Decide what the company should become

Research date: September 15, 2026. All market figures, multiples, and base rates below reflect data available as of that date. Financial markets move; the structural arguments age far better than the numbers. Where a figure is time-sensitive, it is dated inline.

This is not legal, tax, or financial advice. Several sections describe structures — ESOPs, perpetual purpose trusts, Section 1042 rollovers, S-corporation compensation, rollover equity — where the difference between a good outcome and an expensive one is entirely in the execution, and where you genuinely need a qualified professional. Those moments are flagged explicitly.

How this chapter fits with the rest of the library

This chapter is about destinations, not deals.

• Chapter 7 (Funding and Financing) covers acquisition as a financing event (A17), secondaries and founder liquidity (B15), exit planning mechanics (B16), and the raise-versus-independent comparison (B17). If you want to know how a term sheet works or what a Series B costs you in dilution, that is the chapter.
• Chapter 11 (Legal, Operations and Founder Sustainability) covers the operational reality of shutting down (B8), selling (B9), liquidity (B10), returning capital (B11), restarting (B12), and the case for building smaller (B13).
• Chapter 14 (forthcoming) covers transaction mechanics in detail: LOIs, diligence, reps and warranties, escrow and indemnity, earn-out drafting, tax structure, and valuation methodology. When this chapter says "an earn-out typically pays about 20 cents on the dollar," Chapter 14 is where you learn how the clause is actually written.

This chapter sits upstream of all of them. It answers a question the others assume you have already settled: where are you trying to end up, and is that destination still reachable from where you are standing?

The core argument

Most founders do not choose an endgame. They back into one.

The sequence is almost always the same. A founder incorporates because they need a bank account and a way to issue stock to a co-founder. They pick the entity their lawyer's template defaults to, or the one the accelerator requires, or the one a blog post recommended. They take a SAFE from an angel because the angel offered and the money was useful. Two years later they take a priced seed because the SAFE money ran out. Three years after that, with $4M of ARR, flat growth, a board with two investor seats, $18M of liquidation preference sitting above the common stock, and a strategic acquirer offering $22M, they discover that the outcome they would actually have preferred — running the business for cash, or selling it to a mid-sized software holdco for $15M and keeping most of it — stopped being available at some point they cannot precisely identify.

Nothing dramatic happened. No single decision was obviously wrong. But the menu closed, quietly, one line item at a time, and nobody sent a notification.

The thesis of this chapter is simple and, I think, under-argued in startup writing:

The endgame is chosen at incorporation and in the first two financings, not at the exit. Entity type, who is on the cap table, what instrument they hold, what rights they negotiated, and what expectations they were sold determine which destinations remain reachable years later. By the time anyone in the room says the word "exit," most of the menu has already been eliminated.

The corollary is the practical payoff: if you know where you want to end up, a small number of upstream decisions — most of them cheap, most of them reversible only in the first 18 months — keep the door open. Part C of this chapter is that list.

A second thesis, less comfortable:

Most of the menu in this chapter is unavailable to most companies, and the destinations that startup media treats as default are the rarest ones. The IPO is a rounding error. The nine-figure strategic sale is a rounding error. The modal good outcome for a founder who builds something that works is "own a profitable small business and take distributions from it for a long time," and that outcome is almost entirely absent from the discourse because nobody makes money writing about it.

Survivorship bias: read this before you read any number in this chapter

Exit data is the most survivor-biased data in the entire startup field, and it is not close.

M&A databases — PitchBook, Crunchbase, SRS Acquiom, marketplace reports — are built from transactions. A company that never transacts never enters the dataset. A company that quietly dissolved, or is still limping along at $1.2M ARR, generates no deal record and no data point. The population you can measure is defined by the outcome whose probability you are trying to estimate.

Four distortions to hold throughout:

1. Reported multiples are conditional on a sale having happened at all. "Median SaaS acquisition multiple: 3.9x profit" means of the businesses that sold — not of those that listed and failed to sell, or never listed because no buyer would have engaged.
2. Disclosed deals skew large. The sub-$20M strategic sale — the most common venture-backed exit by count — is systematically underrepresented in any dataset built from announcements.
3. Use exit data for pricing, not for probability. The only reliable window into the denominator comes from cap-table platforms, fund return distributions, and stage-graduation data — all in Chapter 7.
4. Earn-out data is the honourable exception, because escrow agents see money promised versus money paid on the same deals. It is ugly. See A5.

Throughout this chapter I label figures as [Verified] (traceable to a primary or near-primary source, cited), [Estimate] (a reasonable inference from cited data, stated as such), or [Analysis] (my own reasoning, which you should argue with). Medians are labeled as medians. Means are labeled as means. Where I only have a mean and the distribution is obviously skewed, I say so, because a mean in a power-law distribution is a number about the top of the distribution wearing a disguise.

The foreclosure clock: what closes, and when

Before the menu, the timing. This table is [Analysis], built from the mechanics described throughout the chapter. It is the argument in one page.

Read the third column again. Three destinations are foreclosed at the first priced round, and a fourth is impaired. That is the whole chapter in one row.

PART A — THE MENU

For each destination: what it is; what it pays and on what timeline; what it costs in control, time, and optionality; who it suits; what has to be true years in advance; the failure modes; and realistic base rates.

A1. Run it indefinitely as an owner-operator cash business

What it is

You own the business, you run it (or run it with a small team), it generates more cash than it consumes, and you take the excess out. There is no liquidity event. The "exit" is that you got paid every year for a long time, and you may or may not sell at the end.

This is the modal good outcome in business generally, and it is the one destination on this menu that requires no counterparty. Everything else on this list requires someone to buy, underwrite, hire, lend, or list. This one requires only that customers keep paying you.

What it pays, and when

The arithmetic is unglamorous and very good. A software business at $2M ARR with a two-person team, 80% gross margin and $700K of operating cost throws off roughly $900K–$1.1M of pre-tax owner earnings a year [Analysis] — consistent with the margin data Acquire.com reports for small SaaS, where average profit margins on listed businesses ran 71% in 2025, up from 67% in 2023 (Acquire.com, Jan 2026) [Verified — margin on listed businesses, a selected sample]. Ten years of that is $9M–$11M of pre-tax cash with no preference stack, no earn-out, no escrow, no acquirer and no board. A services business at $5M revenue with a 15–20% owner-earnings margin yields $750K–$1M, with materially more operational drag per dollar.

The timeline is the point: the money arrives continuously from year two or three, rather than in a single lump a decade out with a high probability of never arriving.

Distributions versus salary, at a high level

The tax structure matters enough to be worth understanding in outline, and enough that you should not act on an outline. This is exactly where you hire a CPA.

• Single-member LLC (default, disregarded entity): all net income flows to your personal return and is subject to self-employment tax on the full amount.
• LLC or corporation taxed as an S-corporation: you must pay yourself "reasonable compensation" as W-2 salary, subject to payroll taxes; profit above that can be distributed without self-employment tax. The savings are real and the IRS scrutiny is real. "Reasonable" is a facts-and-circumstances test based on what you would pay someone else to do your job, and understating it is one of the most commonly examined positions in small-business tax (SDO CPA, S-Corporation Tax Guide) [Verified as a description of the rule; the dollar thresholds are fact-specific].
• The Section 199A qualified business income deduction — a deduction of up to 20% of qualified pass-through income — was made permanent by the 2025 tax legislation, with modified phase-in thresholds (SDO CPA, QBI Deduction 2026) [Verified as of research date]. Specified service businesses (consulting, law, health, and others) face income-based limitations that can eliminate it entirely. This is a material planning difference between a software business and a consulting business at the same revenue.
• C-corporation is generally the wrong default for a cash-distributing business, because distributions are taxed twice. The major exception is Section 1202 QSBS treatment, which requires a C-corp and which matters only if you intend to sell. See Chapter 11 §10 for the QSBS detail including the 2025 changes. This is the central entity trade-off for a founder who is undecided: C-corp optimizes for a sale, pass-through optimizes for distributions, and you cannot fully optimize for both.

What it costs

Control: nothing. This is the only destination where you keep all of it. That is its defining feature.

Time: everything. The business does not run itself, and the cash flow is contingent on your continued attention. You have converted an asset into a job — a very well-paid job that you own, but a job.

Optionality: subtler and more expensive than founders expect. Each year you accumulate key-person dependency, your technology ages, and your market position calcifies — the business becomes less saleable precisely as you become more dependent on its cash flow. Because the money arrives continuously, there is never an obvious moment to reassess. [Analysis] The failure mode here is not losing money; it is spending fifteen years in a business you stopped enjoying in year six because the distributions were too comfortable to leave.

Concentration risk: your income, net worth and professional identity are one undiversified position. A founder taking $900K a year out of a business worth maybe $3.5M on sale should be moving a meaningful share of those distributions into assets that are not the business. Most do not.

What has to be true, years in advance

• No preferred stock with a liquidation preference and no investors who need a fund-returning exit. Angels who understood they were buying a small business are fine. Institutional preferred is not. A fund that owns 20% of your company at a $12M post-money cannot get a fund-relevant return from your dividend policy, and their rights — protective provisions, board seats, sometimes redemption — exist precisely to prevent you choosing this.
• A pass-through-friendly entity, or a plan to convert. Converting a Delaware C-corp to an LLC is a taxable event on appreciated value and is often prohibitive by the time you want it.
• Margins that survive you paying yourself a market salary. If the business is only profitable because you work for free, it is not profitable.
• A market that does not require you to outrun a venture-funded competitor. This is the real constraint, and it is why the destination works far better in unsexy verticals.

Failure modes specific to this destination

1. Undercapitalized growth trap. A funded competitor takes the market while you optimize for distributions. This is the genuine cost of the strategy and it is not always small.
2. Platform dependency. Businesses built on someone else's distribution — an app store, a search engine's rankings, a single API — can go to zero on a policy change. Content businesses learned this brutally through the 2023–2025 search and AI-answer shifts.
3. Complacent pricing, because the current price is working.
4. Owner burnout with no exit valve: you never built toward a sale, so the business is not saleable when you want out.
5. Tax drag from the wrong entity, compounded over a decade.

Base rates

There is no clean statistic for "percentage of startups that become durable owner-operator cash businesses," because nobody collects it — these companies do not raise, do not exit, and do not appear in any startup dataset. [Analysis] What we can observe:

• The US small-business acquisition market gives a shape of the population. BizBuySell's Q1 2026 report covers businesses with median revenue of $713,404 and median cash flow of $165,256, selling at a median price of $350,000 and an average 2.7x cash-flow multiple (BizBuySell Q1 2026 Insight Report) [Verified]. That is the real center of gravity of American small business ownership. A $2M-ARR software business with $900K of owner earnings is, in that population, a genuinely large and desirable asset.
• [Analysis] Among companies that survive five years and never raise institutional capital, "profitable enough to pay the founder well" is plausibly the single most common non-failure outcome — more common than any sale. It is invisible because it is not an event.

Why startup media ignores this

Structurally, not conspiratorially. [Analysis] Venture media reports on the venture asset class. A business that never raises, never exits, and never generates a valuation headline produces no news, and its founder has no incentive to publicize revenue — that invites competitors and acquirers. The accelerators, funds and conference organizers who shape the discourse all have economics requiring some companies to pursue large outcomes. Nobody is lying; the selection effect is simply overwhelming, and the result is that the most likely good outcome is the least discussed one.

A2. Hold the company and hire a CEO — step back to owner

What it is

You stop operating and keep owning. Someone else runs the business; you retain the equity, sit on the board (usually as chair), and take distributions or hold for a later sale. It is the only destination that converts a job you own into an asset you own without selling it.

