Google Recruitment in the Era of Big Data

Applied to employment, the promises of “Big Data” are both exciting and terrifying, as we emphasized recently. In an interview with The New York Times, Laszlo Bock, Google’s human resources manager, spoke on Google’s use of Big Data for recruitment and had much to say. His observations about it — or as we understand him — mainly point to its limits.

Management is often done through instinct, he says, and many people, even the worst, think they do it well. But the reality is that very few people are good team leaders. A few years ago, Google launched a study to find out who the good recruiters at Google were:

“We looked at tens of thousands of interviews and everyone who had done the interviews and what they scored the candidate and how that person ultimately performed in their job. We found zero relationship.”

Google has conducted several studies to find out how many candidates should be interviewed for each position, who are the best-suited to lead the interviews and what kinds of candidate characteristics tend to best predict their success at Google. On the leadership side, Google has tried to understand what people are good leaders and how Google could cultivate these qualities. Google has also observed how people work together and found that the average size of a group at Google was six people:

“Which teams perform well and which don’t? Is it because of the type of people? Is it because of the number of people? Is it because of how they work together?”

On the recruitment side, the brainteasers with which Google challenged candidates — posing questions such as, “How many golf balls can you fit into an airplane?” — proved to be a complete waste of time. Bock seems most satisfied with behavioral interviews, where one asks the candidates, for example, to talk about a difficult problem they solved:

“When you ask somebody to speak to their own experience you get two types of information. You get to see how they actually interacted in a real-world situation and a sense of what they consider to be difficult.”

On the management side, Google has found that for leaders it is important that people know they are consistent and fair in the way they make decisions. If the leader is clear and predictable, his/her teams experience a lot of freedom because they know how far they can go with what they want, “If your manager is all over the place, you’re never going to know what you can do, and you’re going to experience it as very restrictive.” The manager treating people with respect, giving clear objectives, sharing information and treating the team fairly is fundamental.

Twice a year, each person who has a manager is asked about his/her qualities, using a grid containing 12 to 18 different factors. These answers are shared with the manager whose improvement is tracked, “Over the last three years, we’ve significantly improved the quality of people management at Google, measured by how happy people are with their managers.” We make it more difficult to be a bad manager, says Bock. Information leads people to change their behavior:

“One of the applications of Big Data is giving people the facts and getting them to understand that their own decision-making is not perfect. And that in itself causes them to change their behavior.”

Data has shown that academic scores are not a good criterion for hiring, except with brand new university graduates. Furthermore the proportion of people without a degree has increased over time — 14 percent of Google employees never went to college, Bock emphasizes. For Google’s human resources department, this is explained by the artificiality of academic environments, where students must provide the answer expected of them, so it is much more difficult to resolve problems there where there is no obvious answer.

Still, Big Data is not enough “because it will always require an element of human insight.” In the leadership domain, success is very dependent on the context. What works for Google does not necessarily work the same way in another company. Must we do what the system tells us?

“Is this really what we want to do? Is that the right thing?”