经济学人:亚马逊上的评论 五星假好评

 　　Business: Reviews on Amazon Five-star fakes 商业：亚马逊上的评论，五星假好评

　　The evolving fight against sham 1 reviews. 打击虚假评论进行中。
　　“I will post awesome 2 review on your amazon product,” bess98 declared on Fiverr, a website where individuals sell freelance services for $5 or more. “我会对您在亚马逊出售的产品给予好评。”Fiverr网站用户bess98如是说，该网站用户出售这样的自由职业服务，要价为五美元或更多。
　　On October 16th Amazon charged that bess98 and more than 1,000 others were illegally hawking 3 customer reviews. 10月16日，亚马逊网站起诉包括bess98在内的一千多名用户的评论为推销行为的非法评论。
　　The case comes just six months after Amazon sued the operator of four sites peddling 4 similar stuff, including the subtly named buyamazonreviews. com. 六个月以前，亚马逊起诉了四家有类似推销行为的运营者，包括其中一家网址非常微妙的buyamazonreviews.com(意为购买亚马逊好评)。
　　Like Amazon, other websites have fought fakes with lawsuits 5, carefully honed algorithms and even sting operations 同亚马逊一样，其他许多网站也以起诉、过滤评论和安排人员打入内部等方式抵制虚假评论。
　　Yelp 6, a popular review site, has had undercover staff answer ads from firms seeking glowing write-ups. 比如，颇受欢迎的点评网站Yelp就有卧底人员应对那些寻找虚假好评的入驻商家。
　　Yet the problem persists. 但问题仍然存在。
　　For as long as there have been online reviews, there have been fakes. 因为只有有网上评论的存在，就会有虚假评论。
　　The motivation is clear: for example, one extra star on a restaurant's Yelp rating boosts revenue by 5-9%, according to Michael Luca of Harvard Business School. 动机显而易见：例如，据哈佛商学院的麦克·卢卡统计，Yelp点评网上的商家每多一颗星，其营业额变会增加5%到9%。
　　Mr Luca and Georgios Zervas of Boston University have shown that restaurants seeking fake acclaim 7 are likely to be independent 卢卡先生和波士顿大学的乔治·泽瓦斯发现寻求虚假评论的饭店一般为个体商户
　　online reviews matter more to them than to chains with established reputations. 同已经建立起自己声誉的连锁店比，前者更需要网上好评。
　　So some businesses ask friends to post raves 8, seek reviewers-for-hire and offer customers discounts in exchange for praise. 这些个体商户要么请朋友发表溢美之词，要么找付费的评论者，或是给提供好评的客户折扣。
　　For websites that claim to be an impartial 9 resource, such practices are troubling. 对于旨在提供中肯评价的网站，这些行为会损害网站的宗旨。
　　“While small in number,” Amazon contends in its new suit, 亚马逊在最近这次起诉案中辩论道，“尽管目前这类评论为数不多，
　　"these reviews can significantly undermine the trust that consumers and the vast majority of sellers and manufacturers place in Amazon." 但这确实会损害顾客和众多卖家及制造商给予亚马逊的信任。”
　　The problem is particularly irksome for sites dedicated 10 to offering reviews, such as Yelpand TripAdvisor. 对于专门提供点评的网站，如Yelp和猫途鹰，问题更为棘手。
　　Amazon sells everything from books to lawnmowers; Yelp's main offerings are its 83m reviews. 毕竟亚马逊还卖各种产品，从书本到割草机，应有尽有；而Yelp的主要业务便是多达8300万的点评。
　　“If consumers can't rely on the content,” says Vince Sollitto of Yelp, “then the service is of no value.” Yelp发言人Vince Sollitto说，“如果顾客没法信赖点评内容，那么我们的服务便没有意义。”
　　So websites have tried to fight fakes. 所以各网站对虚假评论进行了打击。
　　Algorithms comb reviews for suspicious wording. 运用一定的计算机算法来筛除具有可疑的措辞的评论。
　　Expedia allows hotel recommendations only by those who have paid for a room there. 智游网（Expedia）只许付款入住了酒店的顾客推荐酒店。
　　Amazon tags a review as “verified” if the writer has indeed bought the product. 亚马逊把真正买了商品的顾客评论打上“真实可信”的标签。
　　Presumably such reviews are more reliable, though bess98 is one of many who claim to be able to game Amazon's system. 按理来说，这些评论更加可靠，尽管诸如bess98这样的网友声称能躲过亚马逊的法眼。
　　Yelp may have the most aggressive strategy. Yelp采取了可能最为激烈的打击方法。
　　An algorithm removes a whopping 30% of posts from Yelp's list of “recommended” reviews, though consumers can still see the suspicious ones if they like. 他们采用计算机算法筛除掉网站中多达30%的好评，尽管顾客稍加注意仍能看到疑是虚假的点评。
　　Businesses that try to weasel their way to a higher rating (paying off grumpy clients, for instance) have their Yelp pages branded with a red warning. 为了得到更高评分而搞小动作的商家（如收买打算抱怨的顾客），他们的Yelp页面都被打上了红色的警告字样。
　　Despite all this, some false acclaim and critiques inevitably 11 slip past firms' defences. 尽管如此，仍有虚假的好评或者差评成为漏网之鱼。
　　For websites, fake reviews will remain a stubborn headache. 对于点评网站，虚假评论仍是难以解决的顽疾。
　　Meanwhile businesses are finding new ways to boost their reputations online. 与此同时，商家则在寻找各种新方法在网络上提高声誉。
　　Social bots—lines of code that pose as real accounts—can build buzz on social-media sites like Twitter and Facebook. 代码写成的社交机器人可以模仿真人进行评论，在诸如Twitter和Facebook等社交网站上制造舆论。
　　For the average consumer, it may become ever harder to distinguish real praise from puff 12. 对普通消费者来说，从宣传广告中辨别评论的真实性越来越难。