What it pays, and when

Mechanically: whatever the business threw off before, minus the CEO's fully-loaded cost, minus the performance gap. That last term is the one founders model at zero and that empirically is not zero.

Compensation for a CEO hired into a genuinely small company, as of 2026 [Verified for the ranges; the tiering is the source's framing] (CEO Worldwide, Small Company CEO Salary Guide, 2026):

Add a bonus of 25–40% of base as standard at the lower-mid level, and total compensation including long-term incentive typically running 1.5x–2.5x base for senior roles. Add equity: an experienced outside CEO in a small private company will expect 5–10% of the company, usually as options or profits interests vesting over four years, often with a change-of-control acceleration. [Estimate — this range is the consistent practitioner norm; there is no clean public dataset for private sub-$20M-revenue CEO equity grants.]

So the honest arithmetic on a $2M-ARR business throwing off $1M of owner earnings: a mid-tier CEO at $300K base plus bonus plus payroll costs is roughly $420K–$480K fully loaded. You are handing over 5–10% of the equity. Your distributions fall from ~$1M to ~$520K–$580K, and you have given away a tenth of the asset.

That trade only makes sense if one of three things is true: the business can grow meaningfully faster with a full-time professional operator than with a disengaged founder; your time is worth more than $500K/year elsewhere; or you were going to quit anyway and the alternative is decline.

At $5M+ of revenue and $1.5M+ of owner earnings, the same math gets much better, because the CEO's cost is a smaller fraction of the cash flow. [Analysis] Below roughly $3M of revenue, hiring a CEO usually destroys owner economics. Above roughly $5M, it often creates them.

What it costs

Control: you keep legal control if you keep the shares and the board, but lose operational control immediately and completely. The CEO will change things you would not have changed; countermand them and you have not hired a CEO, you have hired an expensive general manager who will leave.

Time: six to twelve months of overlap is normal, during which you are doing both jobs.

Optionality: mostly preserved and in one respect improved — a company that runs without its founder is dramatically more saleable to PE, search funds and strategics. [Analysis] Removing key-person dependency is the single highest-leverage move available to a small company's valuation, so even when the near-term cash math is marginal, hiring a CEO is often exit preparation disguised as a lifestyle choice.

Why it often fails

The base rates on executive hiring are poor, and there is no reason to think small private companies do better than large public ones.

• Leadership IQ's longitudinal study of 20,000+ new hires across 312 organizations found 46% failed within 18 months and only 19% achieved unequivocal success (Leadership IQ, Executive Failure Rates) [Verified — note this is all new hires, not CEOs specifically, and is a manager-perception measure].
• Aggregated managerial-derailment estimates across twelve studies cluster around a mean of ~47% and median of ~50%, with a wide range (30–67%) reflecting inconsistent definitions (same source) [Verified as reported; treat the precision skeptically — the underlying studies are heterogeneous].
• In the S&P 1500 from 2000–2018, 24.7% of CEO departures were dismissals (1,023 of 4,141) [Verified].

[Analysis] The specific failure modes in a founder-to-hired-CEO transition in a small company:

1. The founder does not actually leave. They keep the Slack, the customer relationships, and the opinions. The new CEO has responsibility without authority, performs poorly, and is blamed for it.
2. The company cannot afford a good enough CEO. The entry-tier hire it can afford is by construction someone who has not done the job, in a company with no bench and no margin for a learning curve.
3. Founder-specific advantages do not transfer. Product taste, key relationships and category credibility often carry more value than process. A professional manager can maintain a machine; they frequently cannot replace judgment.
4. Incentive horizons mismatch. A CEO with four-year vesting and an EBITDA bonus optimizes for four years and EBITDA. If you want a fifteen-year compounding asset, you have misaligned it.
5. Governance vacuum. An owner who has never had a board has no reporting cadence, no scorecard and no early-correction mechanism, so problems surface at eighteen months instead of five.

Who it suits

Owners of businesses above roughly $3–5M of revenue with documented processes, a management layer below the founder, and a founder whose continued involvement is genuinely optional. It suits serial founders who want to start something else while retaining an income stream. It suits owners deliberately preparing for a PE or search-fund sale in two to three years, where founder-independence is directly worth a turn or two of multiple.

It suits almost nobody with an institutional venture cap table, because in that situation the board — not the founder — decides who the CEO is, and the conversation is called something else entirely.

What has to be true, years in advance

Documented operations. A second layer of management. Customer relationships that belong to the company rather than to you. Financial records good enough that an outsider can understand the business in a week. Enough gross profit to carry the salary. These are the same preconditions as a sale, which is the useful insight: preparing to hire a CEO and preparing to sell are the same project.

A3. Sell early — pre-Series A, sub-$20M

What it is

Selling the company before it has raised institutional growth capital: a small strategic sale, a product acquisition, a team-plus-technology deal. Purchase prices typically between $1M and $20M, often mostly stock or a mix, frequently with retention packages that exceed the equity consideration.

This is, by count, one of the most common venture-backed outcomes, and one of the least visible — sub-$20M acquisitions are usually not announced with a price, so they barely appear in M&A datasets. [Analysis] PitchBook/NVCA data showing that 75% of 2025 US venture-backed acquisitions occurred at companies that had raised only through Series A is the clearest available signal of where the volume actually is (Q3 2025 PitchBook-NVCA Venture Monitor) [Verified].

What it pays, and when

Timeline: typically 3–6 months from first serious conversation to close, sometimes faster for small deals, plus a retention period of 1–3 years during which a large share of your real compensation vests.

What you net depends almost entirely on two numbers: the preference stack and the split between purchase price and retention compensation.

A worked case. [Analysis — illustrative, arithmetic is exact]

Company raised $1.5M on post-money SAFEs (converted) and a $4M seed at $16M post. Total preference: $5.5M, 1x non-participating. Founders hold 55% of common after a 12% option pool; investors hold ~30% preferred; employees ~15%.

Sale at $18M, all cash. Preferred is non-participating, so investors compare $5.5M preference against 30% of $18M = $5.4M. They take the preference: $5.5M. Remaining $12.5M splits across common: founders get 55/70 × $12.5M ≈ $9.8M, split between two founders ≈ $4.9M each pre-tax. This is a genuinely excellent outcome.

Same cap table, sale at $7M. Investors take $5.5M. Common splits $1.5M. Founders get $1.18M total, **$590K each pre-tax**, after five years. This is the far more common version, and it is why "we sold the company" tells you nothing.

The second variable is usually larger than founders expect. In small acquisitions a substantial fraction of the economics is delivered as retention equity and salary at the acquirer, which does not flow through the cap table at all. That is good for you and bad for your investors and non-retained employees — a structural conflict discussed at length in A6.

Why early sales are often the rational choice

[Analysis] A $15M sale at seed stage with a clean cap table is frequently a better risk-adjusted outcome than the expected value of continuing. You are undiversified, and converting a 55% stake in an uncertain asset into $5M of cash is a portfolio decision, not a failure of ambition. The conditional probability of a much larger outcome is low — each subsequent stage is a filter most companies do not pass (Chapter 7 §A8). The offer in front of you is information about the market's view of your company, usually better information than your own. And acquirer interest is not persistent: strategics buy when they have a gap, a budget and a champion, and all three decay.

Why investors may block them, and how

Here is the structural conflict, stated plainly. A fund with a $150M vehicle needs individual positions capable of returning a meaningful fraction of the fund. A $15M exit where they own 25% returns $3.75M — irrelevant to fund performance, and it also closes a position that had option value. Their rational preference is for you to keep going even when your rational preference is to sell. Neither party is behaving badly. The incentives simply diverge, and they diverge hardest at exactly the deal sizes that are life-changing for a founder and immaterial for a fund.

The mechanisms: protective provisions requiring majority (sometimes supermajority) preferred consent to a sale, near-universal in priced rounds; board approval, which a sale needs before it reaches stockholders; drag-along and voting agreements, usually drafted so the preferred majority is a required consenting group; and soft pressure, more common than formal blocking and much harder to document — the round you need next will be led by people who talk to your current investors.

[Analysis] The practical implication is not "never take venture money." It is: if a sub-$20M exit is a plausible outcome you would be happy with, the time to negotiate for that possibility is at the term sheet, not at the LOI. Concretely, that means caring about protective-provision thresholds, board composition, and whether your investors' fund size makes a small exit tolerable to them. A $30M seed fund can be delighted by a $15M exit. A $600M multi-stage fund cannot.

What has to be true, years in advance

• A cap table where the preference stack is small relative to plausible sale prices.
• Investors whose fund size makes a small exit acceptable, and ideally an explicit conversation about it before you sign.
• Clean IP assignment, clean contractor agreements, clean open-source hygiene. Small deals die in diligence more often than they die on price, because the acquirer's cost of diligence is fixed and their patience for a $6M deal is limited. Chapter 11 §2 and §6 are the checklist.
• A product or team that is a recognizable "gap-filler" for a specific acquirer.

Failure modes

1. Single-bidder process: no price discovery, no leverage. The cure is a real process, which costs money and time and risks the one bidder walking.
2. Talking to a competitor doing competitive research. NDAs do not un-learn things.
3. Retention package swallowing the deal: a lower headline price for a big retention grant, then you leave or are pushed out before it vests.
4. The preference stack eating everything, discovered late. Model the waterfall before deciding to pursue a sale.
5. Team resentment. Employees whose strike price exceeds the effective common price get zero. Negotiate a carve-out early — it comes out of investor proceeds, so it is a negotiation, not a gesture.

Base rates

[Verified] In 2025, US venture-backed exits ran at roughly 1,100–1,500 per year across all types (1,185 in 2023, 1,259 in 2024, 1,135 through Q3 2025) against a population of tens of thousands of active venture-backed companies (Q3 2025 PitchBook-NVCA Venture Monitor). [Analysis] Against roughly 40,000–50,000 active US venture-backed companies, that is an annual exit rate in the low single digits — and most of those exits are small. Early acquisition is common relative to other exits. It is not common relative to companies.

A4. Sell late — growth-stage strategic M&A

What it is

A sale at Series B and beyond, typically to a strategic acquirer (a larger operating company) or to private equity. Purchase prices from roughly $50M into the billions. This is the outcome that startup media treats as the default, and it is the outcome around which the entire venture financing structure is designed.

What strategics actually pay for, and why

Strategic acquirers are not buying a DCF. They are buying one of a small number of specific things, and knowing which one you are determines your price far more than your metrics do. [Analysis], but consistent with what the 2025 deal record shows:

1. A product gap on their roadmap they would otherwise spend eighteen months building. The most common and most price-disciplined category: they pay a multiple of their build cost, adjusted for time.
2. A customer base or distribution channel — your logos, contracts, or access to a segment. Priced on revenue and retention.
3. A team with scarce capability. The 2025–26 AI market is the extreme case; see A6.
4. Removing a competitor. Rare, expensive, and increasingly awkward given antitrust attention on serial acquirers.
5. A defensive or narrative need. These print the absurd multiples, and they depend entirely on the acquirer's circumstances rather than yours.

The 2025 market was genuinely active. SaaS M&A hit a record 2,698 transactions in 2025, up 28% from 2,107 in 2024, with private equity involved in 58% of transactions and strategics in 48% (Software Equity Group, 2026 Annual SaaS Report) [Verified]. But the pricing environment was much less generous than 2021: the SEG public SaaS index fell from 6.3x EV/revenue at Q4 2024 to 4.8x at Q4 2025 (same source) [Verified], and public comparables anchor private multiples.

