{ "site": "venture-capital.wiki", "origin": "https://venture-capital.wiki", "description": "A machine-readable reference for venture financing: liquidation preference waterfalls with worked formulas, convertible instrument conversion math, dilution and option pool arithmetic, a clause-by-clause map of the standard term sheet, and the valuation algebra that connects them.", "reviewed": "2026-08-27", "license": "CC BY 4.0", "changes": "https://venture-capital.wiki/changes.json", "feed": "https://venture-capital.wiki/feed.xml", "sections": [ { "slug": "liquidation", "title": "Liquidation preferences", "lede": "Who gets paid, in what order, and the exact arithmetic of each variant.", "intro": "A liquidation preference determines the split of exit proceeds between preferred and common. The three structural questions are: what multiple of invested capital comes off the top, whether the preferred also shares in the remainder, and whether that participation is capped. Every named variant is a combination of those three answers.", "tables": [ { "title": "The four structures", "intro": "E is total exit proceeds, I is invested capital, x is the preference multiple, and p is the preferred holder's fully diluted ownership percentage expressed as a decimal.", "columns": [ "Structure", "Preferred receives", "Common receives" ], "rows": [ [ "Non-participating, 1x", "max(I, p*E) - holder elects preference or conversion, whichever is greater", "E less the preferred payout" ], [ "Non-participating, x multiple", "max(x*I, p*E)", "E less the preferred payout" ], [ "Full participating", "I + p*(E - I) - preference first, then shares in the remainder", "(1 - p)*(E - I)" ], [ "Capped participating", "min(I + p*(E - I), cap*I), with the right to convert instead if that pays more", "E less the preferred payout" ] ] }, { "title": "Worked example - 10,000,000 invested for 20 percent", "intro": "I = 10,000,000. p = 0.20. Multiple = 1x. Participation cap where applicable = 3x. Figures are preferred proceeds.", "columns": [ "Exit proceeds E", "Non-participating 1x", "Full participating", "Capped at 3x" ], "rows": [ [ "5,000,000", "5,000,000", "5,000,000", "5,000,000" ], [ "10,000,000", "10,000,000", "10,000,000", "10,000,000" ], [ "25,000,000", "10,000,000", "13,000,000", "13,000,000" ], [ "50,000,000", "10,000,000", "18,000,000", "18,000,000" ], [ "100,000,000", "20,000,000", "28,000,000", "28,000,000" ], [ "200,000,000", "40,000,000", "48,000,000", "40,000,000" ], [ "500,000,000", "100,000,000", "108,000,000", "100,000,000" ] ] }, { "title": "Three-series stack - pari passu versus seniority-ordered", "intro": "Seed 2,000,000 at 1x, Series A 8,000,000 at 1x, Series B 20,000,000 at 1x. Aggregate preference is 30,000,000. Under pari passu each series takes its preference amount multiplied by E/30,000,000. Under seniority the newest series is paid in full first. Figures are proceeds to Seed / Series A / Series B.", "columns": [ "Exit proceeds E", "Pari passu", "Seniority-ordered, newest first" ], "rows": [ [ "15,000,000", "1,000,000 / 4,000,000 / 10,000,000", "0 / 0 / 15,000,000" ], [ "18,000,000", "1,200,000 / 4,800,000 / 12,000,000", "0 / 0 / 18,000,000" ], [ "24,000,000", "1,600,000 / 6,400,000 / 16,000,000", "0 / 4,000,000 / 20,000,000" ], [ "27,000,000", "1,800,000 / 7,200,000 / 18,000,000", "0 / 7,000,000 / 20,000,000" ], [ "30,000,000", "2,000,000 / 8,000,000 / 20,000,000", "2,000,000 / 8,000,000 / 20,000,000" ] ] }, { "title": "Accrued dividends and the preference amount", "intro": "I = 10,000,000 at 1x with a cumulative dividend at rate d = 0.08. The rate is chosen for arithmetic clarity and is not offered as a market level. Simple accrual adds I*d each year; compounding accrual multiplies by (1 + d). The final column is the conversion indifference point E* = preference / p at p = 0.20, using the simple figure.", "columns": [ "Years accrued t", "Preference, simple accrual", "Preference, compounding", "E* at p = 0.20, simple" ], "rows": [ [ "0", "10,000,000", "10,000,000", "50,000,000" ], [ "1", "10,800,000", "10,800,000", "54,000,000" ], [ "2", "11,600,000", "11,664,000", "58,000,000" ], [ "3", "12,400,000", "12,597,120", "62,000,000" ], [ "5", "14,000,000", "14,693,281", "70,000,000" ], [ "7", "15,600,000", "17,138,243", "78,000,000" ] ] }, { "title": "Election flip points in a four-series stack", "intro": "The company from the share ledger after four priced rounds: Seed 2,000,000, Series