Sell or hold

Zuckerberg, 2006. Yahoo offers $1B for Facebook. He's 22. He says no. What he was holding: News Feed and open registration, built and unlaunched, plus the engagement data behind them. Facts the price couldn't see, sitting in the codebase, weeks from shipping. Both launched that September, and users ran from roughly 9M toward 50M within the year. The refusal was still expensive: it cost him most of his management team, who thought he was insane. That's the quieter lesson inside the famous one: even a right refusal bills the people currency. Budget for the social cost of no.

Yang, 2008. Microsoft offers $44.6B for Yahoo, a 62% premium. Yang holds out for $37 a share. What he was holding: a feeling about a company with his own name on it. No standalone plan existed under his counter-number, and Yahoo adopted a severance plan so expensive to acquirers that it was widely read as a poison pill built to repel that specific buyer. Microsoft walked. The stock hit single digits within the year. Core Yahoo eventually sold for $4.83B, nine years and one apology tour later.

Same word. No. One was conviction: private, specific, checkable evidence the market hadn't priced. The other was hubris: a feeling with a name on it, defended with structural sabotage. The test, before any refusal: name what you know that the buyer can't price. Then classify it, fact or feeling. If you can't name it, the offer is probably fair, and your urge to refuse is a feeling wearing a thesis.

There's a third instructive case. Andrew Mason turned down Google's roughly $6B for Groupon while merchant churn ran high, clones multiplied, and the cohort tables visibly decayed: the private evidence argued for selling, and the refusal happened anyway. Why? Look at who was voting. The loudest hold votes had already taken hundreds of millions off the table in pre-IPO secondaries. When the people arguing hold have personally de-risked, the hold isn't conviction. It's someone else's risk appetite wearing your equity.

1. Number first. Set the offer against your written enough. If it clears the number and you can't say what more money would be for, that's arithmetic, and the only live question left is structure: cash at close versus paper and promises. Price stops being the conversation.

2. The private evidence test. What do you know that the buyer can't price? In writing, classified. Fact: it's in the CRM, the product is built, the cohort curves bend upward and you can print them. Feeling: it's my name, we're special, next year is our year. Facts can justify holding. Feelings get priced in the identity currency, honestly and explicitly, and they never get to wear a valuation. Sometimes the identity price is real and worth paying; it still has to be paid at the identity counter, not laundered through a fake revenue projection.

3. The insider check. Has anyone arguing hold already de-risked personally? Secondaries taken, loans against stock, a fund whose markup depends on your refusal. Their vote is incentive information, not judgment information. Discount accordingly, and yes, that includes your own board.

4. Walling's window. Rob Walling's observation, and the most useful sell-timing heuristic in the record: the best time to sell is good growth plus fading founder energy, or good growth plus a fading market window. If growth is strong while your energy quietly goes, that's the top of your personal market. Waiting until both fade means selling the fade, at fade prices. Say which quadrant you're in, out loud. Energy counts as data here, exactly like revenue.

5. The hold-side structure check. Tringas's dodge: finance the company so selling is never forced, and sell-vs-hold converts permanently from crisis to preference. If instead the cap table or the burn is forcing this sale, notice what that means: the real call was made years ago by your capital structure, and what remains today is negotiating well. That's a different, smaller job, and pretending it's still a hold-vs-sell decision just makes you negotiate worse.

6. The personal currencies, honestly. Three checks that decide more deals than the price does:

• *Cash flow versus lump sum.* A company paying you $500k a year is an asset the after-tax lump sum has to beat. Run the arithmetic; it often doesn't.
• *The earn-out reality paragraph.* Write one paragraph about a Tuesday in month 14: reporting to a VP of integration, your product decisions going to committee, your calendar owned by someone else's OKRs. Who are you in that paragraph? If the honest answer is "gone by month 10," price the deal at what survives the clawback, and negotiate cash at close accordingly.
• *The ruin check, inverted.* If the company is most of your net worth and the offer clears your enough, holding is the riskier position, whatever the growth chart says. Concentration that made you rich on the way up is just exposure at the top.

• Reserve prices born in the shower. A number with no plan under it is Yang's number.
• "It'll be worth more next year." That's a hold thesis, and hold theses need hold evidence: cohorts, pipeline, something checkable. The same standard as everything else.
• Structural moves that repel the buyer. Poison pills read as fear, and they salt the ground for every future buyer too.
• Letting the banker set your reserve. Their fee curve isn't your life.
• Selling into burnout you haven't named. Run the energy check before the process starts, or the process will run it for you, at the worst possible moment.

Dev-tools founder, $2.1M ARR growing 45%. Strategic offer: $11M, structured $6M cash at close, $5M over a 24-month earn-out. His enough, written 18 months earlier: $4M liquid. Close cash nets about $4.4M after tax: the number clears on the cash alone. Private evidence: enterprise pipeline, real, in the CRM, worth maybe 30% more in a year. A fact, but a modest one. Insiders: clean. Window: growth good, energy honestly fading, two funded competitors shipped last quarter. Dead center of Walling's window. The earn-out paragraph comes back "I'm gone by month 10," and he believes it, so the earn-out prices near zero and the real deal is $4.4M plus option value. Still clears. He counters once, trades earn-out for cash at close, gets most of it, signs.

First line of his memo: *sold inside my window, on my number, written before anyone was buying.* Eight months later a competitor raises a mega-round and the tech press does what it does. The memo is why that headline reads as weather instead of regret.