[Analysis] The practical translation: in 2026, a growth-stage SaaS business with good retention and 30–50% growth is being discussed in the low-to-mid single digits of forward revenue by disciplined buyers, with premiums reserved for scarcity (AI capability, regulatory moat, category leadership) and discounts for concentration, churn, or services-heavy revenue. Chapter 14 covers how those multiples are actually built.

The preference stack determines what you get — worked numerically

This is the section every founder should read twice.

Liquidation preference terms in the current market are, on their face, founder-friendly. In Cooley's Q2 2026 data, 96.4% of deals used non-participating preferred and 95.8% used a 1x preference (Cooley Q2 2026 Venture Financing Report) [Verified]. Participating preferred — the genuinely punitive structure — is now rare at 3.6%.

That does not save you. Non-participating 1x preferred is perfectly capable of taking everything, because the preference stack is cumulative across rounds, and it is denominated in dollars raised, not in ownership. A company that has raised $95M has $95M sitting above the common stock regardless of how fair each individual term sheet was.

The $100M sale where the founder got nothing

[Analysis — illustrative worked example. The waterfall steps below are exact on the stated inputs. The ownership percentages named are the classes that matter to the waterfall; the unlisted remainder sits in later angel and secondary positions rolled into the preferred classes above, and is not separately modeled.]

Company: enterprise SaaS, founded 2019, sold 2026 for $100M all cash.

Financing history:

Round	Raised	Post-money	Preference
Seed	$3M	$12M	$3M, 1x non-part.
Series A	$12M	$50M	$12M, 1x non-part.
Series B	$30M	$160M	$30M, 1x non-part.
Series C	$50M	$400M	$50M, 1x non-part., senior
Total	$95M		$95M

Cap table at sale (as-converted): founders 18%, employees/option pool 14%, Seed 4%, A 8%, B 13%, C 18%, remainder to later angels and secondary buyers rolled into the preferred classes. (Percentages reflect four rounds of dilution plus pool refreshes — see Chapter 7 §B8 for the dilution mechanics.)

The waterfall at $100M:

1. Series C preference is senior (common where a later round is priced in a weak market): $50M off the top. $50M remains.
2. Series B compares its $30M preference to 13% of $100M = $13M. It takes the preference. $20M remains.
3. Series A compares $12M to 8% of $100M = $8M. It takes the preference. $8M remains.
4. Seed compares $3M to 4% of $100M = $4M. Seed would convert — but it now must share only the residual. Seed takes the $3M preference. $5M remains.
5. Common stock — founders and all employees — splits $5M across 32% of the as-converted shares.
6. Founders' 18% of the company receives 18/32 × $5M ≈ $2.8M, split across two founders: ~$1.4M each, pre-tax, before any escrow holdback, before any earn-out, after seven years.

Employees with options split the other ~$2.2M. Anyone whose strike price exceeds the effective per-share common price receives nothing, which at this waterfall is most people hired after Series B.

The press release says "$100M acquisition." Everyone congratulates the founders. The founders made less than they would have made as senior engineers at the acquirer for the same seven years.

Three things make this scenario worse in practice, and all three are common:

• A management carve-out, negotiated to keep the founders from walking, is funded out of the purchase price before the waterfall. This helps the founders and further reduces what employees receive. It is one of the most under-discussed conflicts in M&A.
• Structure in the last round. Even with 1x non-participating being the norm, distressed and late-stage rounds increasingly carry seniority, ratchets, or multiple preferences. Cooley's Q2 2026 data shows recapitalizations at 1.81% and pay-to-play at 8.4% of deals — down from the 2023–24 peak but not zero (Cooley Q2 2026) [Verified]. Chapter 7 §B13 covers structured rounds; the point here is that a structured round is a mortgage on your exit proceeds.
• Stock consideration, escrow, and earn-out mean the $100M is rarely $100M on close. See A5.

The rule of thumb

[Analysis] Take total capital raised, multiply by roughly 1.2–1.5x to account for seniority, accrued dividends, and transaction costs, and treat that as the sale price below which the common stock is approximately worthless. Then ask honestly whether the sale price you consider likely is above or below that number. Founders who do this arithmetic at the point of raising rather than the point of selling make different decisions.

What it costs

Control: total, on close. Time: a growth-stage process runs 6–12 months, consumes the CEO almost entirely, and has a material chance of failing after you have told your team. Optionality: there is a real and under-discussed opportunity cost in starting a sale process — it distracts the company, informs competitors, and if it fails, leaves you with a demoralized team and a market that knows you were for sale.

Failure modes

1. Selling into a preference stack you did not model. Covered above. This is the single most common way a "successful exit" becomes a personal non-event.
2. The single-acquirer trap at scale. One strategic buyer with a long process and no competing bid will re-trade on price after diligence, and you will have no alternative.
3. Diligence-driven re-trade. Customer concentration, contract assignability problems, unassigned IP, and open-source license issues are the four recurring re-trade triggers. All four are preventable years earlier (Part C).
4. Signing and not closing. Regulatory review, financing failure, or a material adverse change. The gap between signing and closing is where deals die.
5. Telling the team too early or too late. There is no good answer, only trade-offs.

Base rates

[Verified] US VC-backed exits totalled 1,185 (2023), 1,259 (2024), and 1,135 through Q3 2025 (PitchBook-NVCA). Exit value in 2025 was strong — $204.9B through Q3, exceeding 2022–24 combined — but that value is concentrated in a small number of very large transactions, which is the definition of a power-law distribution.

[Analysis] The honest framing: a $100M+ strategic sale is roughly a 1-in-100 to 1-in-200 outcome for a company that raises a seed round, and the conditional probability of getting there is the product of several stage-graduation probabilities each well below 50% (Chapter 7 §A8). Planning your personal finances around it is not a plan.

A5. Earn-outs — as a strategic choice, not a deal mechanic

This is the section founders most need and least often get honestly, so let me state the conclusion first:

An earn-out is not deferred purchase price. It is a contingent claim on a business you no longer control, written by the person who controls it, and on the available evidence it is worth roughly one-fifth of its face value.

What the data says

The best available data comes from SRS Acquiom, which acts as shareholder representative and payments agent on thousands of private-target acquisitions and therefore sees not just what was promised but what was paid. This is the rare dataset that is not survivor-biased in the usual direction, because it follows the same deals forward.

[Verified] From SRS Acquiom's analysis across its 2025 Deal Terms Study, 2025 Life Sciences Study, and 2024 Claims Insights Report (SRS Acquiom, M&A Claims, Undisclosed Liabilities and Earnout Achievement):

• Just over 50% of deals with earnouts see any payment at all. Slightly under half pay nothing.
• Among deals that do pay, the average payout is about 50 cents on the dollar of the earnout opportunity.
• Across all deals with earnouts, including the non-payers, SRS Acquiom's own framing is that "closer to one out of five dollars gets paid" — roughly 20 cents on the dollar. [Verified for the aggregate. The decomposition of that aggregate into a payment rate and an average recovery among payers is not published, so treat any such split as inference rather than data.]
• Lower-middle-market deals (≤$50M upfront) generally experience lower achievement rates than larger ones — which is precisely the segment most founders reading this will be in.
• In life sciences, where milestones are technical and long-dated, the figure is roughly 19 cents on the dollar.

[Verified] Earnouts are also becoming more common: 24% of private-target non-life-sciences deals in 2025 included an earnout, up from 19% in 2014 (SRS Acquiom, Earnout and Milestone Trends). The underlying deal-terms dataset covers 2,300+ private-target acquisitions worth $569B, closed 2020–2025 (SRS Acquiom 2026 Deal Terms Study) [Verified].

Sit with the central number. Twenty cents on the dollar. If a buyer offers you "$40M — $25M at close and $15M in earn-out," the honest expected value of that deal is not $40M. It is approximately $25M + (0.20 × $15M) = $28.0M, and in the lower middle market probably less. If you would not take $28M, you should not take the deal.

Why so much price gets deferred

[Analysis] Four reasons, in rough order of honesty:

1. Genuine valuation disagreement. The seller believes the pipeline will convert; the buyer does not. An earn-out bridges a real gap and lets a deal happen that otherwise would not. This is the textbook justification and it is sometimes true.
2. Information asymmetry. The academic literature on earnouts treats them primarily as a screening device: sellers with private good information will accept contingent consideration, and sellers without it will not. That is a reasonable theory, and it also means agreeing to a large earn-out signals confidence that the buyer then prices against you.
3. Retention. An earn-out is a golden handcuff that costs the buyer nothing if you leave, and that does not have to be expensed the way a retention bonus does.
4. Price optics. A large headline number with a large contingent component lets both sides announce a figure neither expects to pay. This is more common than anyone admits.

Why founders systematically overestimate their earn-out proceeds

[Analysis] A stack of cognitive and structural problems, all pointing the same direction:

• The targets come from your plan, and your plan was optimistic. Earn-out thresholds get negotiated against the forecast you showed the buyer to maximize the headline price. You have anchored the hurdle to your best case.
• You lose control of the inputs on closing day. The acquirer now controls pricing, packaging, headcount, roadmap, sales compensation, channel conflict, and how your product is positioned against theirs. Every one of those levers affects your number.
• Integration destroys the thing being measured. Revenue migrates into their billing system, salespeople move onto their comp plan, the product gets bundled. Six months later "your" revenue is not a well-defined quantity, and the person defining it has an incentive.
• Accounting definitions drift — revenue recognition, shared-cost allocation, transfer pricing, treatment of the acquirer's own sales to your customers — and the acquirer's finance team decides them.
• The counterfactual is unknowable. Even with a "commercially reasonable efforts" clause, proving the acquirer's decisions caused your miss means litigating against a far better-resourced party using their documents.
• Survivorship bias in the stories you hear. Founders who hit their earn-outs write about it. The half who got nothing do not.

What life is actually like inside an acquirer for the earn-out period

[Analysis] Founders should plan for this honestly, because it is usually one to three years of their life.

• You have a boss, probably for the first time in years, and your authority is ambiguous: you "run" your business unit but cannot hire, set prices, change the roadmap, or spend without approval.
• Your team notices you have changed. The founder who fought for them is now explaining corporate policy to them.
• Integration consumes the calendar — systems migration, security review, HR onboarding, legal harmonization — none of which advances the earn-out metric.
• Your champion may leave. The executive who bought you is the one person who cares whether you succeed, and executive tenure at large companies is short. If your sponsor departs, your earn-out is administered by someone with no stake in it and possibly a competing product.
• Attrition is a one-way door, and you cannot backfill at the speed you used to.
• Reorganizations reset everything. A reorg in month nine of a 24-month earn-out can restructure your business unit out of recognizable existence while leaving the clause intact and unmeasurable.

The strategic conclusions

[Analysis] Five practical positions:

1. Discount earn-out consideration to roughly 20–30% of face when comparing offers, then compare. This single adjustment changes which deal is better in a surprising number of cases.
2. Prefer cash at close, even at a materially lower headline price. A $30M all-cash offer generally beats $45M with $20M in earn-out.
3. Prefer metrics you can influence and that survive integration — gross bookings of a distinctly branded product, retention of named accounts, a specific technical milestone — over anything requiring cost allocation or that the acquirer can bundle away. Revenue beats EBITDA; anything with "contribution margin" in it is worse than both.
4. Negotiate the operating covenants as hard as the number: headcount and budget commitments, pricing autonomy, the right to run the business "substantially as conducted prior to closing," an agreed accounting methodology, and an audit right on the calculation. The covenants are worth more than an extra $5M of face value. Chapter 14 covers the drafting.
5. Assume you will not be there at the end. Negotiate what happens if you are terminated without cause or resign for good reason. If the answer is "you forfeit," the earn-out is a retention device and should be valued near zero.

