The Hard Thing About Hard Things

Lead bullets mean doing the hard, direct work instead of seeking clever workarounds. At Netscape, Microsoft released a web server that was five times faster and free. Rather than pursue partnerships to outflank them — silver bullets — engineer Bill Turpin insisted they fix the actual performance gap. They did, and the server line grew into a $400 million business.

Years later at Opsware, competitor BladeLogic was winning key deals. Employees pitched workarounds: go down-market, acquire a simpler product, target different customers. Horowitz rejected them all. Customers were buying — just not Opsware's product. After nine months of grueling development called the Darwin Project, they regained their lead. The lesson: when you're losing an existential battle, there is no side door. You go through the front and deal with the big, ugly problem blocking it.

Dignity preserves culture. Legendary VC Doug Leone told Horowitz he'd never seen a company survive three consecutive layoffs and reach a billion-dollar outcome. The secret was doing it right. Six principles guided each round:
1. Don't delay — leaked rumors poison everything
2. Frame it as company failure, not individual failure
3. Train managers to deliver news personally — never outsource the conversation
4. The CEO addresses the entire company first
5. Stay visible afterward — talk to people, help them carry boxes to their cars
6. Provide generous severance details upfront

Bill Campbell's crucial insight shaped the approach: the layoff speech is really for the people who remain. If survivors see friends treated poorly, they'll never trust the CEO again. Opsware's culture held through all three rounds because departing employees felt respected.

Management debt — Horowitz's coined analogy to technical debt — describes expedient short-term management decisions that create expensive long-term consequences. Three common forms:
1. "Two in the box" — giving two people the same leadership role to avoid choosing, which destroys accountability and slows every downstream decision
2. Matching a competing offer to retain someone — which secretly teaches every employee that threatening to quit is the fastest route to a raise
3. Skipping performance reviews — which means weaknesses go unaddressed and mediocrity becomes the silent default

The danger is invisibility. Unlike technical debt that eventually crashes a system, management debt rots culture quietly. You may think the counteroffer was confidential, but the recipient told a friend who told engineering, and now everyone knows the squeaky wheel gets the grease.

Andy Grove's math is irrefutable. If a manager spends twelve hours preparing training for ten people who will work twenty thousand hours next year, even a 1% improvement yields two hundred hours of gained productivity. Yet most startups invest heavily in recruiting and ignore everything after the hire.

Horowitz discovered this at Netscape when he wrote "Good Product Manager/Bad Product Manager" — a simple expectations document that instantly transformed underperformers into effective contributors. At Loudcloud, he made training mandatory with two hard rules: managers couldn't request new hires without a training plan, and the CEO personally taught management expectations. Without training, there's no basis for performance management — you can't hold people to standards they were never taught. Two main reasons employees quit: they hated their manager's lack of guidance, or they stopped learning.

Mark Cranney broke every pattern. When Opsware needed its fourth head of sales in three years, Horowitz ran sales himself first to learn exactly which strengths mattered. Cranney didn't fit any stereotype: average height, square build, Southern Utah University graduate. Every board member and executive voted no. He made people uncomfortable.

But his strengths were unmatchable. He produced a training manual thicker than a textbook. He'd already mapped competitors' active deals. He gave seventy-five references — every one returned calls within an hour. Horowitz's argument to Marc Andreessen: someone this talented wouldn't join a company trading at $0.35 unless he had flaws. Under Cranney, sales grew tenfold and market cap twentyfold. Consensus hiring swayed by group dynamics almost always selects for polished mediocrity over transformational capability.

Peacetime means large market advantage and growth — the CEO empowers creativity, expands the market, encourages experimentation. Wartime means existential threat — the CEO demands strict alignment, violates protocol when necessary, and obsesses over execution details. Google under Eric Schmidt was peacetime; Google under Larry Page declared war. Apple nearly bankrupt under returning Steve Jobs was pure wartime.

Horowitz calculated he spent three days as a peacetime CEO and eight years at war. Almost no CEOs can function effectively in both modes. Most management books only describe peacetime techniques — broad buy-in, employee empowerment, big hairy goals. Wartime looks radically different: zero tolerance for disagreement on the core mission, detailed personal involvement in execution, and training so employees survive the battle. Knowing which mode you're in determines which playbook will save or sink you.

Football coach Bill Parcells once worried that injuries would ruin his New York Giants season and called his mentor, Raiders owner Al Davis, for sympathy. Davis replied: "Nobody cares, just coach your team." Horowitz found this applied to every CEO crisis — when Loudcloud's customers went bankrupt, when 221 IPOs in 2000 collapsed to 19 in 2001, when his largest client stiffed him $25 million.

The mental trap is elaborating misery. A CEO can burn enormous energy proving circumstances were impossible, but none of it preserves a single dollar or a single job. The energy spent justifying failure would be infinitely better spent finding one seemingly impossible way out. There is always a move — even taking a company public with $2 million in trailing revenue and six weeks of cash. Horowitz made that move. In 2001.

Courage is a practice, not a trait. Horowitz watched a founder agonize over rejecting a rich acquisition offer when his board and team all favored selling. The data was ambiguous — maybe 54% keep the company, 46% sell. The founder chose to stay independent. The moment he decided, the entire team rallied behind him. It turned out their preference for selling was largely driven by reading the founder's own ambivalence — they were following his uncertainty, not leading.

Social pressure creates a dangerous trap. If you decide with the crowd and succeed, everyone remembers your wisdom. If you decide against the crowd and fail, you're finished. This asymmetry pushes CEOs toward safe consensus choices. But going along often transforms a 70-30 decision into a perceived coin flip. The only antidote: practice making the lonely, correct decision until it becomes habit.