大数据拯救招聘官

    对于一家初创公司的创始人史蒂夫•古德曼来说，一切似乎看起来都很顺利，只有一个重大的细节不是这样。

    即便是在经济衰退期间，美国的失业率高达两位数，但古德曼的公司还是找不到既能胜任公司工作、又符合公司企业文化的电脑程序员。

    所说他说：“要想寻找、招聘到最好的人员好比大海捞针。”

    因此，古德曼决定创办一家名叫Bright的公司。包括Bright在内的几家初创公司依靠强大的电脑技术以及互联网上的数十亿份简历信息来构建一种“猎头算法”，他们相信这种算法会使企业的人才招聘变得更有效率，而且更节省成本，同时实现求职者和招聘方的双赢。

    互联网起初的确使找工作变得轻松了不少。只需轻轻点击几次鼠标，求职者就可以同时申请十几个职位，企业的招聘通知也能同时被好几千人看到。有些大公司收到的简历数量甚至增长了五倍之多。

    不过网络求职很快露出了令人失望的一面。首先是许多求职者在发出不少简历后没有收到任何反馈，因此变得意志消沉。同时招聘方也被淹没在简历的汪洋大海里，只能根据一些表面的甚至不可靠的信息来筛选求职者——比如根据求职者的毕业院校来进行选择。

    大约在两年前，古德曼创立了Bright公司，并且收集了数百万份简历和工作描述，创建了一个叫做Bright Score的招聘系统。古德曼称，系统的算法具有不断学习的能力，因此它可以告诉用人单位，对于某一个特定的工作来说，什么样的求职者才是最适合的。

    古德曼表示：“一个招聘者现在只需要读50份简历，因为我们替他们筛选过了。同时它也会告诉求职者，这份工作是否适合自己。如果不适合的话，它也会告诉你为什么。”

    宾西法尼亚大学沃顿商学院（Wharton School of the University of Pennsylvania）的管理学教授彼得•卡培里却对这项新技术持怀疑态度。他认为：“技术本身不是问题，问题是有些公司自己都不知道自己想要的是什么。”

    有些东西就连招聘经理也不能用语言来表达出来，要想量化这种无形素质，几乎是不可能的。同时卡培里也表示，即便真的可以做到这一点，用软件来取代真人招聘也是目光短浅的表现。经济困难时期，向招聘者身上投入大量资金很难说得通，但是说到底，最适合招聘和评价顶级人才的，仍然是真人，而不是电脑。

    另一家招聘网站Path.to的创始人达伦•庞兹表示，他的公司以及一些类似的公司并不打算取代真人招聘，他们只是想让招聘经理们的工作变得更轻松、更有效率。

    庞兹把Path.to比作“招聘的‘相亲网’”，他说：“我们并不是想直接促成‘婚姻’，我们只是想把用户介绍给我们认为能跟他们能‘来电’的人。”

    Path.to也利用评分系统来替招聘人员筛选人才。不过它非常关注用人公司的企业文化，因为它希望除了任职资格之外，求职者在别的方面也能更符合招聘方的要求。

    Path.to的理念是，他们可以为企业节省招聘成本和招聘时间，使招聘人员可以把更多的时间花在最有潜力的求职者身上。虽然Path.to的合作伙伴主要是中小型企业，但它的客户也包括了在线影碟租赁公司Netflix和福克斯新闻（Fox News）等知名企业。

    这些新兴的招聘公司指出，许多企业一开始对求职者进行首轮筛选时，招聘人员采用的衡量信息的方法往往是不恰当的。

    Gild是继Bright和Path.to之后另一家主要的新型招聘公司，它采用了奥克兰运动家队（Oakland Athletics，是一支主场位于加州奥克兰的美国职棒大联盟球队——译注）的经理比利•宾用来给球队选秀的方法筛选求职者。

    Everything seemed to be going right for startup founder Steve Goodman, except for one major detail.

    Even during a recession where the unemployment rate flirted with double digits, he could not find computer programmers that were both qualified and would fit into his tech startup's culture.

    "Sourcing and recruiting the best people is finding the needle in the haystack," he says.

    So Goodman decided to start Bright, one of several new companies that rely on powerful computing and the billions of pieces of job data available on the Internet to create algorithms that they believe make the job hunt more efficient, inexpensive, and successful for both the jobseeker and the recruiter.

    The Internet initially seemed to make job hunting easier. With a few clicks, someone looking for work could apply for a dozen positions, and employers' job postings could reach thousands of eyes. Some large companies saw the number of resumes they received increase five-fold.

    But this democratization of job searching soon proved false. Those looking for openings often grow discouraged when they send out several resumes and receive no feedback. Recruiters are inundated with resumes and are forced to weed out candidates based on superficial and unreliable information such as what college someone graduated from.

    Almost two years ago, Goodman started Bright and gathered millions of resumes and job descriptions to create a system that develops a Bright Score, which tells recruiters what kind of fit a person would be for a specific job based on algorithms that are constantly learning as they devour more data, he says.

    "A recruiter now only has to look at 50 resumes because we cut it down for them," he says. "It also tells job seekers whether they are a good fit, and if not, why."

    Peter Cappelli, the George W. Taylor Professor of Management at the Wharton School of the University of Pennsylvania, is skeptical of this new technology, though. "The technology is not the problem," he says. "The problem is companies that don't know what they want."

    Trying to quantify intangibles that not even the hiring manager can put into words is impossible, Cappelli says, and even if it were, replacing human recruiters with software is shortsighted. The financial investment in recruiters can be hard to justify in tight economic times, but humans, not computers, are best suited to recruit and judge top talent, he says.

    Darren Bounds, the founder of another new job site, Path.to, says that his business and others like it are not trying to replace human recruiters. They are just trying to make their jobs easier and more efficient.

    Billing Path.to as "eHarmony for hiring," Bounds says, "We don't want to follow the relationship through to marriage. We really want to introduce you to someone we feel you have a strong bond with."

    Path.to also uses a score to weed out candidates for recruiters. It focuses heavily on company culture in an attempt to match up candidates that will be a good fit beyond their qualifications.

    The idea is that the cost and time saved for companies will allow recruiters to spend more time with the candidates that have the most potential. Although Path.to works mostly with small- to mid-sized companies, Netflix (NFLX) and Fox News are also among its current customers.

    One of the problems, these new job search companies argue, is that during the initial perfunctory weeding out of candidates, recruiters are weighing information improperly.

    In the same way Oakland A's manager Billy Bean reexamined which statistics make a good ball team, Gild, a third major player in the race for improved job search, is exploring what makes a successful hire.