Finding the best engineers, programmers, and sales repre¬sentatives is a challenge for any company, but it’s espe¬cially rough for a company growing as fast as Google. In recent years, the company has doubled its ranks every year and has no plans to slow its hiring. More than 100,000 job applications pour into Google every month, and staffers have to sort through them to fill as many as 200 positions a week. Early on, the company narrowed the pool of applicants by setting a very high bar on traditional measures such as academic success.
For example, an engineer had to have made it through school with a 3. 7 grade-point average. Such criteria helped the company find a manageable number to applicants to interview, but no one had really considered whether they were the most valid way to pre¬dict success at the company. More recently, the company has tried to apply its quan¬titative excellence to the problem of making better selec¬tion decisions. First, it set out to measure which selection criteria were important.
It did this by conducting a survey of employees who had been with Google for at least five months. These questions addressed a wide variety of char¬acteristics, such as areas of technical expertise, workplace behavior, personality, and even some nonworking habits that might uncover something important about candidates. For example, perhaps subscribing to a certain magazine or owning a dog could be related to success are Google by in¬directly measuring some important trait no one had thought to ask about.
The results of the survey were com¬pared with measures of successful performance, including performance appraisals, compensation, and organizational citizenship (behaving in ways that contribute to the com¬pany beyond what the job requires). One important lesson of this effort was that academic performance was not the best predictor of success at Google. No single factor predicted success at every job, but a combination of factors could help predict success in particular positions. From this information, Google compiled a set of ques¬tionnaires that were related to success in particular kinds of work at Google: engineering, ales, finance, and human re¬sources. Now people who apply to work at Google go on¬line to answer questions such as “Have you ever started a club or recreational group? ” and “Compared to other peo¬ple in your peer group, how would you describe the age at which you first get into {i. e. , got excited about them, started using them, etc. ) computers on a scale from 1 [much later than others] to 10 [much earlier than others]? ” The data are analyzed by a series of formulas that compute scores from 1 to 100. The score predicts how well the appli¬cant is expected to fit into the type of position at Google.
Michael Mumford, an expert in talent assessment at the University of Oklahoma, says that, in general, this approach to predicting performance is effective, but only when it relies on reasonable measures. So, starting a club might be a way to measure leadership behavior, but own-ing a dog (a measure Google abandoned) should be u. sed only if the employer can find an explanation for why it is relevant. Questions 1. Analyse the case 2. Besides the questionnaires, what other selection methods would you recommend that Google use? How would these improve selection decisions?