Who earn-outs suit

[Analysis] Founders who genuinely believe their forecast, intend to stay, have a well-defined and separable product line, and are selling to a disciplined acquirer with a track record of paying earn-outs (which you can and should diligence by talking to founders of their prior acquisitions — this is the highest-value hour in the whole process). Almost nobody else.

A6. Acqui-hire

What it is

An acquisition where the buyer wants the people, not the business. The product is usually shut down or absorbed; the customers are migrated or abandoned; the team is hired. Classically this was a soft landing for a failing company. Since roughly 2024 it has become, in AI, something else entirely — a primary strategy for acquiring scarce talent, at prices that look nothing like distress.

There is also a structurally distinct variant that matters: the "reverse acqui-hire" or license-and-hire, where the acquirer does not buy the company at all. It licenses the technology non-exclusively, hires the founders and key staff directly, and leaves the corporate shell standing with its investors and remaining employees inside it. This structure exists in part because it avoids the merger-review and successor-liability profile of an acquisition.

What it pays, and how it splits

[Verified] Recent landmark transactions, which are the extreme tail and should not be read as typical:

• Microsoft / Inflection (2024): approximately $650M paid to the company, against $1.5B raised — investors recovered roughly 1x–1.5x (Heavybit, Acqui-hires in the Age of AI).
• Google / Windsurf (2025): approximately $2.4B for a licensing arrangement plus the hiring of the CEO and key researchers, against $243M raised (same source).
• Meta / Scale AI (2025): a 49% stake at a ~$14B valuation, structured so that the founder joined Meta (same source).

[Verified] For ordinary acqui-hires — which is what almost every reader of this chapter is actually contemplating — the economics are very different and the key structural fact is this: most of the value flows through employment compensation to founders and key engineers, not through the corporate equity waterfall (CRV, Acquihire: How to Protect Your Team in an Acquisition).

That sentence contains the entire politics of the transaction:

• The corporate purchase price — usually small, often calibrated to roughly cover the preference stack or less — flows through the waterfall. Preferred gets it. Common usually gets little or nothing.
• Signing bonuses, salary, and acquirer equity grants for the people the buyer wants go directly to those individuals, bypassing the cap table entirely.
• Common shareholders who are not hired get the worst of both: no employment package and no equity value.

[Analysis] The standard per-head framing you will hear — a range per engineer, historically a few hundred thousand to low millions of dollars of total consideration, with elite AI researchers commanding far more — is a useful anchor for the buyer's internal budget, but it is not how the money reaches you. What reaches you is a comp package. Negotiate the comp package.

What it signals, and when it beats shutting down

[Analysis] An acqui-hire signals that the market valued your team and not your product. That is not nothing: it preserves your ability to raise again, keeps your team employed, avoids the reputational mess of a disorderly shutdown, and typically returns some capital to investors.

It beats shutting down when: the team wants to stay together; the alternative is a wind-down that returns less to investors than the acqui-hire price; and the buyer's package for the individuals exceeds what they would get in the open market, which it usually does because the buyer is paying a premium for a pre-formed team.

It is worse than shutting down when the terms require you to personally stay for two years in a role you will hate for compensation you could get elsewhere, and when accepting it means employees who built the thing receive nothing while you receive a package. That second scenario is common, and it is a genuine ethical problem, not merely an awkward one.

How to handle the team problem

[Analysis] Three things, in order:

1. Negotiate a carve-out for employees as part of the transaction, funded from the purchase price before the preference waterfall. This requires investor consent because it comes out of their proceeds. Ask early, when you still have leverage, and frame it as necessary to deliver the team the buyer is paying for — which is true.
2. Get written offers for the people, not verbal assurances. CRV's guidance is to secure written protections on vesting, cliffs, post-termination equity treatment, role definition, and reporting lines before signing, because "employee outcomes depend on which people the buyer actually chooses to hire and what terms they receive" (CRV) [Verified].
3. Tell people the truth about what they will get. The reputational cost of a founder who got rich from an acqui-hire while their team got nothing and found out at closing follows that founder for a decade.

Base rates

[Analysis] There is no clean count of acqui-hires because they are rarely disclosed as such and are frequently structured as asset purchases or hiring events. The reasonable estimate is that a meaningful minority of the ~1,200 annual US VC-backed "exits" are acqui-hires or soft landings — and that the number of companies that attempt one and fail to find a buyer is considerably larger, because an acqui-hire requires the acquirer to want your specific people at a moment when they have budget. Do not plan on it as a backstop. It is a lucky outcome, not a floor.

A7. Private equity recapitalization and majority sale

What it is

A private equity firm buys a controlling stake — commonly 60–80% — in a profitable company, usually financed partly with debt. The founder takes most of their value off the table in cash and "rolls over" a minority stake into the new entity. The PE firm professionalizes, acquires add-ons, grows EBITDA, and sells again in roughly five to seven years. The founder's rolled equity participates in that second sale: the "second bite of the apple."

This has become one of the most important destinations for profitable small software businesses, and it is the one most likely to be newly available to a reader who has been quietly compounding a bootstrapped business.

The market, as of 2026

[Verified] (Bain & Company, Global Private Equity Report 2026):

• 2025 global buyout deal value: $904B, up 44% year over year, across 3,018 deals (count down 6%). Average disclosed deal size hit a record $1.2B.
• Exit value: $717B, up 47%, across 1,570 exits.
• Unsold portfolio backlog: ~32,000 companies worth $3.8 trillion.
• Median holding period at exit: ~7 years, up from 5–6 years in 2010–2021.
• Distributions to LPs: 14% of NAV in 2025, below 15% for four consecutive years.
• Dry powder: $1.3 trillion in global buyout capital.

[Analysis] Read those numbers together and you get the strategic picture: enormous capital that must be deployed, a large backlog of unsold assets, and lengthening hold periods. For a founder, the first fact is why PE firms are calling you. The second and third are why the second bite takes longer than the pitch implies.

In software specifically, PE was involved in 58% of the record 2,698 SaaS M&A transactions in 2025 (Software Equity Group) [Verified] — PE is now the majority buyer of software companies by deal count, not strategics.

What it pays, and typical multiples

[Analysis, calibrated to cited data] For a profitable small software business, the pricing conversation is usually run on EBITDA rather than revenue once the company is genuinely profitable, and the multiple depends heavily on size, growth, and retention:

For context, the search-fund segment — which buys small, non-software businesses with similar discipline — paid a median 6.2x EBITDA on a median $2.5M of EBITDA and median $16M enterprise value in 2024–25 acquisitions (Stanford GSB 2026 Search Fund Study data, via CapitalPad) [Verified]. That is a good anchor for what disciplined small-company buyers actually pay.

The second-bite pitch, and how often it pays off

The pitch: you sell 70% today for cash, roll 30%, the firm triples EBITDA in five years, you sell again at a similar or better multiple, and your 30% is worth more than your original 70%.

[Analysis] Three honest observations:

1. The arithmetic can genuinely work. If EBITDA triples and the multiple holds, a 30% stake in the new entity is worth roughly 90% of what 100% of the old entity was worth. This is not a con; it is the actual model and it has made many founders wealthy.
2. It runs on leverage, which cuts both ways. If the plan works, your rolled stake compounds faster than the business does. If EBITDA is flat or falls, debt service comes first and your minority equity can go to zero while the company still exists. Rolled equity is the riskiest money in the structure and it sits behind the debt.
3. The timeline has slipped. With median hold periods near 7 years and a $3.8T backlog of unsold portfolio companies (Bain 2026) [Verified], the "five year" second bite is realistically seven to nine years illiquid, possibly ending in a continuation vehicle rather than a clean sale. Do not roll money you will need.

[Analysis] There is no public dataset on what fraction of founder rollover stakes produce a larger second bite than the first, and I am skeptical of any source that claims one — the data would have to come from PE firms reporting on their own outcomes. Treat the second bite as a genuine call option with meaningful probability of being worthless, and size your rollover so that the first bite alone is an acceptable outcome. If the sponsor pressures you to roll more than you are comfortable with, that pressure is itself information.

What PE ownership does to the company and to your daily life

[Analysis] Consistently, across the founders I have seen described in practitioner accounts and the structural logic of the model:

• Reporting changes immediately: monthly board packs, a defined KPI set, weekly flash reports, a replaced or outsourced finance function, a formal budget process. A CFO arrives early, often before anything else changes; a CRO often follows within a year.
• Debt service becomes the governing constraint. Decisions a debt-free owner makes on instinct — deferring revenue, funding a two-year bet, tolerating a bad quarter — become covenant conversations.
• Pricing goes up and costs come down. The two fastest EBITDA levers; every sponsor pulls them. Some is genuine professionalization; some is extraction that will visibly annoy your longest-standing customers.
• Add-on acquisitions become part of your job under the buy-and-build playbook.
• Your role is defined by the sponsor, and it may not be CEO for long. Founders who stay on after a majority sale frequently become CTO, Chief Product Officer, or a board member. The median outcome is that the founder is not running the company three years later. [Estimate — this is the consistent practitioner pattern; I have no clean dataset for it.]
• You are an employee with a minority stake, subject to a shareholders' agreement that will contain drag-along, tag-along, leaver provisions (good leaver / bad leaver), and non-competes. Read the leaver provisions before you read the price. Bad-leaver treatment can convert your rollover into a return of capital at cost.

Who it suits

Founders of profitable businesses of $2M+ EBITDA who want substantial liquidity now, are willing to trade control for it, and either want to keep operating with more resources or want a defined path out. It suits businesses with recurring revenue, low churn, and fragmented markets where add-on acquisitions are available. It suits founders who are genuinely done with being personally on the hook for everything.

It does not suit founders whose primary value is autonomy, or businesses whose quality depends on choices (unprofitable customer service, under-monetized pricing, long-horizon R&D) that a leveraged owner will reverse.

What has to be true, years in advance

Audited or at minimum reviewed financials, ideally three years. Clean revenue recognition. A management team that is not just you. Customer concentration below roughly 20% for any single account. Documented contracts with assignability and change-of-control terms that do not blow up on a transaction. A cap table simple enough to buy — a venture preference stack larger than the purchase price makes a PE deal impossible, because the sponsor cannot buy control cheaply if the preferred must be paid first.

Failure modes

1. Rolling too much. The most common and most expensive error.
2. Over-leveraging into a downturn: covenant breach, sponsor recapitalizes, your minority stake is crushed or washed out.
3. Signing a shareholders' agreement you did not model. Leaver provisions, drag-alongs and ratchets on the sponsor's preferred return can make your 30% economically far less than 30%. Hire your own counsel, not the company's — a genuine requires-a-professional moment.
4. Cultural collapse. Your team joined a founder-led company; several key people will leave within a year. Fund retention explicitly from the deal.
5. The sponsor's preferred return sitting ahead of your rollover. Many structures place sponsor capital in a preferred position with an 8% accruing return, above which your common rollover participates. Ask directly.

A8. Micro-PE, search funds, holdcos and the small-business acquisition market

What it is

A genuine, functioning market for businesses far too small for institutional PE: self-funded and traditional search funds, ETA (entrepreneurship-through-acquisition) buyers, micro-PE funds, permanent-capital holdcos, and online marketplaces. The buyers are individuals or small firms looking for a business that produces cash and does not require the seller to stay.

Venues as of 2026: Acquire.com (SaaS and small internet businesses, successor to MicroAcquire), Flippa, Empire Flippers (content, e-commerce, some SaaS), Quiet Light, plus broker networks and the search-fund ecosystem for anything with real EBITDA.

