You do not buy a share of a company. You buy a place in a line — and the line is redrawn every time the company raises again.
A venture investor believes he is buying a percentage of a company: a share of its economics, proportional to the number on the cap table. He is buying something stranger and less stable than that. He is buying a rank in a distribution that does not yet exist, against claimants who have not yet arrived, on terms that rounds not yet raised will rewrite. The percentage is the visible face of the asset. The asset is the rank — his place in the line to be paid — and the rank, unlike the percentage, can be pushed back without the number on the cap table ever changing.
A position erodes in two ways, and an investor watches only one of them. The first is dilution, and it is famous. The stake an investor holds at entry is not a holding but a high-water mark; every financing issues new shares and every option pool issues more, so twenty percent at the Series A is nearer twelve after the B and six by the time anyone is paid, unless its holder spends fresh capital each round simply to stand still. The only honest way to read a pre-exit ownership figure is to append two words to it: for now. This much every investor knows, and tracks, and models. It is the erosion that shows up on the cap table.
The second erosion does not show up on the cap table, which is exactly why it is more dangerous. An investor's place in the distribution is not fixed at the moment he takes it. Later money frequently comes in senior to the money already there. Structured rounds add participation rights and preference multiples. Recapitalizations and pay-to-play provisions can reorder the stack outright. All of it grows more common precisely when a company's options are weakest and its existing investors have the least leverage to resist — when the only alternative on the table is a company that does not get funded at all. Each of these pushes an early investor further back in the line while leaving his percentage untouched. He still owns what he owned; he is simply now standing behind capital that arrived after him and insisted on being paid first. Dilution takes a share of the economics. Subordination takes the place in line. The first is visible and discussed endlessly; the second is invisible and discussed almost never, and it is frequently the one that decides whether an investor is paid.
Both erosions matter for the same underlying reason: the proceeds of an exit are not divided by percentage. They are divided by a waterfall — a fixed order of payment in which preferences are satisfied first, by seniority, before a dollar reaches anyone junior. At a modest or middling outcome, which is where the great majority of venture exits actually live, that order is the entire story. The capital at the front of the line is paid in full; the capital at the back is paid what remains, which is often little and sometimes nothing. A holder's percentage tells him what he would receive if the proceeds were split pro rata. They are almost never split pro rata. What he actually receives is set by where he stands when the money is divided — by his rank, which is the asset he was really buying all along, and the one he was least watching.
It is fair to object that none of this is secret; any competent investor knows that preferences exist and that later money can come in senior. True, and it is precisely the point. The mechanics are known. The misallocation of attention is the illusion. The market negotiates hardest over the percentage — the figure on the headline, in the fund's marketing, in the question asked at dinner — and treats the terms that set the rank as boilerplate to be cleared by counsel. It haggles over the proxy and rubber-stamps the substance — fighting for the number it can see and ignoring the one that decides what it is paid. That is not a failure of knowledge. It is a failure of where the knowledge is pointed.
There is a stronger objection: in the largest outcomes, preferences convert, the stack collapses into common, and rank stops mattering — everyone is paid by percentage after all. This is true, and it is why the illusion endures, because it is harmless exactly where attention is naturally fixed, on the home runs. But the home runs are not where most of a portfolio resolves. Across the broad middle of outcomes — the modest acquisition, the soft landing, the respectable-but-unspectacular sale — rank governs and percentage misleads, and the investor who read his ownership as his proceeds is right about the few outcomes he did not need to worry about and wrong about nearly all the rest. None of this makes ownership irrelevant. Ownership determines the size of the claim. Rank determines whether the claim gets paid.
So the cap-table percentage is not ownership, and it is not even, reliably, economics. It is a bet — that the stake holds, that no one cuts ahead, that the exit lands large enough for rank to stop mattering. Not one of those is a fact, and the cap table quotes all three as though they already were. The percentage is a claim on the company's story; what an investor is actually paid is decided by his rank in the line, and the line is redrawn, behind him and largely without his leave, every time the company raises again. Venture capital prices a percentage and discovers, later and sometimes too late, that what it bought was a position — and that the position had been quietly moving the whole time.
The correction is not a trick for being paid more. It is a shift in what an investor guards. It means treating the rank as the asset it is: negotiating seniority rather than assuming it, holding the rights — pro-rata, anti-dilution, the standing to participate in later rounds and resist subordination — that let a position be defended rather than eroded, and reading the percentage as the proxy it is rather than the thing itself. And it is necessary without being sufficient, as these things always are, because rank pays only when there is cash to divide, which returns the question to the judgment no structure can supply. The investors who compound are not the ones who won the largest number at the table. They are the ones who understood what they had actually bought, and spent the intervening years making sure no one quietly moved them to the back of the line.
The percentage is what an investor negotiates for. The rank is what he is paid on. The costliest error in venture is to guard the first while the second slips — and to notice only at the exit, when the line has already formed, and his place in it is whatever the rounds he waved through decided it would be.
The costliest error in venture is to guard the first while the second slips — and to notice only at the exit, when the line has already formed, and his place in it is whatever the rounds he waved through decided it would be.
—QED—
Sources & Methods. As in earlier letters, the argument here is structural rather than empirical; the percentages, round figures, and outcomes in the text are illustrative, chosen to expose the mechanics at a representative scale rather than to report measured averages. The empirical record the analysis is set against:
■ Dilution and ownership erosion by stage — Carta and PitchBook cap-table data on how founder and investor ownership moves from formation through successive financings.
■ Seniority, participation, and the prevalence of preference terms — the Fenwick & West Silicon Valley Venture Capital Survey and Aumni venture-financing data on liquidation-preference structures, stacking, and pay-to-play provisions across rounds and market conditions.
■ The distribution of exit outcomes — PitchBook and Correlation Ventures data on the size and frequency of venture exits, including the predominance of modest outcomes relative to the outliers.