How Google Works
The premise of *How Google Works* is simple and correct: the companies that win in technology don't manage their way to victory, they hire better people and then get out of the way. Schmidt and Rosenberg spent years learning this the hard way inside the Googleplex, and the core lesson — that "smart creatives" are the actual source of competitive advantage — is worth the price of the book.
The concept of the smart creative is the book's best idea. Not just talented engineers, not just creative marketers, but people who combine technical depth with business instinct and genuine curiosity about everything else. The book argues, persuasively, that traditional management instincts break badly when applied to this type of person: they don't want to be told what to do, they chafe in hierarchies, and the surest way to lose them is to surround them with bureaucracy. Schmidt and Rosenberg describe Google's response — flat structures, open communication, meetings where disagreement is expected rather than suppressed — and most of it holds up. The minimum-seven-reports rule exists to prevent managers from turning into micromanagers. The organizing principle of any company should be its best people, not its most senior ones. These are the chapters worth reading slowly.
Where the book wobbles is in its treatment of Google's actual history. The claim that the founders always knew advertising would fund the company sits awkwardly next to the 1996 PageRank paper, which explicitly warned that ad-funded search would be biased toward advertisers and away from users. That paper goes unmentioned here. The Android origin story is handled with enough vagueness to protect the company's preferred narrative around the iPhone. The China chapter is more honest — Schmidt and Rosenberg describe the exit clearly — but the pattern across the book is to sand the rougher edges smooth. This is an inside account, written by insiders, and you feel it. The result isn't dishonest so much as carefully curated.
The book also carries a strain of early-2010s Silicon Valley culture that has aged poorly. Schmidt and Rosenberg dismiss work-life balance as a concept for the unserious, describe mothers checking email at 9pm after getting home as admirable dedication, and present campus perks as evidence of genuine care rather than instruments of retention. This was already a live debate when the book came out. A decade of tech labor discourse later, some of these passages land differently than intended.
For engineering-led founders or technical managers, *How Google Works* remains genuinely useful. The talent and culture sections are strong. The strategy chapters are less novel — bet on platforms, think in five-year horizons, don't optimize for short-term revenue — but are grounded in enough specific anecdote to avoid feeling generic. Schmidt and Rosenberg are better when they're telling you what actually happened than when they're constructing universal principles from it. The reader who gets the most from this is probably someone building their first real engineering organization, trying to understand how a company like Google actually operates versus how business school case studies describe it. For anyone else, it's an entertaining piece of company mythology, told with conviction by two people who were genuinely there.
Read the longer summary
The book has one good idea repeated across 304 pages: when technology accelerates, the only durable advantage is the people who can figure out what to build next, and you should organize the entire company around making them productive. Schmidt and Rosenberg call these people “smart creatives” — engineers with deep technical chops who also have business sense, design taste, and an opinion about what to build. Everything else in the book — flatter org charts, hiring committees that override managers, the obsession with iteration over planning — is downstream of that one thesis.
Whether it’s worth your time depends on whether you trust the messengers. Schmidt was Google’s CEO from 2001 to 2011; Rosenberg ran products from 2002 onward. They were both adults flown into a company already invented by Larry Page and Sergey Brin, and much of the book is a memoir of their adjustment to a place that didn’t run on the management literature they had absorbed at Sun, Novell, Apple, and Excite. That tension — seasoned executives discovering that the founders’ instincts mostly beat their training — is the most honest thread running through it. When Schmidt and Rosenberg are being self-congratulatory, the book is unbearable. When they’re describing what they had to unlearn, it’s actually useful.
The smart-creative thesis is doing all the work
The smart creative is the load-bearing concept. The authors describe a person who is technical enough to build the thing, commercial enough to know whether it’s worth building, and insistent enough to argue with leadership about the answer. The framing borrows from Bell Labs and Xerox PARC mythology, but Google’s contribution is to claim the model can scale: not five geniuses in a research wing, but tens of thousands of them as the operating reality of a public company.
The strongest part of the book argues that if you accept this premise, almost everything else has to change. Smart creatives won’t tolerate org charts that put them under managers who can’t review their work. They won’t write proposals for ideas they can prototype faster. They won’t believe a strategy document written by people who haven’t shipped. So you flatten the hierarchy, hand authority to whoever has the deepest insight, and replace planning rituals with iteration. Rosenberg in particular tells a useful story about arriving at Google with a polished MBA-style product roadmap, having Page reject it, and realizing his entire mental model of “product management” was wrong for a company where the engineers were the strategists.
The advice that actually transfers tends to live in this section. Organize around the people doing the highest-impact work even when they aren’t the senior ones. Require managers to have at least seven direct reports so they can’t micromanage. When entering a new organization, find one smart person and ask them who else is smart, because that’s faster and more accurate than reading the org chart. None of this is unique to Google — Andy Grove and Bill Campbell are the obvious ancestors, and Campbell shows up throughout the book as the coach who forced these lessons on the authors. But the framing is concrete enough to act on.
Strategy: three heuristics worth lifting
The strategy chapter makes a genuinely useful claim: in a market where the underlying technology shifts every eighteen months, a five-year plan is mostly fiction. Schmidt and Rosenberg replace it with a small number of operating heuristics, and at least three of them are worth lifting.
First, build products from a technical insight — something true about the underlying technology that competitors haven’t internalized yet — rather than from market research or feature-gap analysis. PageRank is the example they keep returning to. The authors are right that this is the right way to filter ideas in a technology business, and right that most product roadmaps fail this test. The chapter is at its weakest when it tries to retrofit the same logic onto every Google product. Some of those products were genuine technical bets, and some were defensive responses to Microsoft, Apple, and Facebook that the book quietly recategorizes as bets.
Second, build platforms. The authors define a platform broadly — a venue where multiple sides exchange value — and use it to explain Search, YouTube, Android, and Maps. The point lands. A tool that benefits only its user (Excel, Photoshop) is a different kind of business than a venue where users meet other users (eBay, YouTube, the App Store), and the second tends to compound. The chapter works as a forcing function: if you’re building a tool, can you turn it into a platform, and if so what do the other sides look like?
Third, optimize for growth, not revenue. The authors are careful about how they say this — they don’t mean ignore the P&L — but the practical claim is that in winner-take-most software markets, the cost of pricing or advertising your way to short-term revenue is losing the user base that makes the business interesting later. Google itself is the example, and the example is real: in 2002 you could make more money per query by running banner ads, and the decision not to is what allowed Google to cross the chasm from “search engine” to “default verb.”
These three together are the book’s most exportable idea. They aren’t original to Schmidt and Rosenberg, but they’re stated more crisply than in most strategy texts, and the Google examples make them stick.
Hiring is where the book is most honest
The talent chapter is the longest and the most useful. The authors are unembarrassed that hiring is the work — that an executive who delegates hiring is delegating the company — and they describe Google’s process in unusual detail: structured interview rubrics, hiring committees that override individual managers, a deliberate refusal to lower the bar when groups are short-staffed, and a preference for general intelligence and learning velocity over specialist credentials.
The “minimum seven reports” rule is the most-quoted line from the book and it earns the attention. Inverting the traditional management ratio doesn’t just flatten the chart; it forces managers to abandon micromanagement because there isn’t time, and it forces individual contributors to make their own decisions because no one else will. It also creates real failure modes the book mostly skips — managers who never learn to coach, ICs who flounder without structure — but as a default for a company hiring for autonomy, the heuristic holds up.
The chapter is also where Schmidt and Rosenberg are most willing to disagree with each other in print. Rosenberg defends interview panels of more than a dozen people; Schmidt thinks four is enough. The disagreement is the most interesting page in the book because it shows what management at Google actually felt like — two senior people unable to fully agree on a process they had both run for a decade, working it out on the page rather than papering over it.
The chapters that go quiet when it gets hard
The book has a problem that runs the opposite direction of its strengths. When the topic moves from operating principles to specific decisions Google made, the writing gets careful in a way that’s hard to ignore.
The China episode is the clearest example. Google entered the Chinese market in 2006, accepted government censorship, then pulled out in 2010 after a state-backed cyberattack. The authors tell the story as a moment of corporate principle, and there is a real version of that story to tell — the company genuinely did walk away from a market most competitors stayed in. But the version in the book skips the four years of accommodation that came first, treats the decision as obvious in retrospect when it was contested at the time, and never wrestles with the fact that the company has continued to consider re-entry under different terms. A reader who didn’t know the history would come away thinking Google’s relationship with authoritarian markets is settled. It isn’t.
