Exit Waterfalls

What is an exit waterfall?

An exit waterfall, also called a liquidation analysis, answers one question: when a company is sold, who gets what? It takes the proceeds from an exit, a "liquidity event," and allocates them across every shareholder according to the rights each one holds.

You usually build it forward-looking, as a set of scenarios across a range of exit valuations rather than a single number. The best time to build one is while you're raising a round, because that's when you can see what the new money and its dilution do to everyone's proceeds before you sign. How hard the analysis gets depends entirely on how many different rights the shareholders hold. A company with one class of common is trivial. A company five rounds in, with participation, multiples, and anti-dilution scattered across the stack, is not.

The three steps (the third is the fun part)

1. Build the pre-distribution cap table. Lay out every class of shareholder with the specific rights you'll need for the math: preferences, multiples, participation, conversion ratios, seniority.
2. Figure out how much there is to distribute. Before anything reaches shareholders, settle any debt and any unconverted convertibles. What's left is the distributable proceeds.
3. Figure out who gets what. This is the waterfall itself, walking the proceeds down the stack and testing each class as you go. It's the fun part.

Preferred and common

Most venture cap tables come down to two kinds of shareholder, and the gap between them is the whole reason a waterfall is interesting.

Common shareholders are usually the founders, the employees, and the people running the company. Preferred shareholders are usually the investors, and their preferred stock carries two features common doesn't:

• A liquidation preference: the right to get their invested capital back before anyone junior to them sees a dollar.
• A choice: at exit, a preferred holder can take that preference, or convert to common and take a common share of the proceeds instead.

They'll pick whichever pays more, and that decision depends on what every other investor decides to do.

This is the tradeoff at the heart of a venture deal. Investors buy preferred because its rights sit above common, and those rights, the "structure" of the deal, shape risk and reward independently of the valuation. A higher valuation paired with heavier structure can produce the same headline number and a very different outcome at exit. Structure and valuation are two halves of one negotiation.

The keep-or-convert test

For every preferred holder, the waterfall runs the same test: are you better off keeping the preferred and taking the liquidation preference, or converting to common and taking the common distribution? That test, repeated down the stack, is the most important thing an exit waterfall does. Everything else is setup.

It's also what makes the math circular, because each holder's best choice depends on what the others choose, and theirs depends on this one. More on that below. What you want to build is a structure you can run the same way through every round and every set of terms, so a new deal becomes a new input rather than a new model.

Liquidation preferences

A liquidation preference is the amount a preferred holder is entitled to before junior holders are paid. The plain version is 1x, non-participating: invest $1mm, and you have a $1mm preference. You either take that $1mm or convert to common, not both.

A few terms change the shape:

• A multiple. The preference can be 2x, 3x, or higher. A 2x non-participating preference on a $1mm investment is a $2mm preference. The multiple is downside protection, raising the floor on what the investor recovers if the exit is small.
• Participating preferred. Also called full participation, this lets the investor take the preference and share in the common distribution on their as-converted shares. The "double dip." It lifts the investor's return across every exit outcome, and it does so at the expense of common.
• Capped participation. Participation up to a ceiling, usually expressed as a multiple of invested capital. Past the cap, the investor does better converting to common, so the cap is the point where they stop double-dipping and convert. A 2x cap means participation stops once they've taken 2x their money.

More on the terms: What you need to know about liquidation preferences.

Dividends

Some preferred carries a dividend, set as an annual percentage. It can be cumulative, compounding year over year, or non-cumulative, simple. Venture dividends usually aren't paid out in cash along the way. They accrue, and at exit they get added to the liquidation preference.

That makes a dividend another layer of downside protection. A holder with a 1x preference and an 8% cumulative dividend who exits after five years is owed their capital plus five years of accrued dividend, all of it senior to common. When you model a waterfall, the accrued dividend rides along with the preference through the keep-or-convert test.

Conversion ratios

When a preferred holder converts to common, the conversion ratio sets how many common shares they get. The normal ratio is 1:1, one preferred share for one common share.

It isn't always 1:1. Anti-dilution is the usual reason it moves: once it has fired, a preferred share might convert into 1.5, or 3, common shares. A 3x conversion ratio means one preferred share becomes three common, which changes the as-converted share count and every proceeds-per-share number that depends on it. Use the actual ratio, not the assumed one.

How anti-dilution moves the ratio: Anti-Dilution.

