BackStartups

Lesson 01

The hardest part is not selling

Three of the investments in this section produced very different outcomes from similar entry positions, and the variable was how long the position was held.

The pattern

Put the entries side by side and a pattern appears that has nothing to do with picking. Peter Thiel entered Facebook at a price nobody could improve on and realised roughly two thousand times his money by selling into the listing. Masayoshi Son entered Alibaba at a comparably good price and held for more than two decades. Garry Tan reached the same decision point at Coinbase's listing and publicly chose to hold.

None of these people is a worse investor than the others. What separates the outcomes is a decision made after the hard part — the identification — was already complete, at the moment when the position had grown large enough to make holding it uncomfortable.

That discomfort is structural rather than emotional. A position that has become an enormous share of your net worth is a real risk-management problem. There is tax to fund. There may be obligations to other people. The rational case for trimming is not weak, and the people who trimmed were not being cowardly.

What they mostly were not doing was following a rule written in advance. The decision arrived at the least favourable possible moment for clear thinking — during a liquidity event, under time pressure, with a number in front of them that looked enormous relative to everything they had previously experienced.

For your own cheque

What an angel takes from this

  • Write the sell rule before you need it. A multiple, a time condition and a proportion: consider selling half at 10× or more in any structured secondary after year six. The specific numbers matter far less than the fact that you chose them while nothing was happening.
  • Prefer partial sales. Selling half removes the possibility of a zero and preserves the entire upside case on the remainder, and it requires no view at all about what happens next. Almost every argument for selling is fully satisfied by selling some.
  • Understand which position you are holding. The one you are tempted to sell is, by construction, the one that has worked — which in a power-law portfolio means it is the one paying for everything else. The temptation and the importance have the same cause.
  • Do not let a run of losses drive the decision. Losses cluster in years three to six; the winner resolves in years eight to twelve. Selling the winner to offset the losses is selling the portfolio to pay for the part of it that was always expected to fail.

Checked, not remembered

Sources

Every figure on this page is traceable to published reporting. Where credible sources disagree we say so above rather than picking the most quotable number, and we do not state what a stake is “worth today” — that moves with the share price and is stale the moment it is written.

The hardest part is not selling: common questions

Is it ever right to sell an angel position early?
Frequently. Concentration risk, tax and personal circumstances are real, and an investor eight years into an illiquid position has usually earned the right to take something off the table. The failure is not selling — it is selling without having decided the rule while you could still think clearly.
What should a written sell rule contain?
A multiple, a time condition and a proportion — for example, consider selling half at 10× or more in any structured secondary after year six. Choosing the numbers matters less than choosing them in advance, because the decision itself will arrive at the worst possible moment for judgement.