The Hard Thing About Hard Things

Introduction

Most management books offer recipes for challenges that have no recipes, and this book instead provides hard-won lessons from the author’s actual experience as entrepreneur, CEO, and venture capitalist—arguing that the ‘hard thing about hard things’ is precisely the absence of any formula.

• Conventional management books fail because they treat genuinely unpredictable, dynamic situations as though they have reliable recipes, when in reality there is no formula for leading people through crisis, making a hit product, or motivating a team when the business has collapsed.
  • The hard thing isn’t hiring great people—it’s when those great people develop entitlement and make unreasonable demands.
  • The hard thing isn’t dreaming big—it’s waking up in a cold sweat when the dream turns into a nightmare.
• Hip-hop music provides unexpected insight into entrepreneurship because artists in that genre aspire simultaneously to greatness and commercial success, and grapple with themes—competing, making money, being misunderstood—that parallel the founder experience.
  • The book draws on rap lyrics throughout because hip-hop artists see themselves as entrepreneurs navigating a hostile world.

From Communist to Venture Capitalist

Horowitz’s unconventional upbringing—Communist grandparents, Berkeley counterculture, diverse social circles—gave him the habit of rejecting first impressions and conventional wisdom, a disposition that proved essential when he later had to make decisions that defied market consensus. His path from Silicon Graphics through Netscape illustrates how early encounters with radical uncertainty, competitive crisis, and powerful personalities shaped the instincts he would rely on as a CEO.

• Growing up in Berkeley with card-carrying Communist grandparents and a father who edited the New Left magazine Ramparts gave Horowitz an outsider’s perspective that trained him to separate facts from perception—a skill that later helped him find alternative narratives when conventional wisdom predicted his company’s failure.
  • His grandfather Phil Horowitz lost his job as a schoolteacher during the McCarthy era for being an active Communist Party member.
  • Moving between the football team and advanced math classes in high school, Horowitz observed how radically the same events—like the release of Run-D.M.C.’s Hard Times—registered differently across social groups.
• A childhood encounter—choosing to ask a Black child to share his wagon rather than racially harass him as a troubled older boy demanded—taught Horowitz that being scared does not mean being gutless, and that the surface of a situation conceals the truth underneath; that child, Joel Clark Jr., became his lifelong best friend and best man.
  • The thirty-yard walk felt like thirty miles to the young, terrified Horowitz, but what came out of his mouth was ‘Can I ride in your wagon?’
  • “Following conventional wisdom and relying on shortcuts can be worse than knowing nothing at all.” —Ben Horowitz
• Netscape’s real historical significance was redirecting the computing industry from proprietary Information Superhighway platforms controlled by Microsoft and Oracle toward the open Internet, inventing foundational technologies like JavaScript, SSL, and cookies—a mission Horowitz joined precisely because Marc Andreessen interviewed him on ideas rather than credentials.
  • As late as November 1995, Bill Gates predicted the Information Superhighway—not the Internet—would rule the future, and later quietly revised those references.
  • After the interview, Horowitz phoned his brother and said he thought Andreessen might be the smartest person he’d ever met.
• When Microsoft threatened Netscape’s Web server business by offering a free, faster competing product, Horowitz and Mike Homer responded not with a strategic pivot but with lead bullets—building a dirt-cheap, open alternative to Microsoft BackOffice called SuiteSpot that grew to a $400 million business.
  • Microsoft’s Internet Information Server had every feature Netscape’s product had, was five times faster, and was to be given away free—giving Netscape roughly five months to respond.
  • Mike Homer took every press call during competitive attacks, sometimes with a phone in each ear—the harder the situation, the stronger he got.
• A savage email from Andreessen publicly blaming server product management for Netscape’s $3 billion market cap loss—sent on the same day Andreessen appeared barefoot on the cover of Time—forced Horowitz to confront professional humiliation while also demonstrating that productively combative relationships can endure and strengthen over decades.
  • “Andreessen’s email, CC’d to the CEO and Chairman, ended: ‘Next time do the fucking interview yourself. Fuck you, Marc.’” —Marc Andreessen
  • The relationship works because even after eighteen years, each still upsets the other almost every day by finding something wrong in the other’s thinking.
• A personal crisis—his father’s joke that ‘flowers are cheap but divorce is expensive’—forced Horowitz to recognize that by trying to pursue career, interests, and family simultaneously without priorities, he risked losing everything that mattered most, prompting him to quit NetLabs and begin putting his family first.
  • His daughter Mariah had been diagnosed with autism, making the startup demands a particular burden on the family.
  • He realized he had always thought about himself first and had to stop being a boy and become a man.