A 8,000,000, Series B 20,000,000 and Series C 30,000,000, every series at 1x non-participating, ranking pari passu. Aggregate preference T = 60,000,000. Ownership p is taken from the ledger, where the four series hold 9.600, 16.000, 16.000 and 20.000 percent and common plus the option pool holds 38.400 percent. The third column is the flip point each series would have if it were the only preferred outstanding; the fourth is the exit value at which it actually elects conversion once the other series' elections are solved with it.", "columns": [ "Series", "Preference x*I", "Ownership p", "Flip point solved alone", "Flip point solved jointly", "Understatement" ], "rows": [ [ "Seed", "2,000,000", "9.600", "20,833,333", "68,000,000", "47,166,667" ], [ "Series A", "8,000,000", "16.000", "50,000,000", "82,000,000", "32,000,000" ], [ "Series B", "20,000,000", "16.000", "125,000,000", "130,000,000", "5,000,000" ], [ "Series C", "30,000,000", "20.000", "150,000,000", "150,000,000", "0" ] ] }, { "title": "Four-series waterfall across exit values", "intro": "The same all-1x pari passu stack. The last column records each series' election in ledger order Seed, A, B, C: P means it took its preference, C means it converted. Every row sums to E.", "columns": [ "Exit proceeds E", "Seed", "Series A", "Series B", "Series C", "Common and pool", "Elections" ], "rows": [ [ "20,000,000", "666,667", "2,666,667", "6,666,667", "10,000,000", "0", "PPPP" ], [ "40,000,000", "1,333,333", "5,333,333", "13,333,333", "20,000,000", "0", "PPPP" ], [ "60,000,000", "2,000,000", "8,000,000", "20,000,000", "30,000,000", "0", "PPPP" ], [ "68,000,000", "2,000,000", "8,000,000", "20,000,000", "30,000,000", "8,000,000", "PPPP" ], [ "82,000,000", "4,800,000", "8,000,000", "20,000,000", "30,000,000", "19,200,000", "CPPP" ], [ "90,000,000", "6,000,000", "10,000,000", "20,000,000", "30,000,000", "24,000,000", "CCPP" ], [ "110,000,000", "9,000,000", "15,000,000", "20,000,000", "30,000,000", "36,000,000", "CCPP" ], [ "130,000,000", "12,000,000", "20,000,000", "20,000,000", "30,000,000", "48,000,000", "CCPP" ], [ "150,000,000", "14,400,000", "24,000,000", "24,000,000", "30,000,000", "57,600,000", "CCCP" ], [ "250,000,000", "24,000,000", "40,000,000", "40,000,000", "50,000,000", "96,000,000", "CCCC" ] ] }, { "title": "Accrued dividends by vintage across a four-series stack", "intro": "A cumulative dividend at d = 0.08 accruing from each round's closing. The rate is chosen for arithmetic clarity and is not offered as a market level. Series C closed at the measurement date and has accrued nothing. Simple accrual adds x*I*d each year; compounding multiplies by (1 + d) each year. Figures are rounded to the nearest unit and the aggregate is computed before rounding, so the rounded components differ from the aggregate by one unit.", "columns": [ "Series", "Invested I", "Years accrued t", "Preference, simple", "Preference, compounding", "Difference" ], "rows": [ [ "Seed", "2,000,000", "6", "2,960,000", "3,173,749", "213,749" ], [ "Series A", "8,000,000", "4", "10,560,000", "10,883,912", "323,912" ], [ "Series B", "20,000,000", "2", "23,200,000", "23,328,000", "128,000" ], [ "Series C", "30,000,000", "0", "30,000,000", "30,000,000", "0" ], [ "Aggregate preference T", "60,000,000", "-", "66,720,000", "67,385,660", "665,660" ] ] }, { "title": "Who absorbs each deduction from a 90,000,000 exit", "intro": "The all-1x four-series stack on a 90,000,000 exit. Each deduction is taken before the waterfall runs. The last column is the fraction of the deduction borne by common and the option pool, which hold 38.400 percent of the company.", "columns": [ "Deduction", "Amount", "Proceeds to the stack", "Seed", "Series A", "Series B", "Series C", "Common and pool", "Borne by common" ], "rows": [ [ "None - the whole 90,000,000 reaches the stack", "0", "90,000,000", "6,000,000", "10,000,000", "20,000,000", "30,000,000", "24,000,000", "-" ], [ "Transaction expenses of 3,000,000", "3,000,000", "87,000,000", "5,550,000", "9,250,000", "20,000,000", "30,000,000", "22,200,000", "60.000" ], [ "Escrow holdback of 9,000,000, closing distribution only", "9,000,000", "81,000,000", "4,600,000", "8,000,000", "20,000,000", "30,000,000", "18,400,000", "62.222" ], [ "Carve-out at 10 percent of E", "9,000,000", "81,000,000", "4,600,000", "8,000,000", "20,000,000", "30,000,000", "18,400,000", "62.222" ], [ "Fixed carve-out pool of 