What it pays

[Verified] Current pricing, from the most credible marketplace and academic data available:

• Small SaaS: Acquire.com reports a median profit multiple of 3.9x, stable across 2024 and 2025, with averages in the low-to-mid 4x, focused on enterprise value below $10M. Average time on market: 81 days, with most deals closing within 90 days. Average profit margin on listed businesses: 71% in 2025 (Acquire.com Biannual Acquisition Multiples Report, Jan 2026).
• Content and e-commerce: Empire Flippers values on a monthly multiple of trailing-12-month average net profit, typically 30–50x monthly — roughly 2.5x–4x annual net profit — with a practical floor around $100K of valuation and a preference for businesses with 3+ years of history. Commission is tiered: 15% on the first $700K, 8% from $700K–$5M, 2.5% above (ExitBid, Empire Flippers Review 2026). Note that on a sub-$700K sale the commission is a material haircut.
• Real-EBITDA businesses (search fund / micro-PE territory): the Stanford search fund data shows median 6.2x EBITDA paid on median $2.5M EBITDA and median $16M enterprise value for 2024–25 acquisitions (via CapitalPad).
• Main-street businesses generally: BizBuySell's Q1 2026 data shows a median sale price of $350,000 on median cash flow of $165,256 — an average 2.7x cash-flow multiple (BizBuySell Q1 2026).

[Analysis] Note the pattern: multiples rise sharply with size and with owner-independence. The same dollar of profit is worth 2.7x in a main-street business dependent on its owner and 6.2x in a business with a management team. That gap is the single largest lever a small-business owner controls.

The buyer side, and its base rates

Search funds are the most thoroughly documented small-acquisition buyer, and the data is sobering in a useful way. [Verified] Through December 2025, across 862 core US and Canadian search funds tracked since 1984: an aggregate 33.9% pre-tax IRR and 4.75x return on invested capital; 58% of concluded searches acquired a company (only 48% among 2021–2024 vintages); and 26.4% of the 337 acquired companies experienced total or partial losses (Stanford GSB 2026 Search Fund Study data, via CapitalPad).

[Analysis] For a seller, the operationally important facts are: roughly half of searchers never buy anything, so a searcher's expressed interest is weak evidence of a deal; and a quarter of completed acquisitions lose money, which is why they diligence hard and why seller financing is so often requested.

What makes a business saleable at this scale

[Analysis] In descending order of impact:

1. Clean, separable financials — accrual basis, no personal expenses entangled, quality-of-earnings-ready. Worth more multiple than anything else on this list.
2. Owner-independence. If the business requires you, the buyer is buying a job and will pay job prices.
3. Revenue durability: contracts, recurring billing, low churn, no single customer above ~15–20%.
4. Transferability of everything — domains, repos, cloud and app-store accounts, payment processors, ad accounts — without customer consent or platform approval.
5. Documentation. SOPs, onboarding, runbooks; buyers at this scale are often first-time operators.
6. No platform single point of failure, or an honest acknowledgment of one. Hiding it kills the deal in diligence.
7. Clean IP and open-source hygiene (Chapter 11 §2 and §6).

Who it suits, and the failure modes

[Analysis] Suits bootstrapped founders, solo operators, and small teams with $100K–$3M of annual profit who want a clean cash exit and are prepared to spend 6–12 months preparing. Does not suit anything with institutional preferred stock on the cap table, because the buyers in this market will not navigate a venture waterfall for a $2M deal.

Failure modes: listing before the books are clean (you get a lowball or no offers); a seller who is the business; SBA financing falling through — 45% of brokers report that current lending conditions make deals harder to complete, and new citizenship requirements effective March 2026 further restrict the buyer pool (BizBuySell Q1 2026) [Verified]; and accepting heavy seller financing or an earn-out that converts a clean exit into a multi-year receivable from a first-time operator. 61% of buyers hope for seller participation (same source) [Verified] — which means you should expect to be asked and should decide in advance what you will accept.

A9. ESOP and employee ownership

What it is

An Employee Stock Ownership Plan is a qualified retirement plan that holds employer stock. The company (or a trust it establishes) buys the owner's shares, typically financed with a combination of bank debt and a seller note, and allocates the shares to employee accounts over time. The seller gets paid; the employees become beneficial owners; the company continues independently.

How it works and what it pays

[Verified] (NCEO, ESOP Tax Incentives and Contribution Limits):

• Section 1042 rollover: an owner of a closely held C corporation who sells at least 30% of the company to an ESOP and reinvests the proceeds in "qualified replacement property" (securities of domestic operating companies) within the prescribed window can defer capital gains tax on the sale.
• S-corporation ESOP exemption: the ESOP is not taxable on its share of corporate earnings. A 100% S-corp ESOP effectively pays no federal income tax — an extraordinarily powerful structure for a cash-generating business, because the tax savings service the acquisition debt.
• Contribution deductibility is generally capped at 25% of eligible payroll, with higher effective limits for leveraged C-corp structures; dividends on ESOP-held stock used for specified purposes are separately deductible.

Timeline and proceeds: the owner typically receives part cash at close (funded by bank debt) and the balance as a seller note over 5–10 years, often with warrants. The valuation is set by an independent appraiser at fair market value — which means you should expect a fair price, not a strategic premium. An ESOP will generally pay less than a strategic acquirer and roughly what a financial buyer would pay. The compensation is in tax treatment, control of the outcome, and what happens to your employees.

Scale, and why tech ignores it

[Verified] As of 2023 data (the most recent DOL Form 5500 figures, which run about two years behind): 6,609 ESOP plans across 6,411 companies, 15.1 million participants, over $2 trillion in assets. Privately held companies account for 6,098 plans, 2.87 million participants and $363.4B. In 2023, 309 new ESOPs were formed (NCEO, Employee Ownership by the Numbers).

[Analysis] Why almost none are tech startups: ESOPs require stable, predictable cash flow to service acquisition debt, and venture-stage companies have none. They require a clean equity structure — preferred stock with liquidation preferences and a venture board is fundamentally incompatible. Setup is expensive and slow: feasibility study, valuation, trustee, legal and financing commonly run $150K–$500K+ over 6–12 months [Estimate; practitioner range], disproportionate below roughly $5M of enterprise value. There is a vocabulary gap — tech thinks in options and RSUs; an ESOP is an ERISA retirement plan with a trustee and a fiduciary. And there is an ongoing repurchase liability: the company must buy shares back from departing employees, a real and growing balance-sheet commitment that many sponsors underestimate.

Who it suits

Profitable, stable, people-intensive businesses — engineering firms, manufacturers, contractors, agencies, professional services — with $2M+ EBITDA, a management team, and an owner who wants liquidity, tax deferral, and continuity for staff rather than maximum price. It suits founders for whom "the company keeps existing and my people own it" is worth more than the last turn of multiple.

This is unambiguously a requires-a-professional structure. ESOP formation involves ERISA fiduciary law, independent valuation, trustee selection, and financing simultaneously. Bad ESOPs are a recognized category of litigation, usually about overvaluation at formation.

Lighter alternatives

[Analysis] If full employee ownership is the goal but an ESOP is too heavy: an employee ownership trust (EOT) — a perpetual trust holding shares for employees' benefit, simpler than an ESOP but without the 1042 or S-corp tax benefits; a worker cooperative; phantom equity or profit-interest plans that share economics without transferring control; or simply broad-based equity with a genuine tender-offer program. Each is cheaper and each gives up something.

A10. Steward ownership, purpose trusts, and long-term structures

What it is

A family of structures that permanently separate control from economic extraction, so the company cannot be sold or taken public in the ordinary way and profits are directed to a purpose rather than to shareholders seeking exit value. The umbrella term is steward-ownership; the associated movement in startup circles is the Zebras community and the Purpose network.

How Patagonia actually works — the canonical example

[Verified] (Cole Schotz, Charitable Planning Update: Yvon Chouinard Donates Patagonia Stock; ABA, Patagonia, Purpose Trusts, and Stewardship Trusts):

• 2% of the stock — all the voting shares — went to the Patagonia Purpose Trust, a business purpose trust whose stated object is to protect the company's mission and values. It operates through a stewardship committee directing the trustee, with at least one trust enforcer responsible for holding the trustee to the purpose. The structure uses Oregon's stewardship trust statute.
• 98% of the stock — all non-voting — went to the Holdfast Collective, a newly created 501(c)(4) social welfare organization, which can lobby and engage politically on environmental causes in a way a 501(c)(3) could not.
• Tax treatment: approximately $17.5M of gift tax was paid on the transfer to the purpose trust. No gift tax on the 501(c)(4) transfer. No charitable income tax deduction was available for either, because neither a business purpose trust nor a 501(c)(4) is a "charitable" vehicle for that purpose. Capital gains tax was avoided by not selling. Estimated estate tax avoided: on the order of $1.2 billion.

What these structures actually do, and their real limitations

[Analysis] What they do: make a sale or IPO structurally very difficult; lock a mission in place across generations; direct profit to a purpose; and, in the Patagonia case, transfer enormous value out of the estate tax base. They are serious instruments and not merely symbolic.

What they do not do, and where founders get this wrong:

1. They are not a liquidity event. Chouinard gave the company away. If you need money from your company, this is not the structure. Steward ownership converts an asset into a stewardship responsibility.
2. They require you to already own the company. Existing preferred shareholders have economic rights that a conversion impairs; converting requires buying them out at fair value, which requires capital, which is the problem the structure was meant to solve. This is why steward ownership is effectively foreclosed by the first priced venture round.
3. Legal availability varies by state. Not every US state permits perpetual purpose trusts; some UTC states still apply the rule against perpetuities, and courts may reduce a trust's value if deemed excessive (ABA) [Verified]. Oregon, Delaware and a handful of others are the practical venues.
4. Governance quality is the entire ballgame and it is unsolved. The trust's mission is enforced by a committee and an enforcer. If those people are captured, incompetent, or simply die, there is no shareholder to complain and no market discipline. Steward ownership replaces the tyranny of the shareholder with the risk of an unaccountable trustee. There is no good empirical evidence yet on how these structures perform over multiple generations of leadership, because almost none of them have been through one.
5. Capital access narrows permanently. You can raise debt, revenue-based financing, or mission-aligned capital with capped returns. You cannot raise conventional equity. For most businesses that is a real growth constraint, and it is the trade the structure is making on purpose.
6. Founder compensation becomes a governance question, not a matter of ownership. You are paid a salary decided by a committee.

Who it suits, and base rates

[Analysis] Suits profitable, founder-owned businesses with a genuine mission dependency, a founder who is already financially secure, and no institutional preferred on the cap table. Base rate: vanishingly rare. A few hundred companies globally use recognizable steward-ownership structures; a handful are large. It is an important option to know exists and an unlikely one to use.

A11. IPO

What it is

Selling newly issued (and sometimes existing) shares to the public through an underwritten offering, after which the company's stock trades on an exchange and the company becomes subject to continuous SEC reporting.

An IPO is not an exit. It is a financing event that converts your illiquid stake into a liquid one on a delay, and simultaneously converts your job into a substantially different job. Founders who treat it as a finish line are the ones who have a bad time.

The state of the window, September 2026

[Verified] (Jay R. Ritter, IPO Statistics, University of Florida, updated 2026; Ritter, VC-backed IPOs):

[Analysis] The 2025–26 window reopened, but it reopened selectively and at a much higher quality bar than 2021. Note the two most important columns: 67% of 2025 IPOs were profitable (against 49% in 2021), and median age at IPO was 12 years. The window is open for mature, profitable, large companies. It is closed for the 2021 profile.