“I Will Survive”

The founding and near-destruction of Loudcloud—through the dot-com crash, a harrowing IPO with weeks of cash remaining, consecutive layoffs, the loss of its largest customer, and a forced pivot to selling the cloud business to EDS—demonstrates that startup survival requires ignoring odds, making moves when no good moves exist, and treating employees with dignity even while dismantling what they built.

• Loudcloud was founded on the concept of cloud computing—software developers shouldn’t have to manage security, scaling, and disaster recovery—and grew explosively to $27 million in quarterly bookings and nearly 200 employees within nine months, fueled by the logic that capital was effectively free during the dot-com boom.
  • Benchmark’s Andy Rachleff asked Horowitz to think about how he’d run the business if capital were free, and then the firm raised $45 million in debt with no covenants and no payments for three years—making the question more literal than hypothetical.
  • A Wired cover story billed the company as ‘Marc Andreessen’s second coming.’
• The dot-com crash devastated Loudcloud because most of its customers were dot-com startups that went bankrupt, revealing the fatal fragility of a high-growth infrastructure business built on an imploding customer base—quarterly bookings fell from $100 million forecast to $37 million in reality.
  • The NASDAQ peaked at 5,048.62 on March 10, 2000 and eventually fell below 1,200—an 80 percent drop.
  • “When pitching Softbank Capital, Bill Campbell relayed their reaction: ‘They thought you were smoking crack.’” —Bill Campbell
• The Loudcloud IPO in 2001—executed with six weeks of cash remaining, no celebration from either bank, a stock that had to be re-priced from $10 to $6 per share, and the night before the offering marked by Yahoo firing its CEO—may have been the least celebratory IPO in history, yet it saved the company from certain bankruptcy.
  • The company had $1.94 million in trailing six-month revenue but forecast $75 million the following year, generating press coverage calling it ’the IPO from hell.’
  • While on the road show, Horowitz received a call informing him that his wife Felicia had stopped breathing from an allergic reaction to medication—yet she insisted he stay and finish the IPO.
• I learned the most important rule of raising money privately: Look for a market of one. You only need one investor to say yes, so it’s best to ignore the other thirty who say ’no.’
• Bill Campbell’s role as the ideal board member—the person who is simultaneously the friend you call when something great happens and the friend you call when your life is on the line—proved decisive at every crisis point, including his blunt declaration ‘It’s the fucking money’ when Horowitz was debating whether to pursue the IPO.
  • Campbell began as a college football coach, didn’t enter business until forty, and eventually mentored Steve Jobs, Jeff Bezos, and Eric Schmidt.
  • Campbell’s most important IPO advice was not strategic but operational: stay home during the New York announcement to ensure every employee immediately knows whether they work for Loudcloud, EDS, or neither.
• When Loudcloud’s largest customer Atriax went bankrupt owing $25 million, wiping out the only viable fundraising plan, Horowitz concluded that the cloud business was doomed and secretly began planning to sell it to EDS while pivoting to Opsware—a software company built on the automation technology Loudcloud had developed internally.
  • Project Oxide—the secret effort to separate Opsware software from Loudcloud infrastructure—was disguised as a new product line so that employees wouldn’t abandon the existing business.
  • When two Stanford MBAs presented a 45-slide deck arguing that Oxide was misguided, Horowitz let them finish all 45 slides and replied only: ‘Did I ask for this presentation?’
• The sale of Loudcloud to EDS for $63.5 million—orchestrated with artificial deadlines drawn from Michael Ovitz’s dealmaking philosophy and competitive tension engineered by having IBM and EDS pass each other in the hallway—preserved the Opsware software and team while avoiding bankruptcy, but required immediately and transparently telling 290 employees their fate.
  • “Ovitz’s advice: ‘I believe in artificial deadlines. I believe in playing one against the other. I believe in doing everything and anything short of illegal or immoral to get the damned deal done.’” —Michael Ovitz
  • EDS’s eventual willingness to engage came only after Horowitz and O’Farrell found Jeff Kelly buried in the EDS org chart—the one person who could actually make the decision.