6,000,000", "6,000,000", "84,000,000", "5,100,000", "8,500,000", "20,000,000", "30,000,000", "20,400,000", "60.000" ], [ "Carve-out at 10 percent of proceeds above T = 60,000,000", "3,000,000", "87,000,000", "5,550,000", "9,250,000", "20,000,000", "30,000,000", "22,200,000", "60.000" ] ] }, { "title": "The same stack paid in strict seniority, newest series first", "intro": "Series C ranks senior to Series B, Series B to Series A, and Series A to Seed. S_j is the aggregate preference ranking ahead of series j: S_Seed = 58,000,000, S_A = 50,000,000, S_B = 30,000,000 and S_C = 0. A series receives nothing until E exceeds its S_j, and is paid in full once E reaches S_j plus its own preference.", "columns": [ "Exit proceeds E", "Seed", "Series A", "Series B", "Series C", "Common and pool" ], "rows": [ [ "20,000,000", "0", "0", "0", "20,000,000", "0" ], [ "30,000,000", "0", "0", "0", "30,000,000", "0" ], [ "40,000,000", "0", "0", "10,000,000", "30,000,000", "0" ], [ "50,000,000", "0", "0", "20,000,000", "30,000,000", "0" ], [ "58,000,000", "0", "8,000,000", "20,000,000", "30,000,000", "0" ], [ "60,000,000", "2,000,000", "8,000,000", "20,000,000", "30,000,000", "0" ], [ "68,000,000", "2,000,000", "8,000,000", "20,000,000", "30,000,000", "8,000,000" ], [ "90,000,000", "6,000,000", "10,000,000", "20,000,000", "30,000,000", "24,000,000" ] ] } ], "entries": [ { "id": "conversion-indifference-point", "term": "Conversion indifference point", "definition": "The exit value at which a non-participating preferred holder is indifferent between taking the preference and converting to common. Below it they take the preference; above it they convert.", "formula": "E* = x*I / p", "fields": [ { "label": "Worked", "value": "1x on 10,000,000 at 20 percent: E* = 10,000,000 / 0.20 = 50,000,000", "mono": true }, { "label": "Interpretation", "value": "Between I and E* the preferred is protected and the common is impaired", "mono": false } ], "notes": [ "The dead zone between the invested amount and the indifference point is where founder and employee equity is worth materially less than the headline ownership percentage implies.", "Raising at a high valuation with a high multiple widens that dead zone. A 2x preference doubles E* and doubles the range of outcomes in which common receives little." ] }, { "id": "cap-breakeven", "term": "Where a participation cap stops binding", "definition": "A capped participating holder reverts to converting when straight conversion pays more than the cap. The cap therefore only binds over a finite band of exit values.", "formula": "Cap binds while cap*I > p*E, i.e. for E < cap*I/p", "notes": [ "With a 3x cap on 10,000,000 at 20 percent, the cap ceases to bind above E = 150,000,000, where conversion pays 30,000,000.", "In the worked table above, the capped column tracks full participation until 200,000,000 and then reverts to the conversion value." ] }, { "id": "stacking", "term": "Stacked versus pari passu preferences", "definition": "When multiple preferred series exist, the seniority rule determines the order of payment in a proceeds shortfall.", "fields": [ { "label": "Pari passu", "value": "All series rank equally and share any shortfall pro rata to their preference amounts", "mono": false }, { "label": "Stacked (senior)", "value": "Later series are paid in full before earlier series receive anything", "mono": false }, { "label": "Tiered", "value": "Groups of series rank together, with seniority between groups", "mono": false } ], "notes": [ "Stacking matters only when proceeds are insufficient to cover all preferences. Above that level the distinction is irrelevant.", "Because it only bites in bad outcomes, stacking is frequently conceded in negotiation and then determines the entire result in the outcome that actually occurs." ] }, { "id": "stack-waterfall-algebra", "term": "The algebra of a multi-series stack", "definition": "With more than one preferred series the payout is no longer a single expression. Each series has its own preference amount, and the seniority rule decides whether a shortfall is shared or absorbed in order. Both rules reduce to one line of arithmetic.", "formula": "Pari passu: R_j = x_j*I_j * min(1, E/T) where T = sum of x_j*I_j. Seniority-ordered: R_j = min(x_j*I_j, max(0, E - S_j)) where S_j is the sum of preference amounts senior to j", "fields": [ { "label": "Worked", "value": "Stack 2,000,000 / 8,000,000 / 20,000,000 all at 1x, so T = 30,000,000. At