Motorola is a worse omission. Google bought the handset maker for $12.5 billion in 2012 and sold it to Lenovo for $2.9 billion in 2014, the year the book came out. A management book by the company’s executive chairman that describes itself as a candid look at strategy and decision-making and then declines to explain a roughly $9 billion mistake on its way to press is not a book about strategy. The authors gesture vaguely at culture problems and move on. Steven Poole’s Guardian review caught this at the time and it has aged into a clearer indictment.
Smaller things accumulate the same way. Google Reader’s shutdown in 2013 — the closest thing the company had to a betrayal of its power-user base — gets a line about how products are sometimes “retired.” The “Don’t be evil” motto is defended without engaging the obvious objection that a company with Google’s leverage can’t responsibly aim that low. Workforce demographics get mentioned only to be waved off with a claim that Google is “post-gay” — meaning, the authors say, indifferent to identity — without acknowledging that the workforce was 70% male and 61% white at the time. Max Wallis caught the dodge in his Independent review and he was right to.
The pattern isn’t that Schmidt and Rosenberg are hiding things. They aren’t, quite. It’s that any time the topic forces them to choose between being interesting and being on-message, on-message wins. That’s the price of having current-and-former executives write the book, and it’s the structural reason it can teach you how Google manages but can’t teach you what Google believes.
Where Schmidt’s worldview leaks through
The book isn’t politically positioned and the authors don’t try to be philosophers, but their assumptions occasionally surface and they’re worth flagging because they shape the rest of the argument.
The most striking is a line near the end: “With enough data and the ability to crunch it, virtually any challenge facing humanity today can be solved.” It’s the cleanest statement of techno-optimism in the book and also the laziest. Most of the hard problems Schmidt has actually worked on — Chinese censorship, EU antitrust, content moderation, AI safety, the political economics of advertising — are not data problems and were not solved by more data. The line reads now as the kind of claim a 2014 Google executive made on autopilot, and the book doesn’t push back on it because the book agrees with it.
The “Don’t be evil” defense is similar. The authors clearly believe the slogan is sincere, and the slogan probably was sincere when Page and Brin wrote it. But as a normative bar for a company that mediates a third of the world’s information traffic, “don’t be evil” sets the floor at “not a felony,” and the book never engages with the gap between that floor and the standard you’d actually want. Steven Poole’s joke in The Guardian — that his own moral standard is “don’t be a serial killer” and he therefore expects sainthood — is the right register.
A third tell is small but worth noting. The book casually approves of employees who turn email and messaging back on at 9 p.m. after their kids are in bed, and treats this as a sign of cultural commitment rather than a workforce problem. A reader in 2026, after a decade of conversation about what always-on actually costs, will read that paragraph differently than the authors did when they wrote it.
What you learn, and what you don’t
If you’re trying to figure out how to run an engineering organization, the book is worth the time. The hiring chapter alone is the best short statement of why hiring committees beat hiring managers, and the strategy chapter’s three heuristics — technical insight, platforms, growth over revenue — are a useful checklist for any product review. The argument that smart creatives will demand a different operating model is correct, and most companies that try to copy Google’s surface practices (the perks, the 20% time, the colorful offices) without copying the operating model end up with the perks and the politics and not the productivity.
What you don’t get is a theory of Google as a company in the world. The book is firmly inside the Googleplex looking out, and the view from inside is partial in predictable ways. It can’t tell you why Google’s relationship with regulators has soured every year since publication, why the company keeps shipping product reorganizations the press reads as panic, or why the AI race that began the year after this book came out caught Schmidt’s successors flat-footed despite Google having invented most of the relevant research. The reasons live in the same blind spots the book never fills.
Read it for the mechanics. Read Brad Stone’s The Everything Store or Steven Levy’s In the Plex for the company. Read your own engineering team for what your version of smart creatives actually wants — because the most important thing Schmidt and Rosenberg get right is that the answer is local, and the book is most useful when it stops being a Google book and starts being a prompt to ask the same questions about your own organization.