This Time with Feeling

After selling Loudcloud to EDS, Horowitz rebuilt Opsware from 80 employees and a $0.35 stock price into a $1.65 billion acquisition by HP through a series of near-death crises—a 60-day ultimatum from its largest customer EDS, a competitive product obliteration by BladeLogic, and a final acquisition battle—each resolved through unconventional judgment, personal courage, and the willingness to make decisions that maximized team welfare alongside company value.

• The post-sale off-site in Santa Cruz—where Horowitz told 80 remaining employees the full truth about Opsware’s prospects and invited anyone who doubted the mission to quit that day with his help finding a new job—resulted in only two departures, with 76 of the remaining 78 staying through the eventual HP acquisition five years later.
  • Institutional investors had sold all their shares after the EDS sale, pushing the stock to $0.35—computing to a market cap equal to half of the cash in the bank.
  • NASDAQ sent a letter threatening delisting if the stock did not reach $1 within 90 days.
• When EDS—representing 90 percent of Opsware revenue—threatened to cancel its contract after a failed deployment, Horowitz deployed Jason Rosenthal to fix the technical problems with unlimited company resources while Anthony Wright was tasked with finding what EDS secretly wanted but wasn’t asking for—producing a 60-day deadline to fix everything or face death.
  • “Frank Johnson, the EDS executive in charge, said: ‘You fucking want to know what I think about Opsware? I think it’s the biggest goddamn piece of shit!’—then gave the team 60 days.” —Frank Johnson
  • Anthony’s discovery that Frank desperately wanted to keep the Tangram software product led Horowitz to acquire Tangram for $10 million in under 60 days, purely to give EDS what it really wanted.
• When Opsware acquired Tangram, its CFO John Nelli—who was not going to be retained—was diagnosed with brain cancer, and Horowitz spent roughly $200,000 to keep him on payroll long enough to qualify for COBRA health coverage, a decision he described as motivated by understanding what desperation felt like.
  • Fifteen months later, Horowitz received a handwritten letter from Nelli’s wife saying he had died but that Horowitz had saved her from total despair.
• I guess I did it because I knew what desperation felt like.
• When BladeLogic began consistently beating Opsware in large deals and the stock fell back to $2.90, Horowitz told the engineering team they were ‘getting their asses kicked’ and had one bullet left—producing the Darwin Project, six months of seven-day-a-week intensity that rebuilt the product and restored competitive advantage.
  • “Engineer Ted Crossman later recalled: ‘Of all the times I think of at Loudcloud and Opsware, the Darwin Project was the most fun and the most hard. I worked seven days a week 8 a.m.-10 p.m. for six months straight.’” —Ted Crossman
  • Horowitz was shocked to learn years later that the team experienced the Darwin Project as exhilarating rather than crushing—he had assumed he was asking too much.
• The acquisition of Rendition Networks for $33 million to add network automation—against internal advice that the technology was hard to integrate and the company was old—was validated within three months when Cisco prepaid $30 million for advanced licenses, effectively paying 90 percent of the acquisition cost and demonstrating that conventional financial analysis missed the strategic value.
  • Opsware’s weekly staff meeting included a standing agenda item titled ‘What Are We Not Doing?’—which surfaced the network automation gap.
  • Markets weren’t efficient at finding the truth; they were efficient at converging on a conclusion—often the wrong one, as illustrated by Opsware trading at half its cash value when it had a $20 million/year contract.
• The final sale of Opsware to HP for $14.25 per share ($1.65 billion)—achieved by holding firm at that price even after BMC withdrew and HP’s initial counter was $13.75, using Oracle as a ‘rabbit’ in the dog race, and refusing to let an Ernst & Young accounting dispute crater the deal—validated the strategy of treating the company as the premium market leader until the market agreed.
  • “When the board unanimously advised taking HP’s discounted $13.75 offer, Horowitz refused: ‘HP offered $14.25 because we are the gold standard. The second we accept a discounted offer, this deal falls apart.’” —Ben Horowitz
  • After a two-hour pause with no color returning to CFO Dave Conte’s face, HP accepted $14.25.

When Things Fall Apart

The ‘Struggle’—the psychological and operational abyss every startup CEO eventually enters—has no recipe for escape, but can be navigated through radical transparency with employees, willingness to make the best move even when no good moves exist, and the discipline to keep death in mind at all times so that decisions about hiring, firing, and layoffs are made correctly before the crisis hits.

• The Struggle—the state where self-doubt becomes self-hatred, food loses its taste, and the CEO can’t stop hearing the problems even in conversation—is not failure itself but the condition that causes failure in those who are not strong enough to endure it, and every great entrepreneur from Steve Jobs to Mark Zuckerberg went through it.
  • When Bill Campbell told Horowitz he needed to prepare the company for bankruptcy, his voice and eyes communicated that he believed the contingency plan would be the actual plan.
  • The key survival technique is not playing the odds: ‘It matters not whether your chances are nine in ten or one in a thousand; your task is the same.’
• CEOs who project false positivity to protect employees from bad news actually cause more damage than the bad news itself—because employees can see reality clearly, don’t trust leaders who sugarcoat it, and the CEO wastes the cognitive resources of everyone who could help solve the problem.
  • “Horowitz’s brother-in-law Cartheu, a telephone lineman, described a senior AT&T executive who ‘comes by about once a quarter to blow a little sunshine up my ass’—the moment Horowitz realized he’d been doing the same thing.” —Cartheu
  • Trust and communication are inversely related: the more trust, the less communication overhead is required; destroy trust with dishonesty and no amount of explanation will restore it.
• Layoffs done correctly—without delay, with managers laying off their own people, with the CEO addressing the full company first, and with a message that frames the failure as the company’s rather than the individual’s—can preserve cultural continuity even through three consecutive massive rounds, which is otherwise a statistical impossibility for achieving billion-dollar outcomes.
  • Sequoia’s Doug Leone had never seen a company recover from consecutive layoffs and achieve a billion-dollar outcome in twenty years of venture investing—until Opsware.
  • “Bill Campbell’s principle: ‘The message is for the people who are staying’—how you treat departing employees is what remaining employees will judge you on.” —Bill Campbell
• The Struggle is not failure, but it causes failure. Especially if you are weak. Always if you are weak.
• Firing an executive is almost always a failure of the hiring and integration process rather than the executive—caused by poorly defining the role, hiring for lack of weakness rather than strengths, scaling too soon, or hiring the generic executive rather than the right one for this company at this moment.
  • The wrong way to view an executive firing is as an executive failure; the correct way is as an interview and integration process system failure.
  • The most consistently wrong VC and recruiter advice is to hire ‘bigger’ than you need—someone great in 18 months but poor for the next 18 months will be rejected before they get a chance.
• When companies start losing their competitive battles, executives and employees lie not to investors but to themselves—selectively attending only to positive leading indicators and constructing alternative narratives for negative ones, a pattern Andy Grove identified as the root cause of the great CEOs who catastrophically missed quarters during the dot-com unwinding.
  • “Grove explained that humans who build things only listen to leading indicators of good news, urgently acting on positive signals while explaining away negative ones.” —Andy Grove
  • Common self-deception patterns include classifying departing talent as ‘unwanted attrition,’ blaming competitive losses on the rival giving products away, and projecting hockey-stick revenue despite missed milestones.
• When facing an existential competitive threat, the temptation to find a silver bullet—a product pivot, a market segment escape, an acquisition—must be resisted; the only answer is lead bullets, meaning fixing the actual product, as demonstrated both at Netscape against IIS and at Opsware against BladeLogic.
  • “Engineer Bill Turpin told Horowitz at Netscape: ‘Those silver bullets are fine and good, but our Web server is five times slower. There is no silver bullet that’s going to fix that. We are going to have to use a lot of lead bullets.’” —Bill Turpin
  • The lead-bullet approach at Opsware eventually produced a $1.6 billion company; a silver-bullet pivot would likely have produced one-tenth that value.
• Nobody cares about why a company fails—not the media, not investors, not the board, not employees—and the mental energy spent elaborating misery or assigning blame should instead go entirely toward finding the one seemingly impossible way out.
  • “Al Davis’s advice to Bill Parcells when his team was hit with injuries: ‘Bill, nobody cares, just coach your team.’” —Al Davis
  • A great reason for failing won’t preserve one dollar for investors, save one job, or gain one new customer.

