How Google Works

by Eric Schmidt

The only way to succeed in business in the twenty-first century is to continually create great products, and the only way to do that is to attract smart creatives and put them in an environment where they can succeed at scale. The problem is, as we have discovered, in a large company it becomes more and more difficult to create that environment.

With process comes scale; you absolutely need process to grow a company profitably. At some point, though, process starts to take over. It becomes so entrenched that it can trump common sense and cause executives to, as our head of business operations, Kristen Gil, says, “lose muscle memory.” People stop thinking and instead just depend on the process to make decisions for them. As process gets better, judgment can weaken.

But Susan (and any other start-up CEO) doesn’t think of budget just in terms of operating expenses to run the business. It’s an investment in growing the business. The typical MBA finance jock might want to allocate the budget by estimating the cost of capital, setting that as the target hurdle rate, ranking all projects based on their IRR (internal rate of return), NPV (net present value), or CUF (complex useless formula), and then doling out the cash until it runs out. The typical CEO knows that the budget shouldn’t be set by how much revenue the business will generate in that fiscal year, but rather the potential ultimate market value of the venture some years down the road.

The best products are still the ones that are based on technical insights, those unique ideas that apply one or more technologies in a new way to solve big problems. When it comes to evaluating where to create autonomous units, this is the primary thing we look for: not a business strategy or financial model (although those are necessary), but a strong set of technical insights.

The tailwind of Google’s marketing and PR engine and brand wasn’t nearly strong enough to overcome a headwind of mediocrity. As Jeff Bezos, founder and CEO of Amazon, says: “In the old world, you devoted 30 percent of your time to building a great service and 70 percent of your time to shouting about it. In the new world, that inverts.”

Unfortunately, like Jonathan’s failed gate-based product development framework, most management processes in place at companies today are designed with something else in mind. They were devised over a century ago, at a time when mistakes were expensive and only the top executives had comprehensive information, and their primary objectives are lowering risk and ensuring that decisions are made only by the few executives with lots of information. In this traditional command-and-control structure, data flows up to the executives from all over the organization, and decisions subsequently flow down. This approach is designed to slow things down, and it accomplishes the task very well. Meaning that at the very moment when businesses must permanently accelerate, their architecture is working against them.

The primary objective of any business today must be to increase the speed of the product development process and the quality of its output. Since the industrial revolution, operating processes have been biased toward lowering risk and avoiding mistakes. These processes, and the overall management approach from which they were derived, result in environments that stifle smart creatives. Now, though, the defining characteristic of today’s successful companies is the ability to continually deliver great products. And the only way to do that is to attract smart creatives and create an environment where they can succeed at scale.

We think it’s particularly important for teams to be functionally integrated. In too many places, employees are segregated by what they do, so product managers might sit here but the engineers are kept in that building across the street.

For a meritocracy to work, it needs to engender a culture where there is an “obligation to dissent”.41 If someone thinks there is something wrong with an idea, they must raise that concern. If they don’t, and if the subpar idea wins the day, then they are culpable. In our experience, most smart creatives have strong opinions and are itching to spout off; for them, the cultural obligation to dissent gives them freedom to do just that. Others, though, may feel more uncomfortable raising dissenting views, particularly in a public forum. That’s why dissent must be an obligation, not an option.

We believe in staying functionally organized—with separate departments such as engineering, products, finance, and sales reporting directly to the CEO—as long as possible, because organizing around business divisions or product lines can lead to the formation of silos, which usually stifle the free flow of information and people.

At the most senior level, the people with the greatest impact—the ones who are running the company—should be product people. When a CEO looks around her staff meeting, a good rule of thumb is that at least 50 percent of the people at the table should be experts in the company’s products and services and responsible for product development. This will help ensure that the leadership team maintains focus on product excellence. Operational components like finance, sales, and legal are obviously critical to a company’s success, but they should not dominate the conversation.

What is the culture that defines your company today (not the one described by the mission or value statements, the real one that people live in every day)? What problems has this culture caused with the business? It is important not to simply criticize the existing culture, which will just insult people, but rather to draw a connection between business failures and how the culture may have played a hand in those situations. Then articulate the new culture you envision—to borrow Nike’s advertising phrase from the 2010 World Cup, “write the future”—and take specific, high-profile steps to start moving that way.

