Ben Horowitz on building through chaos with no recipe

Executive overview

Most business advice offers formulas for problems that have no formulas. Ben Horowitz's account of building LoudCloud and Opsware is a rare honest look at the mental and operational reality of founding companies.

The core challenge: how do you keep making decisions when the ground keeps shifting, the money runs out, and there is no playbook?

Determination, not intelligence, is the deciding factor in whether a founder survives.

The relationship between Mark Andreessen and Ben Horowitz

• Worked together across three companies over 18 years
• Their dynamic: each upsets the other almost every day by finding flaws in their thinking
• Most co-founder relationships go one of two ways: too tense to tolerate, or too comfortable to be useful
• Theirs stayed productively uncomfortable

Building LoudCloud and the dot-com collapse

• LoudCloud idea came from a real operational pain: partners' sites crashed when connected to AOL's e-commerce platform
• Raised at a $700M valuation in nine months; months later couldn't raise at all
• Had to IPO to survive — the only viable path to capital in that market
• Abandoning frugality when capital feels free is one of the most dangerous traps for founders
• Went from nearly 300 employees with almost no cash to laying off staff and selling the cloud business to EDS

The pivot from LoudCloud to Opsware

• When facing bankruptcy, Horowitz asked a different question: "What would I do if we went bankrupt?"
• Answer: buy the Opsware software out of bankruptcy and build a pure software company
• That reframe led to a viable path without actually going bankrupt
• To understand the market well enough to build the right product, they had to ship the wrong product first
• Saving the EDS contract: discovered the client's real want (Tangram software), acquired Tangram for $10M, preserved a $20M/year contract

Mental health and the emotional reality of founding

• Horowitz describes the period of near-collapse as one of sustained dread: not sleeping, cold sweats, crying
• Mark Andreessen's description of startup emotions: "euphoria and terror — and lack of sleep enhances both"
• The most useful reframe: spend zero time on what you could have done; spend all of it on what you might do now
• Asking "what's the worst that can happen?" never helped; asking "what would I do if that happened?" did

Lessons on leadership and hiring

• Look for a market of one when raising money privately: you only need one yes; ignore the thirty nos
• This applies to product too — focus on users who value what you built, not on those who don't
• Hire for strength, not absence of weakness: the most important thing is what someone does well, not what they don't do wrong
• Determination outlasts intelligence: a brilliant but unreliable person is less valuable than a determined, ordinary one
• Training is one of the highest-leverage activities a manager can do (Andy Grove: 12 hours of training can produce 200 hours of output improvement across a team)

Founder versus executive mindset

• Big company executives optimise and tune an existing business; most of their work is reactive
• Founders must generate all motion from scratch — nothing happens unless they make it happen
• In a large company, more than three new initiatives per quarter is too much; for a founder, that constraint doesn't apply

On titles, culture, and decision-making

• Two valid but opposite approaches to titles: Andreessen says give the highest titles freely (they're free); Zuckerberg says keep titles below industry standard to enforce internal calibration
• Both work — the lesson is that successful people can reach opposite conclusions that are both correct for their situation
• Culture can be shaped through shock: Bezos built frugality into Amazon by making everyone work at door-desks from Home Depot
• Your unique background is not a liability; it is where your edge comes from

On when not to sell

• If you are early in a large market and have a credible path to being number one, no buyer can afford to pay what you are actually worth
• Google's early acquisition offers looked rich; they were still wrong because no buyer owned a market as large as the one Google was building

Founding Andreessen Horowitz

• After selling Opsware to HP for $1.65B, Horowitz spent a year inside HP and concluded he didn't want to operate again
• The venture firm's founding thesis: founders should advise founders; experience in building companies is a prerequisite, not a credential
• Strategy copied directly from Michael Ovitz's CAA model: build a firm so good it attracts all the top talent, then shift economics toward that talent
• The entrepreneurial community that now exists — founders talking to founders — made word-of-mouth firm-building viable in a way it wasn't in the 1990s