Take Care of the People, the Products, and the Profits—in That Order

Building a company that is genuinely a good place to work—through rigorous training, honest feedback, disciplined hiring for strength, and eliminating political behavior—is not merely a cultural nicety but the survival mechanism that keeps the best people engaged when economic incentives evaporate, as Opsware’s recovery from near-bankruptcy demonstrated.

• Hiring Mark Cranney as head of sales—against unanimous opposition from the board and executive staff who objected to his unconventional appearance, his school (Southern Utah University), and his discomforting intensity—demonstrated that hiring for decisive strength in the critical dimension must override hiring for lack of weakness, especially in wartime.
  • A negative reference warned: ‘Mark Cranney walks up to fifty new recruits and says, I don’t give a fuck how well trained you are. If you don’t bring me five hundred thousand dollars a quarter, I’m putting a bullet in your head.’
  • “Horowitz persuaded Marc Andreessen by arguing: ‘If he didn’t have the things wrong with him that you enumerated, he wouldn’t be willing to join a company that just traded at thirty-five cents per share; he’d be CEO of IBM.’” —Ben Horowitz
• A good place to work—where people can focus on their work with confidence that good performance leads to good outcomes—is the only thing that keeps a company alive when things go horribly wrong, because during bad times all the economic reasons to stay disappear and only genuine job satisfaction remains.
  • GO Corporation employees counted their time there among the greatest work experiences of their lives despite making no money, stalling their careers, and being front-page failures—because Bill Campbell built a genuinely good company.
  • When things go poorly in bad companies, the spiral begins: the company declines in value, the best employees leave, the company declines further.
• Training is among the highest-leverage activities a manager can perform because it compounds across all subordinates’ future work hours—Andy Grove’s calculation showing that 12 hours of preparation yielding 1 percent improvement across 10 employees produces 200 hours of value—yet most high-tech companies fail to invest in it at all.
  • Horowitz’s ‘Good Product Manager/Bad Product Manager’ document—written out of frustration that no one had ever defined the product management job—instantly improved his team’s performance and is still read today.
  • The two primary reasons people quit, based on Horowitz’s review of all Netscape exit interviews: they hated their manager’s lack of guidance, and they weren’t learning anything new.
• In good organizations, people can focus on their work and have confidence that if they get their work done, good things will happen for both the company and them personally.
• Political behavior in organizations almost always originates with the CEO—not through intent but through unstructured processes that reward lobbying over performance—and the antidote is airtight, transparent, and consistently enforced processes for compensation, promotions, and organizational design.
  • If a CEO gives an off-cycle raise to an employee who asked for one, the lesson transmitted to every ambitious employee is that the squeaky wheel gets the grease—independent of performance.
  • Hiring people with the right kind of ambition—ambition for the company’s success as the primary driver, with personal success only as a by-product—is the surest protection against political corrosion.
• Management debt—the pattern of making expedient short-term organizational decisions (putting two people in the same reporting box, matching job offers ad hoc, skipping performance management) that accrue compounding hidden costs—is as dangerous as technical debt and eventually forces a painful lump-sum payment.
  • Overcompensating a key employee to match a competing offer destroys pay equity secretly: the employee tells her friends, word spreads, and soon every ambitious person learns that threatening to quit is the best raise mechanism.
  • Without a performance management process, there is no basis for accountability conversations and the ultimate cost is systematically crappy company performance.
• Metrics create employee behaviors—often unintended ones—so managers must test their descriptions of desired results against the behaviors those descriptions will actually produce, rather than assuming that measuring the right numbers will automatically generate the right outcomes.
  • When Horowitz incentivized salespeople to close deals in the first two months of each quarter, deals simply moved from month three of one quarter to months one and two of the next—producing slightly smaller quarters with no real change in behavior.
  • Managing engineering by schedule, quality, and features produced an on-time product with all features and few bugs that was nonetheless mediocre, because none of the features were great.
• The Law of Crappy People—that talent at any title level converges toward the least competent person holding that title as others benchmark themselves against that floor—can only be mitigated through a rigorous, cross-organizational promotion process that explicitly defines skills required at each level and compares candidates against current titleholders.
  • The Peter Principle holds that people are promoted until they reach their level of incompetence and remain there; as Grove notes, it is unavoidable because there is no way to know in advance at what level a manager will fail.
  • The best analogue to the right promotion process is a top karate dojo: to earn a black belt you must defeat an opponent at that level, guaranteeing no new black belt is ever worse than the worst current one.