The size threshold moved with it. Median last-twelve-months sales for 2025 IPO firms was $180M in 2024 dollars, against $57M in 2024 and $128M in 2021 (Ritter, IPO Sales table, updated January 2026) [Verified].

[Analysis] Practical thresholds for a software company contemplating a 2026-era IPO: roughly $200M+ of revenue, 25%+ growth at that scale, a credible path to GAAP profitability or actual profitability, net revenue retention comfortably above 110%, predictable quarterly forecasting, and two to three years of audited financials with no material weaknesses. Companies have gone public below that. They have generally regretted it.

What it costs

[Verified against PwC's published IPO cost survey; figures rounded] (PwC, Considering an IPO? First, understand the costs):

• Underwriter discount: roughly 4–7% of gross proceeds, with smaller deals at the top of the range.
• Other one-time costs — legal, accounting, printing, advisory, registration — commonly several million dollars, and larger for companies that need remediation work before filing.
• Incremental recurring cost of being public: roughly $1–2M per year, covering audit, SOX 404 compliance, D&O insurance, additional finance and legal headcount, investor relations, exchange fees and reporting infrastructure. For a company below $100M of revenue, that is a material margin hit.

Control: you keep operating control if you keep board support, but you acquire a new constituency with quarterly expectations and the ability to sell. Dual-class share structures preserve founder voting control and are common in tech IPOs; they are also increasingly contested by index providers and institutional investors, and they often sunset.

Time and life: the S-1 process consumes the CEO and CFO for six months. Post-IPO, a meaningful fraction of the CEO's calendar goes permanently to earnings preparation, analyst and investor meetings, and board and audit-committee work. Your compensation becomes public. Your mistakes become public and, if material, litigated.

Lockups: the conventional structure is a 180-day lockup from pricing, increasingly with staged early-release triggers tied to price performance or the second earnings report. So the liquidity you went public for arrives six months later, subject to Rule 144 volume limits for affiliates, insider trading windows, and a 10b5-1 plan you must adopt in advance. [Verified as standard market practice; specific terms vary by deal.]

Failure modes

[Analysis] Going public too small, so that compliance cost eats margin and no analyst covers you. Missing the first or second quarter after IPO, which permanently repositions the stock and is extraordinarily hard to recover from. Going public into a closing window and trading below issue for years, which poisons employee equity and recruitment. And the founder failure mode: discovering that the job you now have — quarterly guidance, investor management, public scrutiny — is not the job you wanted, while being the person least able to leave.

Base rates

[Analysis] Against roughly 40,000–50,000 active US venture-backed companies and 49 VC-backed IPOs in 2025, the annual IPO rate is on the order of 0.1%. Over a company's lifetime, the probability of a venture-backed US startup reaching an IPO is in the low single digits of a percent — and that population is already filtered to companies that raised institutional capital. If you are reading this chapter to decide something, the IPO is almost certainly not the relevant branch of the tree.

A12. Direct listing and SPAC

Direct listing

What it is: the company registers existing shares for resale and lists them on an exchange without an underwritten primary offering. No new capital (unless done as a "direct listing with a capital raise," permitted since 2020), no underwriting discount, and no lockup — existing holders can sell from day one.

[Analysis] It suits a very narrow profile: a company with a consumer-recognizable brand that needs no primary capital, has a large existing shareholder base wanting liquidity, and can attract natural demand without a bookbuilt allocation process. The canonical examples are Spotify, Slack, Palantir, Coinbase and Roblox. The total count of US direct listings since 2018 is in the low dozens (Ritter, Direct Listings in the U.S., updated June 2026) [Verified that Ritter maintains this table; I was unable to retrieve the current count at research time — treat the "low dozens" figure as an estimate].

For almost every company reading this, the direct listing is a footnote. It is worth knowing it exists mainly because it clarifies what an IPO's underwriting fee and lockup are actually buying: price stabilization, distribution, and an orderly float.

SPAC

What it is: a blank-check company raises money in an IPO, then merges with a private operating company, taking it public. In 2020–21 this was sold to founders as a faster IPO with negotiated pricing and the ability to use forward projections.

The record, stated plainly. [Verified] (Ritter, SPAC statistics, University of Florida):

[Analysis] That is not a bad patch. That is a structural result, stable across five consecutive cohorts and across a rising equity market. The de-SPAC vehicle has destroyed roughly 60% of post-merger shareholder value per year, on average, every year since 2021. The mechanism is well understood: sponsor promote dilution, redemption-driven collapse of the trust, PIPE overhangs, and a selection effect in which companies that could do a conventional IPO generally did one.

Note that SPAC IPO issuance recovered sharply in 2025 to 144 vehicles and $26.9B — meaning the supply of sponsors looking for targets in 2026 is large, and some of them will call you. The 2025 vintage is structured more conservatively than 2021 (higher sponsor at-risk capital, more overfunded trusts, shorter deadlines), but its post-merger record is not yet observable — the returns column reflects de-SPACs completing their first year after merging, not the cohort that IPO'd that year — so treat the structural improvement as untested rather than disproven.

[Analysis] The honest advice: if a SPAC is your only route to being public, that is strong evidence you should not be public. The one defensible use case is a capital-intensive business with a genuine strategic sponsor and a committed, non-redeeming anchor investor — and even then, model the fully diluted outcome after maximum redemption before you sign anything.

A13. Wind down with capital returned

What it is

Deciding the company will not work, stopping while money remains, paying creditors and obligations, returning the remaining cash to shareholders, and dissolving. Chapter 11 §B8 covers the mechanics — dissolution filings, final payroll, WARN, IP disposition, record retention. This section is about why it is a strategic choice rather than a failure event.

What it pays

Usually little, in absolute terms. What it preserves is worth more than what it pays.

[Analysis] Returning $2M of a $5M seed round means your investors got 40 cents back. That is not a good outcome, but it is a categorically different outcome from zero, and the people on the receiving end know the difference. What you buy:

1. Fundability. The single most reliable path to raising your second company is having handled the end of your first one well. Investors talk about this specifically and remember it for a long time.
2. Reputation with your team. People who were given honest notice, full final pay, extended healthcare where possible, and active help finding roles will work for you again and will say so publicly.
3. Your own capacity. A clean shutdown takes two to four months. A slow-motion one takes two years and leaves you depleted.
4. Legal cleanliness. Dissolving properly, with creditor notice and a proper claims process, closes off personal exposure. Abandoning a company — letting it go dark with unpaid obligations and unfiled returns — creates liabilities that follow you.

Why founders delay past the point of rationality

[Analysis] The reasons are well-documented in behavioural terms and universally recognized by anyone who has done it:

• Sunk cost. Years of work and other people's money create an obligation to keep trying that feels moral and is simply an error.
• Identity. "Founder of a company" and "person who shut a company down" feel like different people. They are not.
• Optimism as a job requirement. The disposition that let you start makes you the worst-calibrated person to judge when to stop.
• The pivot mirage. There is always another pivot. Some work. Most are a way of spending the money that should have been returned.
• Fear of the investor conversation, which is almost universally far less bad than expected. Professional investors have had it dozens of times. What they resent is not failure; it is finding out late.
• Employees' livelihoods — the only honourable reason on this list, and one that argues for deciding earlier, because an earlier decision means more severance and more notice.

[Analysis] The practical test: write down, in advance, the conditions under which you will stop — a runway threshold, a growth threshold, a date — and share them with your board. Founders who set this tripwire at a moment of clarity make better decisions than founders deciding in month 34 with eleven weeks of cash. The tripwire should be set at a point where enough cash remains to wind down properly and return something, which in practice means roughly 6 months of runway, not 6 weeks.

A14. The zombie — default alive but going nowhere

What it is

The most common unacknowledged endgame. The company is not failing: it has revenue, it covers its costs or nearly does, the team is small and stable. It is also not going anywhere: growth is in the low single digits, the market is not expanding, no acquirer is interested at a price that clears the preference stack, and no investor will lead a new round. It can continue in this state for a very long time, and frequently does.

How to recognize it

[Analysis] Honest diagnostic questions, all of which should be answered with numbers:

• Has ARR grown less than 15% in each of the last two years, with no identified cause you are actively fixing?
• Would a new investor, told the truth, lead a round at or above your last valuation? (If you have not tested this in 18 months, you do not know.)
• Is total liquidation preference greater than a realistic sale price?
• Are you spending your time maintaining rather than building?
• Have you had the same headcount for two years?
• Is the most exciting thing that happened this year a renewal?
• Would you invest your own money in this company today at the last round's price?

Three or more yes answers and you are in it. The distinguishing feature of zombie status is not distress — it is the absence of a forcing function. Nothing makes you decide, so you do not.

The options from here

[Analysis] There are five, and the first is the one founders default to:

1. Continue. Legitimate only if you are honest that you are choosing it, you are being paid adequately, and your investors know. It is not legitimate if you are quietly hoping something turns up while burning the option value of your own next five years.
2. Convert to an owner-operator cash business (A1). This requires restructuring the cap table, because a preference stack and a venture board are incompatible with a distribution policy. The mechanism is a negotiated buyback or recapitalization: offer investors a cash payment, a note, or a structured payout to retire the preferred. Many investors will engage, particularly funds near end-of-life that would rather mark something to a real number than carry a zombie position. Chapter 7 §B15 and Chapter 11 §B11 cover adjacent mechanics. This is under-attempted and it is frequently the highest-value move available.
3. Sell for whatever it fetches. Micro-PE, a competitor, a holdco, an acqui-hire. If the preference stack exceeds the price, you need investor consent and a management carve-out; ask for both explicitly.
4. Wind down and return the remaining cash (A13). If the business is not going to produce a return and is not going to produce income for you, the capital is worth more in your investors' hands and your time is worth more elsewhere.
5. Change the business. A real pivot, resourced properly, with a deadline. Different from the pivot mirage in that it has a date attached.

[Analysis] The base rate here is my strongest claim in this chapter and the one I can least support with data, precisely because these companies generate no events: the zombie state is probably the single most common destination for companies that raise a seed round and do not fail outright. The stage-graduation data in Chapter 7 implies a very large population of companies that raised once, did not raise again, and did not exit. Those companies went somewhere. Most of them went here.

A15. Succession, family transfer, and management buyout

Family transfer and succession

What it is: transferring ownership to family members, with or without operating control. Rare in technology, common in the broader economy.

[Analysis] The widely quoted statistic that roughly 30% of family businesses survive into the second generation and around 12% into the third is repeated everywhere and is methodologically weak — it originates from small mid-century samples and conflates "family retained ownership" with "business survived." Family business researchers have criticized it directly (Family Business Consulting Group, Family Business Survival: Understanding the Statistics). Treat the direction as real and the precision as fictional: generational transfer is hard and most attempts do not produce a durable multi-generation business.

The recurring failure modes are governance, not tax: no clear successor, multiple heirs with unequal involvement and equal ownership, a founder who does not actually hand over, and the absence of any mechanism for family members to exit their stake. The cure is structural — a shareholders' agreement with buy-sell provisions, a defined valuation mechanism, and a governance body that is not the dinner table — and it must be built years in advance. Estate and gift tax planning here genuinely requires a professional.

Management buyout

What it is: your existing management team buys the company, typically funded by a mix of bank or SBA debt, a seller note, and a small equity contribution from the managers. Often supported by a mezzanine lender or a small PE firm.

What it pays: fair value, usually less than a strategic would pay, and usually heavily deferred. A typical structure is 50–70% cash at close from debt and 30–50% as a seller note over 3–7 years. That means you are financing your own buyout and you are the most junior creditor — behind the bank, ahead of nobody. If the business struggles under the new debt load, your note is the thing that does not get paid.

