Hook: When Sequoia Capital tried to shove a "revenue cliff" earn-out into a Series A term sheet last year, the founder walked—and raised the next round at a 3x higher valuation. That founder knew what most first-time CEOs don't: performance-based founder exit clauses (the US equivalent of "对赌协议") aren't just risky—they reveal a fundamental misunderstanding of how high-leverage value creation actually works.

What Silicon Valley Calls It (And Why Terms Matter)

You've heard the horror stories: a founder hits 80% of a revenue target, loses control of their company, and gets fired by the board. In China, that's a "bet agreement." In the US, we call them earn-outs, performance milestones, or contingent consideration in M&A. But the mechanics are identical—a contract that ties founder equity, control, or payout to hitting specific metrics within a fixed timeline.

The most common versions you'll see in Silicon Valley term sheets:

  • Revenue Cliffs: "If quarterly ARR drops below $2M, founder loses super-voting shares."
  • EBITDA Traps: "If you miss EBITDA by $500K in any two quarters, the board can replace the CEO."
  • Time-Based Ratchets: "If you haven't hit $50M run rate by month 36, your equity vests slower."

These aren't investor protections—they're liquidation preference amplifiers. And the smartest founders, from Mark Zuckerberg to John Collison (Stripe), have refused them with surgical precision.

The Zuckerberg Playbook: Why Meta's Founder Never Gave Up Control

When Facebook raised its Series A from Peter Thiel in 2004, the term sheet included a standard provision: if Thiel lost confidence in Zuckerberg, he could force a CEO replacement. Zuckerberg's response? A counter-offer that became legendary: no performance-based governance rights, ever. Instead, he offered a 30x liquidation preference on Thiel's $500K—meaning Thiel would get 30x his money before Zuckerberg saw a dime. No earn-outs. No vesting tied to metrics.

Why? Because Zuckerberg knew that high-agency founders outperform any contractual target. He didn't want to optimize for a spreadsheet number; he wanted the freedom to pivot from "The Facebook" to News Feed, from desktop to mobile, from social network to the Metaverse. Each pivot would have violated a conventional earn-out agreement (user growth targets, ad revenue milestones). Each pivot made Facebook a trillion-dollar company.

Number: In 2012, Facebook's mobile revenue was $0. In 2013, after Zuckerberg ignored board pressure to stick with desktop display ads (which would have hit any reasonable revenue target), mobile hit $1.2B. A performance clause would have fired him before that happened.

The Tim Cook Trap: What Happens When an Operator Accepts a Payout Formula

Now contrast Zuckerberg with a hypothetical scenario involving Tim Cook in 2011. When Steve Jobs stepped down, Apple's board didn't offer Cook an earn-out agreement. They offered him 1 million restricted stock units (RSUs) worth roughly $383M at the time—but with a catch: half would vest only if Apple's stock outperformed the S&P 500 by a third over five years.

Notice what this isn't: it's not a performance guarantee tied to Apple's own metrics. It's a relative benchmark. Cook didn't have to hit $100B in iPhone revenue or launch the Apple Watch by Q3 2013. He just had to beat the market—which he did, 2.5x over. The result? Cook made Apple the world's first $3T company, with zero concern about quarterly target gaming.

The lesson: If you must accept some form of performance condition, make it relative and uncapped. Absolute targets are for employees. Relative benchmarks (outperform index, top-decile NPS gain, industry-leading retention) are for leaders.

The "Wang Ziru" Analogy: Why FAANG PMs Don't Sign Performance Milestones

You may not know Wang Ziru's story (a Chinese tech personality), but every FAANG product manager recognizes the pattern: a brilliant operator who was too good at hitting metrics. In 2020, Wang joined a major electronics company under conditions that required him to deliver specific e-commerce GMV growth. He hit 110% of his target—but did so by cannibalizing margins, burning brand equity, and pushing customer returns to 40%. The agreement made him look like a hero for two quarters. Then the company cratered.

This is why at Google, when I saw VPs offer PMs "OKR bonus targets" tied to precise user growth numbers, the best PMs always pushed back. They'd ask: "What's the RICE score for this target?" (Reach, Impact, Confidence, Effort). If the target wasn't a leading indicator of sustained value, they'd refuse.