The preference stack: senior, junior, pari passu

Preferred holders sit in an order relative to each other, and that order is the preference stack.

Senior means a holder's claim on proceeds is paid ahead of another's. Junior means it's paid behind. Seniority usually tracks the order of investment, so a later round is senior to the rounds before it: last money in, first money out. Pari passu means two classes rank equally and share proceeds at the same level, even though they're different classes.

The stack is what "waterfall" refers to. Proceeds fill the most senior claim first, then the next, on down.

Subseries within a round

A single round often holds multiple subseries of preferred (Series A-1, A-2, A-3) when SAFEs and convertible notes converted at different prices alongside the new money. Each subseries is pari passu with the others, so seniority among them is equal, but their liquidation preferences are sized to the dollars actually invested by each holder rather than to the converted share count times the round price. The keep-or-convert test runs separately for each subseries: an A-2 holder might convert to common while an A-1 holder takes the preference, depending on each one's invested dollars and conversion ratio.

Background on why subseries exist on the post-round cap table: SAFEs and Convertible Notes.

Preferences all the way down

The cleanest way to picture a distribution is proceeds flowing downhill. They enter at the top of the stack, fill the most senior class, and what's left flows to the next class, and the next, down through preferred to common shareholders and option holders at the bottom. At each class you weigh that holder's rights and terms, decide what they take given what everyone else is doing, and pass the remainder down.

This is why the average proceeds per share is a number to distrust. Two shareholders in the same company, at the same exit, can land in very different places, and the terms are exactly what create that spread. Shaping returns beyond the raw exit valuation is the entire point of preferred structure.

Calculating proceeds to options

Options need their own handling at exit.

In a change of control, issued-but-unvested options vest according to their agreements. Watch this one: they usually don't fully accelerate, so don't assume 100% vesting. Unissued options still in the pool are either cancelled or redistributed back to shareholders under a formula that compensates them for the dilution the pool created.

An option is in the money if proceeds per share clears its strike price, and out of the money if it doesn't. Out-of-the-money options won't be exercised, which pulls those shares out of the proceeds-per-share denominator. In-the-money options will be exercised, and the cash the holders pay to exercise (strike price times shares) typically goes to the company and gets added to the proceeds available to distribute.

Both effects loop back on proceeds per share. Exercising changes the share count, and the exercise cash changes the proceeds, so testing every strike against the exit is genuinely circular. Iterative calculations again.

Calculating proceeds to warrants

Warrants are the right to buy shares at a preset price, usually handed to investors, and often to lenders, as a sweetener. They get the same in-the-money test as options, though in practice their strike price is often set low enough to be negligible.

Warrants are usually exercised "cashless": the holder doesn't actually wire the strike price times the share count, they just receive the net difference in value.

Economic rights and control rights

Everything above is an economic right, the kind that shows up as dollars in a waterfall. Deal structure has a second half that doesn't: control rights.

Control rights govern the strategy and the running of the company. Board composition (how seats split between investors, the company, and outsiders, and which decisions need board approval) and protective provisions (veto and blocking rights over specific decisions) are the main ones. They don't change the waterfall math, but they shape whether an exit happens at all and on what terms, so they belong in the picture even though they never enter the arithmetic.

Proceeds per share is the crux

If there's one number to anchor on, it's proceeds per share. Everything else falls out of it: whether a preferred holder keeps or converts, whether an option is worth exercising, what each shareholder walks away with, and the return metrics investors calculate on top of all that. Get proceeds per share right and the rest is bookkeeping.

It's also why the math is circular. Proceeds per share depends on the share count, the share count depends on who converts and which options exercise, and those choices depend on proceeds per share. You can't break the loop cleanly, so don't try. Build a repeatable structure that walks capital down the stack the same way every time, ignore the spreadsheet's circular-reference warnings, turn on iterative calculations, and let it solve.

Returns metrics like MOIC and IRR are built on top of these proceeds: Venture Capital Metrics.

For the full mechanics, see the Cap Table and Exit Waterfall Tool.

Where this lives in your legal docs

Liquidation preferences, participation rights, conversion ratios, and seniority all live in the Certificate of Incorporation, typically in Article IV. The Voting Agreement governs drag-along rights, which can force minority holders into the transaction. The NVCA model documents are the canonical templates.

Adding more shareholder classes

If you're using the Cap Table and Exit Waterfall Tool and need more shareholder classes, you can add them, though for anything elaborate a custom waterfall may serve you better. Here's how to add classes: