Actually, the talent acquisition team at any one of Silicon Valley's major players could probably tell you. Companies like Google use people analytics to figure out the predictive value of interview questions, using follow-up data on factors like new hire job performance and attrition to match questions and candidate answers with positive outcomes. 

The analysis is performed by people scientists, an evolving function at large companies that leverages data about employee performance, accomplishment, job satisfaction and other metrics to bring new and highly useful intelligence into the always-mysterious art of managing people. Successful people management always has been and always will be a type of alchemy, a mix of self-knowledge, experience, empathy, and strategic thinking. Adding credible data can only help: wouldn't you love to identify the three behaviors that reliably predict success or failure where you work?

Lesson One on data is always the same: credibility depends on statistical significance. For data to matter, you need an adequate sample size. The fact that the last guy you had to let go spent fifteen minutes a day staring at the fridge in the break room doesn't mean that anyone mesmerized by appliances is bad at their job. On the other hand, if the last 25 employees you fired all had a thing for the fridge, maybe it's worth looking into.

For a company like Google, boasting over 54 thousand employees worldwide, sample size is not a problem. But how can a smaller organization ever collect enough data on its employees to make people analytics a fruitful part of its operations? Enter the cloud. When a company uses a cloud-based HR platform, it contributes to and draws from data aggregated across hundreds of organizations.

In effect, we're starting to crowd-source HR. The data is anonymized to protect participants from inadvertently revealing sensitive information, but the analytics are there. Finally, employers can start to understand the connection between interview questions, resume content, job description terminology and other factors to new hire performance. Here at Unitive, for example, our data shows that job listings which have been scrubbed for traditionally masculine descriptors ("active," "outspoken," "hyper-competitive") have higher appeal to job candidates -- both male and female!

Of course, crowd-sourced data won't be specifically tailored to your organization, but the more data gets collected, the quicker it can get broken down into useful categories: what works for companies employing under 50 people? What works in consumer services versus B2B? What works for consumer services companies employing under 50 people? Soon we'll know. Thanks to cloud-based HR platforms, the data we need to bring the hiring process into the 21st century is being gathered right now.