Who it suits: owners with a genuine management team who care about continuity, are comfortable with a lower price, and can afford to wait for the money. It suits businesses with stable cash flow and enough debt capacity to fund most of the purchase.

[Analysis] Failure modes: management that can run the business but cannot own it (capital allocation, financing, and pricing are different skills); a seller who stays involved and undermines the new owners; a note structured with no security and no covenants; and the unpriced awkwardness of negotiating hard against people you have worked with for a decade. Get independent counsel and a third-party valuation. The valuation is where MBOs generate lasting resentment in both directions.

A16. Licensing, franchising, and productizing instead of selling

[Analysis] An under-considered branch: monetize the asset without transferring the company.

Licensing. If the valuable thing is technology rather than the business around it, license it: high-margin revenue with no transfer of ownership, and the structure underneath many 2025–26 "reverse acqui-hire" deals. It requires IP that is genuinely separable, clearly owned (Chapter 11 §2), and defensible enough that a licensee will pay rather than rebuild. Risks: creating a competitor with your own technology; a non-exclusive license that collapses your enterprise value; and a licensee who does not renew once they have learned what they needed.

Franchising. Converts an operating playbook into a royalty stream. It works for location-based, replicable, brand-dependent businesses and essentially never for software. It is heavily regulated in the US — a Franchise Disclosure Document under the FTC Franchise Rule plus state registration in about a dozen states — and building a franchise system is itself a business requiring dedicated staff, training infrastructure, and ongoing support. Treat it as starting a second company, not as an exit. Genuinely requires specialist counsel.

Productizing. Turning a services business into a software or fixed-scope product. This is the most commonly attempted and most commonly failed of the three: agencies have been "building a product on the side" for as long as agencies have existed. It fails because the services business pays the bills and therefore always wins the resource conflict. It works only when the product is separately staffed, separately funded, and given a real deadline. The strategic value when it does work is large, because it converts a business worth 2.7x cash flow into one worth 4–8x — which is a reframing of the whole exercise: productizing is not an alternative to selling, it is a way of doubling the price when you do.

PART B — A DECISION FRAMEWORK

The purpose of this framework is not to tell you which endgame is best. It is to tell you which ones are still available, which is a factual question you can answer today, and what you would have to change to open the others.

Work through it in order. Each step eliminates.

Step 1 — Establish the cap table constraint (this is the biggest filter)

Answer three questions with numbers, not impressions:

1. Total liquidation preference, in dollars, including accrued dividends and any seniority or multiples. Pull the actual charter; do not estimate.
2. Who holds blocking rights, and at what threshold. Read the protective provisions in the certificate of incorporation and the voting agreement. Specifically: what percentage of preferred must consent to a sale, and who controls the board?
3. Fund size of your largest investors, and where they are in fund life. A $40M fund in year 3 and a $900M fund in year 8 will behave completely differently about a $20M offer.

[Analysis] If you land in the bottom row, the highest-value work available to you is not growth. It is cap table restructuring — a negotiated buyback, a preference reset, or a recapitalization with a management carve-out. That conversation is uncomfortable and it is usually worth more than a year of execution.

Step 2 — Classify the business

Step 3 — State your personal financial requirement

Write down three numbers:

• Floor: the amount you need to not be financially stressed. Often smaller than founders think.
• Number: the amount at which you would genuinely stop, diversify, and do something else.
• Runway: how long you can personally continue at current compensation.

[Analysis] This step matters more than any other for what you should actually do, and founders skip it because it feels unambitious. If your Number is $3M and you own 40% of a business with $900K of annual owner earnings, you are already at your Number on a five-year distribution basis, and the entire venture-scale apparatus is irrelevant to you. If your Number is $3M and you own 12% of a company with $18M of preference, you need a $43M+ exit to clear it, which sharply narrows the plausible set.

Step 4 — State your time horizon and appetite

• Do you want to be doing this in five years? In ten?
• Do you want to work for someone else for two years? (This single question eliminates or admits earn-outs, most acqui-hires, and most strategic sales.)
• Do you want to manage people, or to build?
• Does control matter to you more than money, or the reverse? Answer honestly; most founders get this wrong about themselves in the optimistic direction.

Step 5 — Intersect, and name two or three

The intersection of Steps 1–4 usually leaves two or three viable destinations, not one. Name them explicitly, in writing, and share them with your co-founder and your board. The act of naming them is most of the value of this exercise, because it converts a drift into a decision and makes the upstream choices in Part C legible.

Step 6 — Identify what you would have to change to open one more

For each closed destination you actually want, write the specific unlock:

Step 7 — Set the tripwires

[Analysis] Write down, now, the conditions that would make you switch destinations: a growth threshold, a runway threshold, an offer threshold, a date. Put them in a document, share them with your board, and review them quarterly. This is the entire mechanism by which founders avoid the zombie state, and it takes an hour.

Worked examples

Example 1 — Bootstrapped $2M ARR micro-SaaS, solo founder

Situation. Solo founder, $2M ARR, 80% gross margin, two contractors, no outside capital, LLC taxed as S-corp, ~$1.0M of annual owner earnings. Founder is 38, moderately tired, Number is $4M, floor is $150K/yr.

Step 1 (cap table): no constraint. Full menu. Step 2 (business type): high-margin software, owner-run, sub-$3M. Natural: owner-operator, marketplace sale, micro-PE. Step 3 (finances): already earning $1M/yr. Number reached in ~4–5 years of distributions, or immediately via sale — at Acquire.com's median 3.9x profit multiple, roughly $3.9M, likely 70–90% cash at close. Step 4 (appetite): does not want a boss; does not want to manage a team.

Available: (a) continue and take distributions; (b) sell on a marketplace for ~$3.5–4.5M; (c) hire a general manager and step back.

[Analysis] Recommendation: continue for now, but spend six months doing sale preparation anyway — clean books, documented operations, reduce personal dependency, move customer relationships onto the company. That work raises the sale multiple, makes the "hire someone and step back" option real, and costs almost nothing. Then reassess with both options genuinely open. The mistake to avoid is drifting for five more years with the business unprepared, because the option to sell decays as the founder's dependency deepens and the technology ages.

What would open more: converting to a C-corp for QSBS would be valuable only if a large sale is plausible, and it is not here — the QSBS gain exclusion is worth nothing on a $4M sale relative to the double-taxation cost of the intervening years of distributions. Stay pass-through. That is the single highest-value entity decision in this scenario and it was made correctly at incorporation.

Example 2 — Seed-funded AI startup, $500K ARR

Situation. Delaware C-corp. Raised $2.5M on post-money SAFEs and a $6M priced seed at $30M post. Total preference $8.5M, 1x non-participating. Two founders own 48% combined; option pool 15%; investors ~37%. $500K ARR growing 15% month over month off a small base. 18 months of runway. Lead investor is a $250M fund with two board seats out of five.

Step 1: preference $8.5M. Blocking rights held by preferred majority. The lead fund needs a $250M+ outcome to matter to its portfolio; a $25M sale is immaterial to them and life-changing for the founders. This divergence is the defining feature of the situation and it should be discussed openly now, not in a deal room. Step 2: early software, high growth, AI category. Natural: strategic sale, acqui-hire, future Series A and beyond. Step 3: founders have modest savings; Numbers of ~$5M each. At 48% combined ownership, clearing $5M each requires a sale of roughly $22M — and above about $23M the preferred is better off converting, so the preference stops binding and the founders' 48% applies to the whole price. Call it a $22–27M exit at current ownership, pre- and post-tax, and considerably more after a Series A's dilution. Note how much lower that threshold is than founders in this position usually assume: a non-participating 1x preference stops mattering surprisingly early. Step 4: high appetite, early career, willing to work for an acquirer for two years.

Available: raise a Series A and continue; sell now to a strategic (a live option in the 2026 AI market, where acquirer appetite for capability is unusually high); acqui-hire; wind down with ~$5M returned.

[Analysis] Recommendation: the realistic choice is between raising and continuing and selling into the current AI acquisition window. The window argument is real — 2025–26 pricing for AI capability is anomalous and anomalies close. Take a serious meeting with two or three plausible acquirers before raising, precisely because raising forecloses the small-sale branch. Then decide with information rather than assumption.

Critical action regardless: ask the lead investor directly, in writing, what size of exit they would support, and get the answer before the Series A. If the answer is "nothing below $200M," that is a fact about your available menu and you should know it now. Also: negotiate a management carve-out policy at the Series A, not at the exit. A board-approved carve-out plan of 8–12% of exit proceeds for management, established when everyone is optimistic, is dramatically easier to obtain than the same conversation during a distressed sale.

Example 3 — Profitable services business, $5M revenue

Situation. Digital agency, $5M revenue, 18% owner earnings margin (~$900K), 22 employees, S-corp, no outside capital. Two owners, 60/40. Largest client is 31% of revenue. Owners are 52 and 49. Numbers: ~$3M each.

Step 1: no cap table constraint. Step 2: services. Natural: owner-operator, MBO, ESOP, small strategic sale. Step 3: distributions of ~$540K/$360K per year; the business itself is worth perhaps 3–5x owner earnings, so $2.7M–$4.5M — and the 31% client concentration will push it to the bottom of that range or below, because a buyer prices that as a single point of failure. Step 4: both want to work 3–5 more years, then stop.

Available: continue and take distributions; MBO to the senior team; ESOP; sale to a larger agency or holding company.

[Analysis] Recommendation: the binding constraint is client concentration, and it is worth more than anything else they could do. Reducing the largest client below 15% of revenue over three years plausibly moves the multiple from ~3x to ~5x — on $900K of earnings that is $1.8M of created value, roughly equal to two full years of the owners' combined distributions — and unlike the distributions, it is permanent. Do that first.

Then, with a three-to-five year horizon and a genuine management team, run a real comparison between an MBO (continuity, deferred payment, moderate price), an ESOP (Section 1042 deferral if they convert to C-corp and sell 30%+, S-corp exemption if structured at 100%, staff continuity, fair-value price, $150K+ of setup cost), and a strategic sale (highest price, least continuity, likely earn-out — which per A5 should be discounted to ~20 cents on the dollar). For a $4M business, the ESOP setup cost is a meaningful fraction of value and the analysis is close; for a $10M business it usually is not. Get a feasibility study before falling in love with the ESOP.

Example 4 — VC-backed, $15M ARR, real preference stack

Situation. Raised $52M across seed, A and B. Total preference $52M, 1x non-participating, Series B senior. $15M ARR growing 22% year over year, burning $600K/month, 14 months of runway. Founders own 16% combined; pool 16%; preferred 68%. Board: two founders, two investors, one independent. NRR 104%. Not profitable.

Step 1: preference $52M against a business a disciplined buyer will value at, say, 3–5x ARR = $45M–$75M given 22% growth and 104% NRR in a market where the SEG public SaaS index sits at 4.8x revenue. The company may be worth less than its preference stack. This is the central fact. Step 2: growth-stage software with decelerating growth. Natural: strategic sale, PE, recapitalization. Step 3: at a $60M sale, preferred takes the full $52M; common splits $8M; founders' 16% of a 32% common base gets roughly $4M total, ~$2M each. At $45M, common gets nothing. Step 4: founders are tired; growth is decelerating; the Series C market for 22%-growth, unprofitable SaaS in 2026 is thin.

Available: (a) cut burn hard and reach profitability, which converts the company into a PE-recap candidate in 24–36 months; (b) sell now to a strategic or PE buyer, with a negotiated management carve-out; (c) raise a structured round, which adds preference and makes everything worse for common; (d) continue and become a zombie.

