Google is well-known for hiring only the best of the best. They're incredibly selective, hiring only 0.25% of candidates who apply, making it harder to get a job at Google than it is to get into Harvard University (Work Rules!, page 69).
In his book Work Rules!, Laszlo Bock (Senior Vice President of People Operations at Google from 2006 to 2016) explains why Google's hiring process is so rigorous –– it's more effective to spend time hiring the right people than it is to hire average performers and then try to train them to be top performers (Work Rules!, page 59).
Given that philosophy, Google spends a lot of time and energy on their hiring process. They've had rough patches along the way, but the lessons they've learned are invaluable. Let's take a look at some of their advice and practices below.
When you have an open job position to fill, it's tempting to rush to hire someone. It becomes even more tempting when a position has stayed open for a couple months or more. After all, you have work that needs to be done, and surely someone is better than no one, right? This kind of thinking leads to hiring average or below average performers, which negatively impacts your team.
Instead, hire slowly (Work Rules!, page 62). Don't compromise on your standards. Wait to find the right person –– it will be worth it.
When hiring, you might feel threatened by an amazing candidate. Will they outperform you? Will they end up taking your job? Will having this person on your team or at your company hurt your career or stifle your growth?
Those are entirely the wrong thoughts to have! One of the best ways to be promoted is to build up your team and prepare people to replace you so that you can move on to bigger and better work.
By hiring people who are better than you, at least in one meaningful way, you help increase the overall talent level and diversify the skillset of your team. If each additional hire is "above average", each new person raises the bar and make the company an even better place.
Managers at Google aren't allowed to decide on their own who to hire. While this may be frustrating to many managers, it helps ensure that the hiring process is fair.
At Google, hiring committees are made up of the candidate's potential manager, peers, and subordinates. The committees are also cross-functional, including people from multiple different departments. Each interviewer submits individual feedback, and then all the pieces of feedback are reviewed by the committee.
Google has done meta-analyses on thousands of their interviews, comparing the actual hiring decision against the scores that reviewers gave for whether or not a candidate should be hired. They've found that over time the wisdom of the crowd always outperforms a single individual's judgment (Work Rules!, page 109).
Just about anyone applying for a job can find one or two people who will give them a glowing review. For that reason, references provided by the candidate really aren't very helpful.
To combat this, Google checks for connections on LinkedIn or in college alumni databases to find current Google employees who know the candidate. Asking these "backdoor" references what they think about the candidate and whether they'd like to work with that particular person provides a much more honest viewpoint (Work Rules!, page 71).
If you were to search a site like LinkedIn or Indeed for job postings for "software engineer" right now, you'd likely find hundreds of open positions that are filled with meaningless jargon and vague expectations for candidate requirements. You'd likely struggle to find what the specific position is actually for and what you'd be working on if you took the job.
If you're posting open positions, you should make job postings "excruciatingly specific" (Work Rules!, page 83). What specifically does this job entail? What team is this candidate going to be on? What is the team like? What are the day-to-day responsibilities? What exactly does someone in this position work on? What skills are actually needed in order for the candidate to be successful?
Not only does including this level of detail help give the candidates applying for the job more information, it also helps you in your hiring process because you will (hopefully) be attracting the right people with the right skillset.
Have you ever been asked a question like the following as part of a job interview?
You are standing outside of a room with no windows. The room has three light bulbs and three switches outside of the room. Each switch controls one of the light bulbs. You may only enter the room one time. How can you find out what switch goes to each light bulb?
Depending on who you are, this kind of question either excites you or makes you cringe. Either way, it turns out that brainteasers are incredibly poor interview questions. Research shows that there is no direct link between job performance and being good at insight problems.
This is partially because the brainteaser problems are almost always irrelevant to the job, can be practiced or searched for online, and don't require fluid intelligence (and fluid intelligence does in fact correlate well with job performance) (Work Rules!, page 89).
(In case you're wondering, the answer to the problem is to turn on the first switch, wait a few minutes, turn the switch off, turn on the second switch, and then go into the room. The lightbulb that is off but warm is controlled by the first switch, the lightbulb that is on is controlled by the second switch, and the lightbulb that is off and cold is controlled by the third switch.)
So, what sort of questions should you ask in an interview? Google looked at all sorts of question types and interview methods to see which ones correlated best with actual job performance. They found that the best indicator of future job performance is how well a candidate does in a "work sample test" (Work Rules!, page 91).
This test, simply put, involves the candidate doing something that resembles what they would actually do in their job. For a software engineer, this might be a take-home assignment, an online coding assessment, or a whiteboard interview.
Tied in second place for best predictors of job performance are 1) testing for general cognitive ability, and 2) using structured interviews with pre-defined questions and grading criteria (Work Rules!, page 91).
This shouldn't come as a surprise, but Google also found that using a combination of assessment techniques throughout the interview process was more effective than using a single technique (Work Rules!, page 94). So, given what we learned above, interviews should consist of a work sample test, a general cognitive ability test, and include structured questions and grading criteria.
Google used to have an extremely long and drawn-out interview process. Interviewing with Google consisted of anywhere between 15 and 25 interviews and could take six months or longer before you were finally hired or rejected (Work Rules!, page 76). This was a major pain point both in terms of the amount of hours Google employees spent interviewing people and in terms of creating a negative experience for candidates.
As always, Google did a bunch of research and analyzed their data. They found that "four interviews were enough to predict whether [they] should hire someone with 86 percent confidence. Every additional interview after the fourth added only 1 percent more predictive power" (Work Rules!, page 103).
Improving your hiring process will help you find quality candidates to add to your team and save you time.
In summary, here are the nine hiring tips from Google:
- Hire slowly
- Only hire people who are better than you in some meaningful way
- Hire by committee
- Use "backdoor" references
- Make job postings excruciatingly specific
- Don't use brainteasers as interview questions
- Use work sample tests
- Use a combination of assessment techniques
- Four interviews are enough to know if you should hire someone or not
Thanks for reading, and happy hiring!