Summary of Work Rules
This is a summary of Work Rules, a 2015 book about Google’s culture by Laszlo Bock, HR head:
There are multiple ways to build a successful company, and the purpose of this book is to share what worked for Google, not to imply that that’s the only way.
Sometimes companies object to what Google does saying they don’t have as much profit as Google. But a lot of what Google does costs little. Culture is not about colorful offices or foosball tables, but about underlying values and assumptions. Even supermarket chains have been able to implement these values. And generated higher financial returns, so treating people well is both a means and an end.
Founder mentality: Being a founder is a question of mentality, not stake. Founders take responsibility for what they see around them and improve them, whether or not it’s part of the job description, whether or not it’s even permitted. Founders create room for other founders alongside them. Being a founder is about attitude, not equity.
Culture
Culture eats strategy for breakfast.
Google’s cultural cornerstones are:
Mission: The right kind of mission — organise the world’s information and make it universally accessible — motivates people. The wrong kind, like Procter & Gamble’s about making better products to increase sales and profit, doesn’t.
Transparency: Engineers see most code, board of directors presentations, snippets describing people’s work this week, product plans and so on. If you send an email complaining about someone, they can be CCed on the email, forcing them to work out their differences.
Voice: Employees have a voice, such as asking questions at TGIF, or emailing VPs suggestions.
Highly mobile engineers demand an inspiring mission, transparency and a voice.
Give people slightly more trust, freedom and authority than you’re comfortable. If you’re not nervous, you haven’t given them enough.
Hiring
Acquihire means acquiring a company for its employees, who have proven success, and shutting down the product. It’s a way of hiring. And it’s costly — a million dollars per employee plus ongoing costs like salaries and bonuses.
2/3rds of mergers and acquisitions fail.
Don’t be too narrow in your hiring. For example, hire software engineers who can learn, not frontend engineers. Some teams may require specific skills, but even in those teams, mix in some generalists.
Don’t do more than four interviews. Each such interview improves outcomes by only 1%.
When a manager is being hired, one of the subordinates should interview him, not just people above him and his peers. Likewise for promotions.
Management
Project Oxygen found that great managers have the following attributes:
Be a good coach.
Empower the team and don’t micromanage.
Care about subordinates’ success, career development and personal well-being.
Be productive and results-oriented.
Be a good communicator; listen and share information.
Have a clear vision and strategy for the team.
Have important technical skills that help advise the team. This is less important than the above, but still essential to be a great manager.
Managers can have three types of conversations managers with their subordinates:
Evaluative conversations (“How you’re doing”)
Development conversations (“How can you improve?”)
Pay conversations (“I want more pay!”)
Don’t mix them up. When you’re having a conversation, be clear which type of conversation you’re having.
Managers should pay attention to both tails — underperformers and overperformers.
People are motivated by a purpose and value, which transcends the day to day work. Managers should help people find their meaning.
Avoid status symbols like reserved parking for VPs.
Training
The average American worker spends half an hour a week learning. American companies spend more than $100B a year on training, which is more than the GDP of many countries. But all this money is wasted.
To learn, you need a plan. My tennis plan is to a) not hold my breath while hitting the ball, which is a bad habit that tires me out quickly. b) to hit the ball before it comes to my body. c) pay attention to my footwork. These are the only three areas I’m focusing on now. After I get better, I’ll focus on other areas. This is a plan. You need a similar plan for effective learning at work: what your goals are, what skills you need for that, what skills you already have, and how you want to develop. Corporate training goes waste because people simply show up and attend training without a plan.
To learn, you also need feedback, like a tennis coach gives on your game. Otherwise, you’ll repeat your mistakes forever. Corporate training doesn’t give feedback, so it goes waste.
Pay
There are two types of distributions in statistics:
A normal distribution, as this infographic shows:
A power law:
The power law applies to engineers, where the best engineer generates 300 times the value of an average engineer, according to Alan Eustace, former SVP of Engineering at Google. Or 10,000 times, according to Bill Gates. If you want to write a formula, such extremes require exponents to characterise, hence the name power law.
Another example of a power law is wealth, where the richest 1% push up the average wealth significantly.
On the other hand, the tallest 1% of people don’t push up the average height significantly — nobody is 10,000 feet tall. So wealth should be characterised by a power law, and height by a normal distribution.
Many people make the mistake of overusing normal distributions to situations it doesn’t apply. For example, HR departments apply it to pay, thinking the best performer can be only somewhat better than the average. In too many companies, people who do great things are not justly rewarded. Sometimes, this is because their contributions are not recognised. Sometimes, this is because the budget is spread over many people, leaving insufficient budget for the exceptional performer.
Paying exceptional performers only a little more than average performers forces people to switch jobs every few years to earn their potential. At Google, an engineer can contribute 100x that of another engineer, and Google did give one engineer $1,000,000 in stock while giving his teammate, who was at the same level, $10,000. This happens rarely. This means that the outperformer earned more than people one level above him.
This requires you to understand what each person’s contribution is. If a person’s contribution is 10x, it doesn’t follow that you need to pay him 10x. That depends on market conditions and what he could have earned elsewhere. But you should pay him at least 5x.
Rewarding the best people fairly will take away budget from rewarding average performers, and that’s a good tradeoff. It also encourages others to contribute more.
When you give extreme rewards, managers should be prepared to explain why, if it leaks out, with specifics on what the person approaching the manager can do to earn the same reward. Otherwise, it breeds jealousy.
Celebrate achievement, not rewards.
Otherwise, rewards can backfire. An example is Founders Awards. The company introduced these awards telling employees they are startup-like rewards, and that people can earn a million dollars. These awards were given to teams that made exceptional contributions like building Chrome. This made everyone unhappy. Founders Awards were often given to engineering and product teams, making the business half of the company feel undervalued. Even in engineering, some teams like ads and Chrome had clearly visible impact, while the non-core teams felt undervalued. Even in core teams, while some awards were high, like a million dollars of stock, the minimum was only $5000, disappointing people who heard the hype about million-dollar Founder Awards. In fact, most people got far less than a million dollars. Even those who won a million dollars immediately wanted to transfer to another team since they knew they can’t win the award twice for the same product. The Founders Awards made everyone unhappy and had negative consequences. So Google quietly stopped giving them. The mistake was in celebrating pay, such as a million dollars or startup-like pay.
Rewards should have distributive justice: are the outcomes fair? In other words, are rewards being allocated based on contributions? People also expect them to have procedural justice: the process by which they’re allocated should also be perceived as fair. In this example, excluding non-engineers who may have delivered outsized impacts immediately violated procedural justice. People are very sensitive to these.
Reward people with experiences (a trip to Bali) or products (a tablet) over cash. People consider the former memorable and thoughtful. On the other hand, cash is evaluated rationally like “This is not even one month’s pay!” Money disappears, while memories last.
Reward smart and thoughtful risk-taking to overcome the stigma around failure that would discourage others from taking even a considered risk. Rewards can also soothe the pain of failure thus allowing people to pay attention to the learning that comes from failure. Larry says that if you have a wildly ambitious goal and fail, you may achieve more than if you didn’t have an ambitious goal in the first place, so it’s not a failure.
HR
HR teams shouldn’t hire only people with a HR background, since they’ll get a lot of incompetent people. Google hires only a third of HR team members from a HR background, the second third from strategy consulting, and the final third from analytics backgrounds, who crunch data to measure and improve the effectiveness of the HR team as a whole. Google’s HR teams has lawyers, PhDs, olympic athletes, and world-record holders.