Concerning the Going Concern

As companies scale from small to large, the CEO’s primary organizational challenges shift from building to maintaining: minimizing politically corrosive behaviors, designing culture through shock-value mechanisms rather than perks, managing the transition from wartime to peacetime leadership, and giving ground grudgingly on communication and decision-making efficiency without letting the organization collapse under its own complexity.

• Effective company culture is built not through perks or yoga classes but through carefully chosen shock-value mechanisms—trivially implementable policies with far-reaching behavioral consequences—such as Amazon’s door-desk frugality signal, Andreessen Horowitz’s $10-per-minute fine for being late to entrepreneur meetings, or Facebook’s ‘Move fast and break things’ motto.
  • Jeff Bezos’s door desks from Home Depot aren’t ergonomic or status-appropriate for a $150 billion company, but every shocked new employee gets the same withering explanation: ‘We look for every opportunity to save money so that we can deliver the best products for the lowest cost.’
  • Andreessen Horowitz’s $10-per-minute lateness fine was designed to communicate that entrepreneurs’ time matters more than VCs’ time—a reversal of the traditional power dynamic.
• Peacetime and wartime CEOs require radically different management styles—peacetime leaders empower broad creativity across multiple objectives while wartime leaders demand strict adherence and alignment to a single mission—and the same behaviors that are virtuous in one context are destructive in the other.
  • “Andy Grove once told an employee who arrived late to a meeting in front of the full room: ‘All I have in this world is time, and you are wasting my time’—behavior that would be catastrophically inappropriate in peacetime but appropriate in a survival fight.” —Andy Grove
  • Google’s peacetime innovation required letting every employee spend 20 percent of time on personal projects; Apple’s wartime under Jobs returning required everyone to move with precision on his exact plan.
• Scaling a company requires giving ground grudgingly—introducing specialization, organizational design, and process only as necessary, accepting the communication and coordination costs these impose, because the alternative of holding a monolithic structure is like an offensive lineman holding his ground against a faster defensive player.
  • Organizational design should be treated as a communications architecture: people who report to the same manager communicate; people distant in the org chart barely communicate at all.
  • The first rule of organizational design is that all designs are bad—any structure optimizes communication among some parts at the expense of others, so the goal is choosing the least of evils.
• Peacetime CEO sets big, hairy, audacious goals. Wartime CEO is too busy fighting the enemy to read management books written by consultants who have never managed a fruit stand.
• CEO succession is almost always difficult because most organizations are built around ‘Ones’ (directors who love making decisions and setting strategy) who surround themselves with ‘Twos’ (operators who excel at execution), meaning internal promotion typically promotes a Two while external promotion disrupts the entire executive team.
  • Microsoft promoted Steve Ballmer—Bill Gates’s literal number two—when Gates stepped down in 2000, leaving natural Ones like Paul Maritz and Brad Silverberg to eventually depart.
  • GE promoted Jack Welch two levels down past all of his superiors in 1981, causing massive executive turnover but ultimately succeeding because Welch was demonstrably a One.
• Great leadership requires three attributes in combination: the ability to articulate a compelling vision especially when things fall apart (the Steve Jobs attribute), the right kind of ambition—for the company rather than oneself (the Bill Campbell attribute), and demonstrated competence in achieving the vision (the Andy Grove attribute).
  • Jobs’s greatest leadership achievement was getting talented people to follow him through NeXT long after it lost its patina, then getting Apple employees to buy into his vision when the company was weeks from bankruptcy.
  • Grove’s competence attribute can be learned; his tolerance for incompetence was legendarily low, and he wrote management books and taught management classes throughout his tenure at Intel.
• The most important and underappreciated CEO skill is managing one’s own psychology—the lonely condition of having no one who fully shares the knowledge needed to validate a decision, combined with responsibility for all failures and a mean test score of 22 out of 100—which is why the greatest CEOs, when asked their secret, consistently answer ‘I didn’t quit.’
  • Three techniques that helped: making friends who have survived similar decisions (psychological grounding), writing the logic out to separate oneself from the decision (clarity), and focusing on the road rather than the wall (not being paralyzed by what could go wrong).
  • Every problem in the company was the founding CEO’s fault: if someone was promoted wrong, if the quarter was missed, if a great engineer quit, if the product had too many bugs—all of it.
• Courage is the most tested CEO quality because the social credit matrix systematically rewards going with the crowd—when you’re right after defying consensus, few remember you made the decision; when you’re wrong after following consensus, blame is shared—yet every time a CEO makes the hard correct decision they become more courageous, and every capitulation makes them more cowardly.
  • “Cus D’Amato: ‘What is the difference between a hero and a coward? No difference. Only what you do. They both feel the same. The man who is yellow refuses to face up to what he’s got to face. The hero is more disciplined and fights those feelings off.’” —Cus D’Amato
  • When a founder called Hamlet refused to sell his company despite most of his board and management preferring the sale, the entire team immediately embraced the decision once made—suggesting their preference had been driven by his initial ambivalence rather than independent analysis.