[Analysis] Recommendation: (a) and (b), run in parallel, and negotiate the management carve-out before starting a process, not during one. The arithmetic that drives this: reaching profitability at $15M ARR makes the company sellable to the PE buyers who accounted for 58% of 2025 SaaS M&A (SEG), and a profitable $15M-ARR business with 22% growth and improving margins is a materially different asset from an unprofitable one — plausibly 6–9x EBITDA rather than a distressed revenue multiple. It also removes the fundraising dependency that currently gives the preferred all the leverage.

Option (c) is the trap. A structured round at this point adds seniority and preference on top of $52M and converts a bad-but-survivable common position into a certainly worthless one. Chapter 7 §B13 covers the mechanics; the strategic point is that a structured round is how a $2M-each outcome becomes a $0 outcome.

What the founders should have done three years earlier: raised less at the Series B, or raised the same amount at a lower valuation with a smaller preference, or taken the $40M acquisition offer that almost certainly existed at $8M ARR. All three were available then. None are now.

PART C — HOW THE ENDGAME CONSTRAINS UPSTREAM DECISIONS

This is the practical payoff. Each item below is cheap or free early and expensive or impossible later.

C1. Entity and jurisdiction

The trade-off: a Delaware C-corporation is required for institutional venture investment, enables Section 1202 QSBS treatment on a sale (see Chapter 11 §10 for the current rules including the 2025 changes), and supports multiple share classes — but taxes distributions twice, making it a poor vehicle for a cash-distributing business. A pass-through (LLC, or S-corp election) is far better for distributions and for the QBI deduction, but is effectively uninvestable for institutional venture and forecloses QSBS.

[Analysis] The rule: choose the entity that matches your most likely endgame, and be aware that converting LLC → C-corp is usually straightforward, while C-corp → LLC is a taxable liquidation and is often prohibitively expensive. If genuinely undecided and the business may produce cash before it produces an exit, starting as an LLC preserves more optionality than starting as a C-corp — with the significant caveat that the QSBS five-year holding clock does not start until you convert, and that some investors will not engage with an LLC at all. This is a decision to make with a CPA, once, deliberately. Chapter 11 §1 has the mechanics.

C2. Whether to take venture money at all

[Analysis] Venture capital is not a funding source; it is a commitment to a specific class of outcomes. Taking it converts "sell for $20M and keep most of it" from a good outcome into a failure, and removes owner-operator, ESOP, steward ownership and marketplace sale from the menu more or less permanently.

The honest test: is there a plausible path where this company is worth $500M+, and do I want that? If the answer to either half is no, venture capital is the wrong instrument regardless of how available it is. Chapter 7 §B17 has the full comparison.

C3. Investor selection

Beyond price and brand, three things determine your future menu:

1. Fund size, which determines what exit size is acceptable. A $40M fund can be thrilled by a $25M exit. A $900M fund cannot. This is the single most predictive variable for whether your investors will support a small sale.
2. Fund vintage and life. A fund in year 8 wants liquidity. A fund in year 2 wants you to keep going.
3. Behaviour in bad situations. Ask founders of the fund's companies that did not work — not the successes. This is the highest-signal reference call available and almost nobody makes it.

C4. Board composition

[Analysis] Board control is the mechanism through which every other constraint operates. Practical guidance: keep founder-plus-independent at or above investor seats for as long as possible; choose the independent director jointly and choose an operator rather than another investor; and understand that a 2-2-1 board with an independent aligned with your investors is an investor-controlled board. Chapter 7 §B10 has the detail.

C5. Option pool design

Pool size dilutes founders (it is almost always created pre-money), and pool design determines whether employees get anything in a small exit. Two specifics with outsized effect:

• Extended post-termination exercise windows (e.g. 7–10 years rather than 90 days) let departed employees retain value. This costs the company nothing at grant and matters enormously at exit.
• A written management/employee carve-out policy, adopted by the board early, that reserves a percentage of exit proceeds for employees ahead of the preference waterfall. Negotiate it when times are good. In a distressed sale you will have no leverage and the people who built the thing will get nothing.

C6. Customer concentration

[Analysis] The most reliably value-destroying, most fixable problem in small-company M&A. Any single customer above ~20% of revenue causes a buyer to discount the multiple, structure more of the price as contingent, or walk. Above ~35% it can make the business unsaleable at any reasonable price. Measure it quarterly; treat 20% as a hard ceiling and act on it two years before you need to.

C7. Key-person dependency

The gap between the 2.7x cash-flow multiple of an owner-dependent main-street business (BizBuySell Q1 2026) and the 6.2x EBITDA paid for businesses with management teams (Stanford search fund data) is, to a first approximation, the price of being indispensable. [Analysis]

What to do, starting years early: document processes; put customer relationships in the company's name and CRM; build a second layer of decision-making; take a genuine two-week holiday without access and see what breaks; and fix whatever broke. Founders resist this because indispensability feels like security. It is the opposite.

C8. Code and IP hygiene

[Analysis] Diligence kills more small deals than price does. The recurring, entirely preventable killers:

• Unassigned IP. Every founder, employee and contractor must have a signed, present-tense assignment ("hereby assigns"). Chapter 11 §2 is explicit about this and it is the most common early mistake in the field.
• Open-source license contamination. Copyleft code in a proprietary product is a deal-stopper. Maintain an SBOM and run license scanning from early on. Chapter 11 §6.
• Unclear trademark and domain ownership, especially assets registered in a founder's personal name.
• Undocumented third-party dependencies and data rights — increasingly the central diligence question for AI companies, where the provenance of training data and the license terms of model weights are now standard diligence items.
• Customer contracts without assignability, or with change-of-control termination rights, which hand every major customer a veto over your sale.

C9. Financial record-keeping

[Analysis] The cheapest multiple expansion available. Accrual-basis books from day one; a clean chart of accounts; personal expenses out of the business; documented revenue recognition; monthly close; and, once you cross roughly $3M of revenue, a reviewed or audited set of financials. A buyer's quality-of-earnings process will find everything. Finding it yourself first, two years earlier, is worth a turn of multiple and several months of process.

What this chapter cannot tell you

Three honest limitations.

First, the base rates are worse than they look, and I cannot tell you how much worse. Every dataset here is built on transactions that happened. The denominator — companies that tried and did not transact — is not observable. When I write "median SaaS acquisition multiple 3.9x," the honest sentence is "3.9x among businesses that listed on one marketplace and found a buyer," and I do not know what fraction of listings that is.

Second, multiples are the most perishable content in this chapter. The pricing figures reflect September 2026. The structural arguments — that preference stacks determine founder proceeds, that earn-outs pay a fraction of face, that key-person dependency costs you multiple, that the endgame is chosen at incorporation — will still be true when the numbers are not.

Third, this is not advice about your company. Several structures described here (ESOP formation, purpose trusts, Section 1042 rollovers, PE rollover equity, entity conversion, generational transfer) have outcomes determined almost entirely by execution details this chapter deliberately does not cover. Get a qualified professional for: entity choice and conversion; any ESOP; any trust structure; rollover equity and shareholders' agreements; management buyouts; estate and gift planning; and the tax structure of any sale. Chapter 11 §22 lists the moments that genuinely require one, and every item in that list applies here.

The one-sentence version

Decide where you are trying to end up before you sign your first financing document, because that document — more than your product, your market, or your execution — determines which endings remain available to you.

Sources

All URLs accessed September 15, 2026 unless otherwise noted.

M&A deal terms, earn-outs and outcomes

• SRS Acquiom — M&A Undisclosed Liability Claims and Earnout Achievement
• SRS Acquiom — Earnout and Milestone Trends in Private-Target M&A Deals
• SRS Acquiom — M&A Trends: 2026 Deal Terms Study
• American Bar Association — Announcing the 2025 Private Target Mergers & Acquisitions Deal Points Study
• Bates, Neyland & Wang — Financing Acquisitions with Earnouts (Journal of Accounting and Economics)
• Thompson Coburn — Lessons from Recent Decisions on Earnout Disputes

Venture exits, deal terms and market data

• PitchBook-NVCA Venture Monitor, Q3 2025
• Cooley — Q2 2026 Venture Financing Report
• Cooley — Q4 2025 Venture Financing Report
• Carta — State of Private Markets: 2025 in Review
• Software Equity Group — 2026 Annual SaaS Report

IPOs, direct listings and SPACs

• Jay R. Ritter — IPO Statistics (University of Florida)
• Jay R. Ritter — VC-backed IPOs
• Jay R. Ritter — Median Sales of IPO Firms
• Jay R. Ritter — SPAC IPOs and de-SPAC returns
• Jay R. Ritter — Direct Listings in the U.S., 2018–2026
• Jay R. Ritter — IPO Data index
• SEC — Initial Public Offerings (IPOs) statistics
• PwC — Considering an IPO? First, understand the costs
• Foley & Lardner — SPAC 4.0: From Spectacular Failures to a Disciplined Renaissance

Private equity

• Bain & Company — Global Private Equity Report 2026 (press release)
• Bain & Company — Global Private Equity Report 2026 (full report)
• Chesapeake Corporate Advisors — Demystifying the "Second Bite of the Apple"

Small-business acquisition market, search funds and marketplaces

• Acquire.com — Biannual Acquisition Multiples Report, January 2026
• BizBuySell — Q1 2026 Insight Report
• BizBuySell — Q4 2025 Insight Report
• ExitBid — Empire Flippers Review 2026: Fees & Valuation Multiples
• CapitalPad — Search Fund Statistics (2026 Stanford GSB Search Fund Study data)
• Search Funds News — 2026 Search Fund Study (Stanford GSB)
• CapitalPad — Private Equity Holding Period Statistics

Acqui-hires

• CRV — Acquihire: How to Protect Your Team in an Acquisition
• Heavybit — Acqui-Hires in the Age of AI

Employee ownership and ESOPs

• NCEO — Employee Ownership by the Numbers
• NCEO — Understanding Tax Benefits and Contribution Limits for ESOPs
• ESOP Association — S Corporation ESOPs
• NCEO — Trends in New ESOP Creation

Steward ownership and purpose trusts

• American Bar Association — Patagonia, Purpose Trusts, and Stewardship Trusts
• Cole Schotz — Charitable Planning Update: Yvon Chouinard Donates Patagonia Stock to Two Nonprofits
• Purpose — The Patagonia Structure in the Context of Steward-Ownership
• Columbia Journal of Law and Social Problems — Set It in Stone: Patagonia and the Evolution toward Stakeholder Governance

Leadership, succession and CEO hiring

• Leadership IQ — Executive Failure Rates
• CEO Worldwide — Small Company CEO Salary 2026: What Each Pay Tier Buys
• Family Business Consulting Group — Family Business Survival: Understanding the Statistics
• MIT Sloan Management Review — When Is an Outsider CEO a Good Choice?

Tax and entity

• SDO CPA — S-Corporation Tax Guide: Salary, QBI, Distributions
• SDO CPA — QBI Deduction 2026: Section 199A Guide for S-Corps
• CSG Partners — 1042 Rollovers: Deferring Capital Gains on ESOP Sale Proceeds

Sources consulted but treated with caution

The following were reviewed and are not relied on for any numeric claim in this chapter, because they aggregate secondary statistics without traceable methodology or are marketing content from firms with a direct interest in the transactions they describe: SEO-optimized "startup exit statistics 2026" compilations; broker and advisory blogs publishing valuation multiple tables without disclosed sample sizes; and vendor-published earn-out "benchmark" guides. Where a figure from such a source agreed with a primary source it added nothing; where it disagreed, the primary source was used.