When you raise your first round, the cap table feels abstract. Percentages in a spreadsheet. You do the math, decide the dilution is worth it for the capital and the partner, and move on to building.
By the time you've raised three rounds, the cap table is no longer abstract. It's a structure of claims on the company's value that will determine who gets what in every possible exit scenario — and that structure was shaped by dozens of decisions made in the earliest rounds, when the decisions felt small.
The decisions weren't small. They compound.
After raising $40M at Realm across seed through Series B, here's what I understand about dilution that I didn't understand at the beginning.
The dilution you model is not the dilution you get
Every founder runs a dilution model before a round. It goes like this: we're raising $X at a $Y valuation, which means selling Z% of the company. Simple math.
What this model doesn't capture:
The option pool refresh. VCs typically require that the company's option pool be topped up before the round closes — which means the refreshed options come from the existing shareholders' equity, not from the new investors' equity. If you're raising at a $20M pre-money with a 20% option pool refresh required, the "effective" pre-money is closer to $16M. This is called the "pre-money option pool shuffle," and it's standard but confusing in every first round.
Pro-rata rights. Early investors often negotiate the right to maintain their percentage ownership in future rounds. When you're raising your Series A, your seed investors might exercise pro-rata, which means the Series A is effectively smaller than it appears — some of the round is going to maintaining prior investors' percentages. This is fine and often useful (early investors with pro-rata are signaling continued conviction) but it affects the total size of the fresh capital entering the company.
Pay-to-play provisions. In down rounds or bridge rounds, some investors will have provisions requiring participation to avoid conversion to common — meaning non-participating preferred converts to a weaker equity class. These are uncommon in good markets and common in hard ones, which means you rarely have to understand them until you're already in a hard situation.
The option pool shuffle hit us hardest in the Series A. We had modeled our dilution cleanly — we knew what percentage we were selling at the negotiated valuation — but the option pool refresh requirement that came in the term sheet added a meaningful gap between the dilution we had planned for and the dilution we actually took. It wasn't hidden; it was in the term sheet in plain language. We had just not run the math on the effective pre-money. It was one of those things where the structure makes sense once you understand it, but the standard model every founder uses doesn't capture it. We caught it before closing and negotiated the pool size down from what was initially requested — but only because someone on our legal team flagged it explicitly. Without that flag, we would have signed it and done the math later and been frustrated without knowing exactly why.
The preference stack is the other cap table
The percentage ownership conversation is about upside. The liquidation preference stack is about who gets paid first, and it matters more in the scenarios that are most likely.
Here's the structure: every preferred investor gets a return of their investment before common shareholders see anything. Most deals are 1x non-participating — the investor gets either 1x their money back or their equity stake, whichever is higher. Standard, reasonable, fine.
The variation that changes outcomes dramatically is participating preferred: the investor gets their 1x back AND participates pro-rata in the remaining proceeds. This means in moderate exit scenarios — the $50M acquisition of a company that raised $20M — participating preferred shareholders capture a disproportionate fraction of the exit value.
At Realm, we were careful about preference stack structure. But I've seen founder friends who weren't — who took participating preferred in an early round when the market was hard, didn't fully model the implications, and then found in the acquisition process that the preference stack was extracting most of the value in the middle range of exit scenarios.
The simple test: model three exit scenarios — a modest acquisition at 2-3x the last round's post-money, a successful acquisition at 5-8x, and an IPO-level outcome at 15x+. Run the preference stack through each. In the modest scenario, how much common gets? In the successful scenario? This is where the structure actually shows up in outcomes, and it's almost never modeled clearly in the moment a term sheet is signed.
Investor rights compound differently than equity
The percentage-based dilution model misses a second form of compounding: the governance rights and protective provisions that accumulate across rounds.
Every preferred investor typically gets:
- Board seats (or board observer rights)
- Protective provisions requiring preferred approval for major decisions
- Information rights (financials, board materials)
- Sometimes drag-along rights in acquisition scenarios
In your seed round, one new board member with a protective provision list feels manageable. By Series B, you have multiple preferred investors, each with their own rights, each with their own list of decisions requiring their consent. The governance overhead is real, and the constraint on founder decision-making compounds with each round.
This doesn't mean you shouldn't raise — the capital matters and the right investors add more than they constrain. But the "investors are partners" framing should be understood concretely: they have specific rights over specific decisions, and those rights don't go away.
The practical implication: the governance structure you set up in your seed round is approximately the governance structure you'll have in your Series B, scaled by the number of investors. Treat early governance conversations as seriously as valuation conversations.
The right mental model for dilution
After working through three rounds and an acquisition, the mental model I use for dilution is this:
You're not selling a percentage of your company. You're selling a percentage of a specific set of outcomes — and the outcomes you're selling percentages of are the ones where the company creates meaningful value.
The right question is not "how much are we diluting?" It's "what outcomes does this round make possible, what is the probability of each outcome, and what does my ownership stake represent in the expected value calculation?"
A 20% dilution that takes a company from a 5% chance of a successful exit to a 30% chance of a successful exit is a very different transaction from a 20% dilution that extends runway by six months on a trajectory that was already weak.
Founders who feel bad about their dilution are usually the ones who took it without fully modeling what the capital was supposed to accomplish. Founders who are happy with the dilution they took are usually the ones who raised at the right moment, with the right investors, and can point to specific outcomes the capital made possible.
The goal isn't to minimize dilution. It's to make every dilution decision in the context of a clear model of what you're buying with it.