How to Lead Even When You Don’t Know Where You Are Going

After the near-obliteration of Opsware’s stock price and the NASDAQ delisting threat, Horowitz’s decision to personally go on the road and tell the company’s story—rather than reverse-split or capitulate—illustrates the defining CEO pattern: focusing relentlessly on what needs to get right while resisting the psychological pressure of uncertainty, and evaluating the CEO role through the two questions of whether they know what to do and whether they can get the company to do it.

• When Opsware stock fell to $0.35 and NASDAQ threatened delisting, Horowitz rejected reverse-splitting as capitulation and instead went on the road to tell the company’s story—securing investment from Herb Allen, who said investing in courage and determination when every other CEO was hiding under their desk was an easy decision.
  • Ron Conway directed Horowitz and Andreessen to see Herb Allen of Allen & Company, whose annual invitation-only conference regularly attracts Bill Gates, Warren Buffett, and Rupert Murdoch.
  • Allen & Company buying Opsware stock, Herb Allen personally buying, and several clients becoming major investors propelled the stock from $0.35 to $3 in months.
• CEOs should be evaluated on three dimensions: whether they know what to do (strategy and decision quality), whether they can get the company to do what they know (leadership and operations), and whether they achieve results against appropriate objectives—with the first two being leading indicators that predict future success better than results alone.
  • The best decision Horowitz ever made—selling Loudcloud to EDS and becoming Opsware—would have lost by a landslide if put to a vote of employees, investors, or customers.
  • “Robin Li of Baidu described sitting terrified on IPO day because the stock closed at $122 rather than the $27 he’d prepared to deliver against, making him responsible for results tied to expectations he hadn’t set.” —Robin Li
• The company story—a full articulation of why the company exists, not a mission statement or tagline—is the CEO’s most essential strategic output because it provides context for every employee’s work, aligns interests, and enables distributed decision-making; a company without a story is usually a company without a strategy.
  • Jeff Bezos’s three-page 1997 letter to Amazon shareholders remains the model for a great company story: extended, specific, and compelling enough to get everyone who matters on the same page.
  • When companies fail to tell their story clearly, you hear: ‘These reporters don’t get it’ and ‘We have great technology but need marketing help.’
• The Freaky Friday management technique—literally swapping two warring executives into each other’s jobs permanently—resolved chronic conflict between Opsware’s Sales Engineering and Customer Support organizations more effectively than any mediation, because experiencing the other role’s pressures produced understanding that mere conversation could not.
  • After just one week in each other’s roles, both executives quickly diagnosed the core process issues causing the conflict and implemented fixes that made the two teams the best-collaborating groups in the company through the HP acquisition.
• Predicting whether an executive can ‘scale’ with the company—judging their future fitness before that future arrives—is counterproductive because it corrupts present management, retards development, causes premature replacement with worse executives, and still requires re-evaluation when the scale is actually reached.
  • Managing at scale is a learned skill rather than a natural ability—nobody is born knowing how to manage a thousand people, and it was not obvious that Bill Gates would learn when he was a Harvard dropout.
  • The right question is never ‘Can this executive scale?’ but ‘For this company at this exact point in time, does an executive exist whom I could hire who would be better?’

