Career Advice HR Focus

What is the ideal number of job interviewers? Google’s Laszlo Bock explains his data-driven direction

Internet giant’s SVP of people operations takes a scientific approach to HR

Laszlo Bock runs “people operations” at Google, an apt title for a human resources department that seems far more like a data-driven lab than a mere home for HR administrivia. Only one-third of the people he hires for the department have a traditional HR background. The rest are strategy consultants or hold advanced degrees in subjects such as organisational psychology and physics.

Since joining Google in 2006, Bock and his team have examined questions such as: What is the ideal number of interviewers to assess job candidates? His upcoming book, Work Rules!, shares many of those findings – as well as Google’s approach to handling more traditional HR matters, such as promotions and performance reviews. Here he reveals some of those findings.

What led you to write this book?

I have always been a little frustrated and disappointed not only in myself as a leader, but also in how leadership and management work. We spend more time working than we do with our loved ones, than with our kids, than sleeping. With a little science and comparing notes with other organisations and testing at Google, we have been able to figure out ways to make work better. The hope is that through the book we can make work better everywhere.

What is the biggest mistake managers make when it comes to conducting interviews?

Relying on their own opinion. We all think we are amazing at assessing character and candidates, but the research shows that we make an assessment in 10 seconds, based on a first impression. The rest of the time is spent trying to confirm that, even though we don’t know that is what our brains are doing.

The best thing you can do to fix it is to have a bunch of people (we say four) interview every candidate. Make sure it’s not just the manager, but people who are going to work for, and around, this person. Have every person assign a score, average that score, and make your decision based on that. Everybody has some level of error in their assessment, but when you combine the errors together, they cancel out. Some people are a little soft on candidates, some a little hard on them, some are biased in one way, some in another.

The second thing you should do is only hire people who are better than you in some way. Unless you walk away thinking, “That person is better than me at organising things, or running a process, or solving a problem, or selling to customers,” you shouldn’t hire that person.

Google is known for its perks. Do you worry about creating an environment of ever-escalating expectations?

All the time. Human beings are incredibly adaptable. A lot of the things Google does that were special have become standard in Silicon Valley.

There’s a couple of ways to guard against it. We had a gentleman named Shawn Achor come to our re:Work conference last year, and he talked about how you can make yourself happier by spending a few minutes a day sending thank you notes. So we’ve encouraged people to do that. When people behave in a way that appears more grateful, they actually become less entitled.

The last thing is: If it doesn’t make sense anymore, we take it away. We used to offer a $5,000 reimbursement if you bought a hybrid car. It stopped being an incentive, and merely became a subsidy to Toyota to buy Priuses. Googlers were cranky and we explained it. Then we moved on.

Last year Google released its diversity statistics. A lot of other companies followed suit. How did you come to that decision?

It was just the right thing to do. There’s a tremendous diversity problem in the technology industry, in academia, in the paths leading to this industry. But also, when women get into this industry with computer science degrees, they tend to opt out of it after a while, and at pretty significant rates.

We realised that, unless we were open about that, we wouldn’t be able to make the kind of progress that we wanted to at Google and that we thought our industry needed to. So we decided to go public with the data.

We had a lot of conversations about it: Is this going to trigger a lot of lawsuits? What’s going to happen as a result of this? We decided whatever the consequence, the right thing to do is to start this conversation and follow it up with action.

A project at Google called Project Oxygen looked into whether managers mattered. Did the research show any underappreciated leadership traits that managers have?

Two things. One was this idea of emergent leadership. It is stepping in and solving a problem and, importantly, being willing to step back when your piece is done. That made a big difference.

When I joined Google, we used to look for traditional outward metrics or indicators of leadership when we were hiring. What we discovered was our best leaders were not necessarily captain of the football team or president of the chess club. Instead, they were the ones who, when they saw a problem, were conscientious enough to step in and fix it – and low-ego enough and self-aware enough to then let other people fix the next phase.

The second thing, and this was a surprise, was that once you start measuring something, you don’t need to teach people a whole bunch about it. Every six months we ask teams what they think of their manager. What someone’s team thinks is reality: If I think my boss is a jerk, it doesn’t matter if he’s nice to puppies and gives me a gift on my birthday.

So we started publishing this data internally, showing where people were in relation to other Google managers. And even before we built training around it, they started improving. If you survey your people on what they think about your managers, then tell the managers where they fall on the distribution, most managers are going to work to get better.

In the book, you talk about an idea called “Dublin goes dark”. What is that and what is Google doing to help prevent technology from being all-consuming in employees’ lives?

It came out of our Google DNA survey. We found there were “segmenters”, or people for whom work and life are two separate things. They can check their email before going to bed, but they don’t worry about it. Then there were “integrators”, where work and life are bundled. They do the same things, but psychologically it affects them differently. They can’t get away from work. What’s interesting is two-thirds of the company are integrators, but half of those wish they were segmenters.

Our Dublin team said: “Why don’t we try forcing everyone to be a segmenter just for one night a week?” It grew virally. First it was just the people operations team in Dublin, then people operations in all of Europe, and then the Dublin office, which is thousands and thousands of people. People loved it.

But the insight is that you need a peer group to support you. We’ve had people say they’re going to have no-meeting Thursdays, and that works until they’re dependent on another team. Then suddenly you’re getting pinged from outside your group and your Thursdays are shot. I decided I was not going to send emails to my team on weekends – Friday nights to Monday mornings.

The effect: My direct reports don’t send out as many emails, and their direct reports don’t either. The volume has gone down. People just focus on what’s really urgent. 

The Washington Post