’s world for the sprawling economic ecosystem of Web sites that depend on their ranking in search results. What is it like to live this way, in a giant’s shadow? The experience of its inhabitants is nuanced and complex, a blend of admiration and fear.
The relationship between Google and Web sites, publishers and advertisers often seems lopsided, if not unfair. Yet Google has also provided and nurtured a landscape of opportunity. Its ecosystem generates $80 billion a year in revenue for 1.8 million businesses, Web sites and nonprofit organizations in the United States alone, it estimates.
The government’s scrutiny of Google is the most exhaustive investigation of a major corporation since the pursuit of Microsoft in the late 1990s.
The staff of the Federal Trade Commission has recommended preparing an antitrust suit against Google, according to people briefed on the inquiry, who spoke on the condition they not be identified. But the commissioners must vote to proceed. Even if they do, the government and Google could settle.
Google has drawn the attention of antitrust officials as it has moved aggressively beyond its dominant product — search and search advertising — into fields like online commerce and local reviews. The antitrust issue is whether Google uses its search engine to favor its offerings like Google Shopping and Google Plus Local over rivals.
For policy makers, Google is a tough call.
“What to do with an attractive monopolist, like Google, is a really challenging issue for antitrust,” says Tim Wu, a professor at Columbia Law School and a former senior adviser to the F.T.C. “The goal is to encourage them to stay in power by continuing to innovate instead of excluding competitors.”
SPEAKING at a Google Zeitgeist conference in Arizona last month, Larry Page, the company’s co-founder and chief executive, said he understood the government scrutiny of his company, given Google’s size and reach. “There’s very many decisions we make that really impact a lot of people,” he acknowledged.
The main reason is that Google is continually adjusting its search algorithm — the smart software that determines the relevance, ranking and presentation of search results, typically links to other Web sites.
Google says it makes the changes to improve its service, and has long maintained that its algorithm weeds out low-quality sites and shows the most useful results, whether or not they link to Google products.
“Our first and highest goal has to be to get the user the information they want as quickly and easily as possible,” says Matt Cutts, leader of the Web spam team at Google.
But Google’s algorithm is secret, and changes can leave Web sites scrambling.
Consider Vote-USA.org, a nonprofit group started in 2003. It provides online information for voters to avoid the frustration of arriving at a polling booth and barely recognizing half the names on the ballot. The site posts free sample ballots for federal, state and local elections with candidates’ pictures, biographies and views on issues.
In the 2004 and 2006 elections, users created tens of thousands of sample ballots. By 2008, traffic had fallen sharply, says Ron Kahlow, who runs Vote-USA.org, because “we dropped off the face of the map on Google.”
As founder of a search-engine optimization company and a recipient of grants that Google gives nonprofits to advertise free, Mr. Kahlow knows a thing or two about how to operate in Google’s world. He pored over Google’s guidelines for Web sites, made changes and e-mailed Google. Yet he received no response.
“I lost all donations to support the operation,” he said. “It was very, very painful.”
A breakthrough came through a personal connection. A friend of Mr. Kahlow knew Ed Black, chief executive of the Computer & Communications Industry Association, whose members include Google. Mr. Black made an inquiry on Mr. Kahlow’s behalf, and a Google engineer investigated.
The problem, Mr. Kahlow learned, was that the site’s state Web pages also had information for national candidates — reasonable to a person seeking voting information at one digital location. But to Google’s algorithm, duplicate content on a site suggests a shady shortcut to try to make a site look bigger than it is.
Mr. Kahlow fixed that, so that someone on a state page must go to another page to see the ballot and information about national candidates. Vote-USA.org quickly moved off Google’s black list, and during this election season people have been viewing 333,000 pages a day.
Google will not comment on its relations with specific Web sites. But a person briefed on its interactions with Vote-USA.org said Google had found additional duplication, like Michigan candidates showing up on the Iowa Web site. The person also said that hardly any other Web sites linked to Vote-USA.org, one of the most important elements in the algorithm.
EARLY last year, operators of small local news sites nationwide found that their number of readers had plummeted. Why? Their Web sites had disappeared from Google News, which in many cases was their No. 1 source of traffic.
Their owners wrote e-mail messages to Google and scoured online forums, to no avail.
“There was no explanation why or place you could go for more information,” says Hal Goodtree, editor and publisher of CaryCitizen, a local news site in Cary, N.C.
Google News, he says, brings traffic and credibility. “You’re legit if you’re on Google News,” he explains, “so it was painful.”
Lance Knobel, co-founder of Berkeleyside, a local news site in Berkeley, Calif., complained about the problem on Twitter, and a Berkeleyside reader who was a Google engineer eventually saw it. He promised Mr. Knobel he would look into it, unofficially.
Twelve hours later, Berkeleyside reappeared in Google News. Google said dropping the site was an error.
In North Carolina, Mr. Goodtree heard nothing for months about what had happened. In June, he received an e-mail from Google saying it was “unable to provide specifics” about why the site had been dropped.
Then, in August, Google wrote, “We recently reviewed your site again and have decided to add it to Google News once more.” No further explanation was given. Traffic to CaryCitizen jumped 24 percent. Mr. Knobel, in California, says of Google: “It’s a totally opaque corporation.”
Mr. Cutts of Google says it is impossible for the company to respond individually to every Web site owner with a question, because of sheer scale; there are 240 million domain names and people search Google more than 3.3 billion times a day.
Yet Google is trying to do more, he says. Last year, it started hosting video chats for Web site owners to ask questions. It has also started publishing blog posts about changes in the algorithm and broadcast eight minutes of a top-secret meeting about it.
“We try very hard to make sure we communicate with Webmasters,” Mr. Cutts says. For instance, he says, he reassured Gawker Media that it would remain in search results after Hurricane Sandy caused it to crash.
Google does not compete with sites like CaryCitizen and Berkeleyside in the labor-intensive chore of reporting and writing local news. But local news sites, notes Mr. Knobel of Berkeleyside, do compete in their way with Google for local advertising.
As the company builds up Google Plus Local, its local business listing and review service, Mr. Knobel says Google’s search engine could give the Internet giant “a tremendous advantage.” That would result, he says, if a person seeking information about a local business is steered to Google Plus Local rather than to small sites like Berkeleyside or sizable services like Yelp.
Google's overriding goal, Mr. Page suggests, is continual improvement of its product, which, he says, means adding more services that collect and parse data. Some competitors may suffer. But, he adds, “our job is to serve users.”
Google has argued that forcefully to regulators because in antitrust, consumer benefit weighs heavily.


沒有留言:

網誌存檔