First Rule of Entrepreneurship: There Are No Rules

The final acquisition of Opsware by HP—nearly derailed by Ernst & Young’s national office demanding contract amendments within 48 hours that required three large banks to act in under a day—illustrates the book’s central theme: when the sky turns purple and every rule you thought you could count on breaks, you must deal with the new reality rather than arguing about the old one, and sometimes the right answer has no precedent.

• Ernst & Young’s national audit partner reversed its own partner’s judgment on Opsware’s revenue recognition of three contracts mid-acquisition, threatening to force a restatement that would destroy a $1.6 billion deal—demonstrating that external parties on whom companies depend will act in their own interest at the worst possible moment.
  • The accounting involved no material cash flow difference between the two interpretations; the restatement itself, not the accounting choice, would have killed the deal.
  • Horowitz, Dave Conte, Mark Cranney, and the legal team stayed up all night drawing relationship maps to find contacts at three large banks who could amend contracts in under 24 hours—and succeeded.
• When BMC withdrew and HP lowered its offer to $13.75 after the accounting scare, Horowitz refused to accept below the original $14.25 on the grounds that any discount would signal Opsware was not the market’s gold standard—the entire premise on which HP had valued the company—and HP eventually accepted the original price.
  • The board unanimously advised taking $13.75; Horowitz said he had not slept and didn’t know if he was right or wrong, but knew he couldn’t get much wronger.
• The accountability versus creativity paradox—whether to punish the engineer who slips her schedule from three months to nine while solving a real architectural problem—resolves by distinguishing accountability for effort (non-negotiable), accountability for promises (contextual by degree of difficulty), and accountability for results (dependent on seniority, difficulty, and risk quality).
  • If you hold people strictly to commitments on genuinely hard problems, your people will stop taking important risks and spend their time covering their butts instead.
  • The difference between mediocre and magical companies is often the difference between letting people take creative risk and holding them too tightly accountable.

The End of the Beginning

The founding of Andreessen Horowitz grew directly from Horowitz’s experience of being told as a founding CEO that he needed to be replaced by a ‘real’ CEO—a systemic failure of the venture capital industry that the firm was designed to correct by combining experienced operator-partners with a systematized network modeled on Michael Ovitz’s Creative Artists Agency, betting that the now-networked entrepreneurial community would spread word of a genuinely differentiated offering.

• The premise of Andreessen Horowitz was that technical founders are the best people to run technology companies—as demonstrated by every long-lasting technology company Horowitz admired—but that the VC industry was better designed to replace founders than help them succeed, creating an opportunity for a firm explicitly built to do the opposite.
  • Benchmark partner David Beirne asked Horowitz in front of his cofounders: ‘When are you going to get a real CEO?’—a moment that tormented him for months and ultimately inspired the firm.
  • All of the long-lasting technology companies they admired—HP, Intel, Amazon, Apple, Google, Facebook—had been run by their founders, specifically the innovator running the company.
• Andreessen Horowitz copied Michael Ovitz’s Creative Artists Agency operating model—building an integrated firm-wide network across large companies, executives, engineers, press, and investors rather than a collection of individual partner networks—on the theory that a networked entrepreneurial community would now spread word of a genuinely better offering.
  • Ovitz’s breakthrough at CAA was that any agent could connect any client to a firm-wide grid rather than relying on individual agent relationships; the result was that within 15 years CAA represented 90 percent of top Hollywood talent.
  • When Marc discovered that traditional VC firms did no PR because the original investment banks funded wars and sometimes both sides, the founders decided to counterprogram by launching Andreessen Horowitz with great fanfare.
• The final lesson Horowitz offers is to embrace the unusual elements of one’s own background—the Communist grandfather, the football team, the diverse social circles—because those distinctive perspectives are the only source of genuinely novel approaches; the keys to dealing with hard things, if they exist at all, are found in one’s own particular formation.
  • His grandfather Phil Horowitz’s tombstone bears his favorite Karl Marx quote: ‘Life is struggle’—which Horowitz identifies as the most important lesson in entrepreneurship: embrace the struggle.
  • As a venture capitalist freed from the unrelenting public confidence required of a CEO, he can now share fears and weaknesses—and it is those fears and controversial opinions that hold the clues to dealing with hard things.