In fact, it’s fine to have a plan, but understand that it will change as you progress and discover new things about the products and market. This rapid iteration is critical to success, but equally important is the foundation upon which the plan is built. The tectonic, technology-driven shifts that characterize the Internet Century have rendered some of the commonly accepted strategic fundamentals we learned in school and on the job incorrect.59 So although your plan might change, it needs to be based on a foundational set of principles that are grounded in how things work today and that guide your plan as it shape-shifts its way to success. The plan is fluid, the foundation stable.

Product leaders create product plans, but those product plans often (usually!) lack the most important component: What is the technical insight upon which those new features, products, or platforms will be built? A technical insight is a new way of applying technology or design that either drives down the cost or increases the functions and usability of the product by a significant factor. The result is something that is better than the competition in a fundamental way. The improvement is often obvious; it doesn’t take a lot of marketing for customers to figure out that this product is different from everything else.

You will never disrupt an industry or transform your business, and you’ll never get the best smart creatives on board, if your strategy is narrowly based on leveraging your competitive advantage to attack related markets.

The right strategy has a beauty to it, a sense of many people and ideas working in concert to succeed. Start by asking what will be true in five years and work backward. Examine carefully the things you can assert will change quickly, especially factors of production where technology is exponentially driving down cost curves, or platforms that could emerge.

Smart coaches know that no amount of strategy can substitute for talent, and that is as true in business as it is on the field.

Most people, when they are hiring for a role, look for people who have excelled in that role before. This is not how you find a learning animal.

One way to make expanding the aperture work is to judge candidates based on trajectory. Our former colleague Jared Smith notes that the best people are often the ones whose careers are climbing, because when you project their path forward there is potential for great growth and achievement. There are plenty of strong, experienced people who have hit a plateau. With those candidates, you know exactly what you are getting (which is good) but there is much less potential for the extraordinary (bad).

The simple way to keep recruiting in everyone’s job description is to measure it. Count referrals and interviews. Measure how quickly people fill out interview feedback forms. Encourage employees to help with recruiting events, and track how often they do. Then make these metrics count when it comes to performance reviews and promotions. Recruiting is everyone’s job, so grade it that way.

In general, when you are CEO you should actually make very few decisions. Product launches, acquisitions, public policy issues—these are all decisions that CEOs should make or heavily influence. But there are many other issues where it is OK to let other leaders in the company decide, and intervene only when you know they are making a very bad call. So a key skill to develop as the CEO or senior leader in a company is to know which decisions to make and which to let run their course without you.

Computer scientists hate inefficiency, so over the years Eric’s team developed a series of rules for meetings that we found to be quite effective: Meetings should have a single decision-maker/owner. There must be a clear decision-maker at every point in the process, someone whose butt is on the line. A meeting between two groups of equals often doesn’t result in a good outcome, because you end up compromising rather than making the best tough decisions. Include someone more senior as the decision-maker. The decision-maker should be hands on. He or she should call the meeting, ensure that the content is good, set the objectives, determine the participants, and share the agenda (if possible) at least twenty-four hours in advance. After the meeting, the decision-maker (and no one else) should summarize decisions taken and action items by email to at least every participant—as well as any others who need to know—within forty-eight hours. Even if a meeting is not a decision-making meeting—for example it’s designed to share information or brainstorm solutions—it should have a clear owner. Again, that owner should ensure that the right people are invited to the meeting, that there’s a clear agenda, that the necessary prep work has been done in advance, and that action items are circulated promptly. Meetings are not like government agencies—they should be easy to kill. Any meeting should have a purpose, and if that purpose isn’t well defined or if the meeting fails to achieve that...

One of the most important decisions any business leader makes is how to spend his or her time. When Eric became CEO of Novell in 1997, he got some great advice from Bill Gates: Spend 80 percent of your time on 80 percent of your revenue. But this rule can be deceptively hard to actually follow.