E = 18,000,000 the pari passu ratio is 0.60: Seed 1,200,000, Series A 4,800,000, Series B 12,000,000", "mono": true }, { "label": "Worked, seniority", "value": "Same stack at E = 24,000,000: B takes min(20,000,000, 24,000,000) = 20,000,000; A takes min(8,000,000, 4,000,000) = 4,000,000; Seed takes min(2,000,000, max(0, -4,000,000)) = 0", "mono": true }, { "label": "Check", "value": "Both rules distribute exactly min(E, T); above T the two are identical", "mono": false } ], "notes": [ "The seniority rule is a single word in the charter and it decides the entire result in a shortfall. Model both before agreeing to either.", "Pari passu shares pro rata to preference amounts, not to invested capital. A series carrying a multiple therefore recovers a larger fraction of its money than a 1x series ranking alongside it, without any seniority.", "A stack built round by round accumulates preference faster than it accumulates valuation. Track T against the exit values actually being discussed, not against the last post-money." ] }, { "id": "seniority-zero-threshold", "term": "When seniority wipes a series out completely", "definition": "Under seniority-ordered payment a series receives nothing at all once proceeds fail to cover the preferences ranking ahead of it. Under pari passu the same series always receives something as long as proceeds are positive. The threshold is exact and easy to compute.", "formula": "Series j receives zero under seniority whenever E <= S_j, the aggregate preference senior to it. Under pari passu it receives x_j*I_j * E/T for any E > 0", "fields": [ { "label": "Worked", "value": "Seed is junior to 8,000,000 + 20,000,000 = 28,000,000, so Seed receives nothing under seniority for any E <= 28,000,000", "mono": true }, { "label": "Same point, pari passu", "value": "At E = 28,000,000 Seed receives 2,000,000 * 28,000,000/30,000,000 = 1,866,667", "mono": true } ], "notes": [ "The seed investor who agreed to junior ranking has an all-or-nothing claim: full recovery above 30,000,000 and nothing below 28,000,000, with a 2,000,000 band in between.", "This is why later investors ask for seniority and earlier investors often grant it: the earlier investor is comparing outcomes that already look poor, while the later investor is buying protection in the region where its money is actually at risk.", "A pay-to-play or recapitalisation that converts the junior series to common produces the same result as seniority, by a different route. Check whether the charter already achieves what the new term asks for." ] }, { "id": "capped-participation-regimes", "term": "The three regimes of a capped participating series", "definition": "A capped participating preference behaves as three different instruments above the point at which it has recovered its money: it participates in the residual, then plateaus at the cap with proceeds that do not move at all, then converts to common. Both boundaries are closed-form.", "formula": "Participation reaches the cap at E_cap = I*(1 + (cap - 1)/p). Conversion overtakes the cap at E_conv = cap*I/p. Between them proceeds are constant at cap*I", "fields": [ { "label": "Worked", "value": "I = 10,000,000, cap = 3, p = 0.20: E_cap = 10,000,000*(1 + 2/0.20) = 110,000,000; E_conv = 3*10,000,000/0.20 = 150,000,000", "mono": true }, { "label": "Check at E_cap", "value": "I + p*(E - I) = 10,000,000 + 0.20*100,000,000 = 30,000,000 = cap*I", "mono": true }, { "label": "Check at E_conv", "value": "p*E = 0.20*150,000,000 = 30,000,000 = cap*I", "mono": true }, { "label": "Plateau width", "value": "E_conv - E_cap = I*(1 - p)/p = 10,000,000*0.80/0.20 = 40,000,000 of exit value over which the holder's proceeds do not change", "mono": true } ], "notes": [ "Inside the plateau the holder is economically indifferent to price. A 40,000,000 improvement in the sale price is worth nothing to it, which removes its incentive to push the buyer and can quietly misalign it from common in a live negotiation.", "The cap is usually presented as a founder-friendly limit on participation. It is, but it also creates a range where the holder stops caring about the outcome. Check where that range sits relative to the exits being modelled.", "The plateau width simplifies to I*(1 - p)/p and does not depend on the cap at all. At I = 10,000,000 and p = 0.20 a 2x cap, a 3x cap and a 4x cap all produce a 40,000,000 flat band; the cap only decides where the band starts. Ownership