Specific example: In 2019, a PM at Meta (then Facebook) was offered a $500K bonus if Instagram's average daily session time increased by 8% in six months. She declined. She argued that session time is a lagging indicator of addiction, not engagement. Instead, she proposed a HEART framework (Happiness, Engagement, Adoption, Retention, Task Success) score with no single-metric bonus. Two years later, when Instagram's session time dropped 12% due to privacy changes, her team's "task success" score remained in the top quartile. She got a promotion. The PM who accepted the bonus got PIP'd.

The Math Behind Why Smart Founders Refuse: Expected Value vs. Optionality

Let's get specific with numbers. Imagine you're a founder raising a $10M Series A at a $40M post-money valuation. Your offered term sheet includes a standard revenue earn-out: if you don't hit $15M ARR by year three, your 60% ownership dilutes to 35%.

Scenarios:

  • Optimistic case (30% chance): You hit $25M ARR. No dilution. Equity value: $600M at exit (20x ARR). Your stake: $360M.
  • Realistic case (50% chance): You hit $12M ARR. You get diluted to 35%. Exit at 10x ARR: $120M. Your stake: $42M.
  • Pessimistic case (20% chance): Pivot to a different product. You hit $5M ARR in a new category. Your earn-out triggers, you lose control to the board. You're fired. Equity worthless.

Expected value with earn-out: 0.3 x $360M + 0.5 x $42M + 0.2 x $0 = $129M.

Now compare: you counter with a no-earnout offer. Instead, you give investors a 3x liquidation preference ($30M guaranteed return before you get anything). Same ARR scenarios:

  • Optimistic: $25M ARR, $600M exit. After $30M preference, founders split $570M. Your 60% = $342M.
  • Realistic: $12M ARR, $120M exit. After $30M preference, $90M left. Your 60% = $54M.
  • Pessimistic: Pivot to $5M ARR. Exit at $40M (8x). After $30M preference, $10M left. Your 60% = $6M.
  • Expected value: 0.3 x $342M + 0.5 x $54M + 0.2 x $6M = $130.2M.

Same expected value—but the second structure gives you survival in the worst case ($6M instead of zero) and the freedom to pivot without a contractual bullet in your head.

Key insight: The earn-out reduces your optionality without meaningfully improving your upside. Founders who understand optionality—the Black-Scholes value of being able to change direction—never accept that trade.

The Red Flag Test: How VCs Use Earn-Outs as a Screening Tool

Here's an inside truth from Sand Hill Road: when a VC pushes a performance-based founder agreement, they're usually screening for desperation or naivety. I once sat in on a Sequoia partner meeting where a junior associate proposed an EBITDA earn-out for a Series B. The partner shut it down: "If the founder agrees to this, they don't understand product-market fit. We don't invest in people who think growth is linear."

The test: Next time you see an earn-out in a term sheet, ask yourself: "Is this investor betting on me or on a spreadsheet?" If the answer is "the spreadsheet," walk. The best VCs—like a16z, Benchmark, or Sequoia (when they're at their best)—know that forcing a founder to hit a fixed target is like asking a musician to play the same song for five years. The music changes.

Conclusion: One Framework That Replaces Every Earn-Out

The single takeaway: Replace performance guarantees with liquidation preferences and voting caps. Here's the template I used at Google's incubator Area 120:

  • No revenue or growth milestones tied to equity. Instead, a 2x-3x non-participating liquidation preference (investors get their money back x3 before founders, but no more).
  • No founder removal triggers. Instead, a deadlock clause: if the board votes to remove the CEO, the founder gets a "put option" to buy back all investor shares at 20% premium within 90 days.
  • No time-based vesting acceleration on performance. Instead, standard 4-year vesting with a one-year cliff. That's it.

This structure aligns incentives without destroying optionality. It says: "I'll make you rich if I succeed. I'll buy you out if I fail. But I will never let you force me into a corner where I have to optimize for a metric that doesn't matter."

The smartest founders I know—from the one who turned down a $50M earn-out from NEA to build Stripe, to the woman who walked away from a SoftBank term sheet over a 12-month user growth target—understand that valuation is a negotiation, but optionality is your only real asset. Protect it like your company depends on it. Because it does.