Leadership teams often underestimate how long it takes for revenue from a new product area to ramp up. That shiny new stuff can be much more interesting than the boring old core business stuff, but it’s the core stuff that pays the bills, and if you make a mistake there, you probably won’t be able to recover.

The right approach is to look for the outstanding smart creatives who are already progressing rapidly through the ranks. Ask the question, Could one of these people be running the company in ten years? When the answer is yes, give them a lot of compensation and make sure their career doesn’t bog down. Losing these high-potential employees (especially to competitors) is very costly to the company, so be proactive and aggressive in your efforts to keep them happy.

This is where the “share everything” rubber hits the “but you can’t possibly mean everything, can you?” road. Every quarter, well-meaning and otherwise Googley Googlers highlight sentences and paragraphs with the red marker of death (digitally, of course—no paper was inked in the evisceration of this letter). “We can’t put that in the letter,” they say. “What if it leaks? It would cause problems.” Or “We can’t tell employees this, even though it’s true and it’s what we told the board. It might hurt morale.” Fortunately, the people running this process understand that “share everything” doesn’t mean “share everything that wouldn’t look bad if it leaked and that doesn’t hurt anyone’s feelings,” it means “share everything except for the very few things that are prohibited by law or regulation.” Big difference! This is why we make everyone who wants something to be removed justify exactly why it needs to come out, and the reason better be very good.

“The essence of being human involves asking questions, not answering them.”144 Eric likes to put this concept to the test when he walks the halls of Google or of the other companies with which he’s involved. When he runs into an exec he hasn’t seen in a while, the pleasantries don’t last long. After a cordial hello he’ll get to the point: “What’s going on in your job? What issues do you have? Tell me about that deliverable you owe me.” This has a couple of results: It helps Eric keep on top of the details of his business, and it helps him know which of his executives are on top of the details of their business. If someone is in charge of a business and can’t rattle off the key issues she faces in a matter of ten seconds, then she’s not up to the job. A hands-off approach to leadership doesn’t cut it anymore. You need to know the details.

Bill also suggests a nice format for 1:1s, which we have adopted with good results: 1. Performance on job requirements a. Could be sales figures b. Could be product delivery or product milestones c. Could be customer feedback or product quality d. Could be budget numbers 2. Relationship with peer groups (critical for company integration and cohesiveness) a. Product and Engineering b. Marketing and Product c. Sales and Engineering 3. Management/Leadership a. Are you guiding/coaching your people? b. Are you weeding out the bad ones? c. Are you working hard at hiring? d. Are you able to get your people to do heroic things? 4. Innovation (Best Practices) a. Are you constantly moving ahead… thinking about how to continually get better? b. Are you constantly evaluating new technologies, new products, new practices? c. Do you measure yourself vs. the best in the industry/world?

Board members want to talk about strategy and products, not governance and lawsuits. (If this isn’t true, get new board members.) Remember this when you establish the rules and agenda of the board, and stick to it even when the going gets tough.

But focus on the user is only half the story. The full sentence should read “focus on the user and all else will follow.” This means that we will always do what’s right for the user, and we trust that our smart creatives will figure out how to make money from it. It could take a while, so sticking it out requires a lot of confidence.166 But it is usually worth it.

Focus on the user… and the money will follow. This can be particularly challenging in environments where the user and customer are different, and when your customer doesn’t share your focus-on-the-user ethos.

At Google, our users are the people who use our products, while our customers are the companies that buy our advertising and license our technology. There are rarely conflicts between the two, but when there are, our bias is toward the user. It has to be this way, regardless of your industry. Users are more empowered than ever, and won’t tolerate crummy products.

The program doesn’t mean that the campus turns into summer camp every Friday, with all the engineers goofing off in (hopefully) creative ways. In fact, 20 percent time is more like 120 percent time, since it often occurs on nights and weekends. But it can also be stored up and used all at once—Jonathan had one product manager take a summer to work on a 20 percent project. Regardless of when you take your 20 percent time, assuming it doesn’t get in the way of doing your regular job, no one can stop you from doing it. Twenty percent time is a check and balance on imperial managers, a way to give people permission to work on stuff they aren’t supposed to work on.

Our constant advice to anyone who wants to launch a 20 percent project is to start by building a prototype, because that’s how you get people excited about the project. Coming up with an idea is pretty easy. Getting a few of your colleagues to join your project and add their 20 percent time to your 20 percent time is a lot harder. This is where the Darwinian process begins. Finding collaborators can be difficult in a nonhierarchical organization, especially for newcomers, because it’s much harder to figure out where to go to get things done. Relationships become highly critical, and since they take time to build (and this isn’t everyone’s natural skill set), you can end up with a lot of great ideas dying on the vine.

The innovation is flowing; good ideas are a-bubblin’. Most of them die before they see the light of day, but a precious few are good enough to reach the promised land. You cue the blog post, push the green button, and pop a cork to celebrate with the team. Then you get back to work, because if you did the job well, the product is not yet done. Voltaire wrote, “The perfect is the enemy of the good.”191 Steve Jobs told the Macintosh team that “real artists ship.”192 New ideas are never perfect right out of the chute, and you don’t have time to wait until they get there. Create a product, ship it, see how it does, design and implement improvements, and push it back out. Ship and iterate. The companies that are the fastest at this process will win.

To be clear, ship and iterate doesn’t mean you have license to push crummy products out the door and then hope to make them better. In fact, Jonathan often warned his team not to launch crummy products and depend on the Google brand name to get early user traction. The products should be great at what they do, but it’s OK to limit functionality at launch. Withholding significant marketing and PR resources at launch helps with this, as customers are much more likely to feel disappointed by a highly hyped product than one that is quietly launched. Then you can expand functionality with new features (and tweak the existing ones) later. As Eric wrote to Googlers in February 2006: “There should be a plan for a ‘wow’ feature reasonably soon after launch.”194 With this approach, users get accustomed to seeing products launch with a high-quality, if somewhat limited, set of functions that they know will expand rapidly after launch.

To innovate, you must learn to fail well. Learn from your mistakes: Any failed project should yield valuable technical, user, and market insights that can help inform the next effort. Morph ideas, don’t kill them: Most of the world’s great innovations started out with entirely different applications, so when you end a project, look carefully at its components to see how they might be reapplied elsewhere.

And don’t stigmatize the team that failed: Make sure they land good internal jobs. The next innovators will be watching to see if the failed team is punished. Their failure shouldn’t be celebrated, but it is a badge of honor of sorts. At least they tried.

Management’s job is not to mitigate risks or prevent failures, but to create an environment resilient enough to take on those risks and tolerate the inevitable missteps.

So it seems to us that incumbent businesses have a choice to make. They can continue to operate as they always have, existing in a world where technology is something to be used not as a tool of transformation but simply to optimize operational efficiency and maximize profits. In a lot of these incumbent businesses, technology is that interesting thing run by that slightly odd group in the other building; it isn’t something that anchors the CEO’s agenda every week. And the impending disruption caused by new competitors entering their markets is something to be fought with battalions of lobbyists and lawyers. Although it might take a long time (and cost a lot of money), this dig-a-moat-and-bury-your-head-in-the-sand approach is bound to end tragically. The forces of technology and disruption are too powerful.

So forgo conventional wisdom, crank up that imagination, and ask yourself what could happen in your industry in the next five years. What could change most quickly, and what will not change at all? Then once you have an idea of what the future could hold, here are some more hard questions to consider. How would a very smart, well-capitalized competitor attack the company’s core business? How could it take advantage of digital platforms to exploit weaknesses or skim off the most profitable customer segments? What is the company doing to disrupt its own business? Is cannibalization or revenue loss a frequent reason to kill off potential innovation? Is there an opportunity to build a platform that can offer increasing returns and value as usage grows?

Do your decision-making processes lead to the best decisions, or the most acceptable ones? How much freedom do employees have? If there is someone who is truly innovative, does that person have the freedom to act on his ideas, regardless of his level? Are decisions on new ideas based on product excellence, or profit? Who does better in the company, information hoarders or routers? Do silos prevent the free flow of information and people?