is what narrows it - at p = 0.50 the same band is 10,000,000 wide." ] }, { "id": "participation-vs-multiple", "term": "Full participation at 1x versus a higher non-participating multiple", "definition": "These are the two standard ways to give a preferred holder more than its money back, and they cross exactly once. Below the crossover the multiple pays more; above it participation pays more.", "formula": "1x full participation equals an N x non-participating preference at E_eq = I*(1 + (N - 1)/p)", "fields": [ { "label": "Worked", "value": "I = 10,000,000, p = 0.20, N = 2: E_eq = 10,000,000*(1 + 1/0.20) = 60,000,000", "mono": true }, { "label": "Below the crossover", "value": "At E = 40,000,000: participating pays 10,000,000 + 0.20*30,000,000 = 16,000,000; 2x non-participating pays max(20,000,000, 8,000,000) = 20,000,000", "mono": true }, { "label": "At the crossover", "value": "At E = 60,000,000 both pay 20,000,000", "mono": true }, { "label": "Above the crossover", "value": "At E = 100,000,000: participating pays 28,000,000; 2x non-participating pays max(20,000,000, 20,000,000) = 20,000,000", "mono": true } ], "notes": [ "An investor asking for participation and an investor asking for a 2x are not asking for the same thing, and which is worse for common depends entirely on the exit range. Compute E_eq before conceding either.", "A higher multiple is bounded: it can never pay more than N*I unless conversion is better. Participation is unbounded. Founders trading a 2x away for participation are trading a known cost for an open-ended one.", "The comparison is only valid at a fixed p. Any change to the fully diluted denominator moves E_eq, so run it on the post-round cap table, not the pre-round one." ] }, { "id": "conversion-election-interdependence", "term": "Conversion elections across series are interdependent", "definition": "Each series elects preference or conversion to maximise its own proceeds, but a converting series joins the residual pool and changes what every other converting series receives. The elections must be solved together, not one at a time.", "formula": "A converting series j receives s_j / (s_common + sum of s_k over converting k) * (E - sum of preference amounts of non-converting series)", "fields": [ { "label": "Setup", "value": "Common 5,000,000 shares; Seed 1,000,000 shares carrying 2x on 2,000,000 (preference 4,000,000); Series A 4,000,000 shares carrying 1x on 8,000,000. E = 25,000,000", "mono": true }, { "label": "Worked", "value": "Seed takes its 4,000,000 preference (conversion would pay 1,000,000/10,000,000 * 25,000,000 = 2,500,000). Series A converts: 4,000,000/9,000,000 * 21,000,000 = 9,333,333, against an 8,000,000 preference", "mono": true }, { "label": "Result", "value": "Seed 4,000,000; Series A 9,333,333; common 5,000,000/9,000,000 * 21,000,000 = 11,666,667. Total 25,000,000", "mono": true } ], "notes": [ "Solving series by series in charter order gives the wrong answer, because an early series' election changes the residual that a later series is comparing against. Iterate until no series wants to switch.", "The equilibrium is a mixed one here: the series with the multiple takes cash, the series with the larger ownership converts. Mixed outcomes are the normal case, not the exception.", "A waterfall model that hard-codes each series' election rather than solving it will be wrong precisely in the exit range where the answer matters. Test it by sweeping E and checking that every series' proceeds curve is non-decreasing." ] }, { "id": "management-carve-out", "term": "Management carve-out and how it reorders the waterfall", "definition": "A carve-out reserves a slice of proceeds for management ahead of the preferred stack. Because it is paid before the most senior preference, it inverts the payment order the charter otherwise establishes.", "formula": "Proceeds available to the stack = E - carve-out. Each series then receives x_j*I_j * min(1, (E - carve-out)/T)", "fields": [ { "label": "Worked", "value": "E = 18,000,000 with a carve-out of 10 percent of E = 1,800,000. Available to the stack = 16,200,000 against T = 30,000,000, a ratio of 0.54", "mono": true }, { "label": "Result", "value": "Seed 1,080,000; Series A 4,320,000; Series B 10,800,000; management 1,800,000", "mono": true }, { "label": "Without the carve-out", "value": "Seed 1,200,000; Series A 4,800,000; Series B 12,000,000", "mono": true } ], "notes": [ "A carve-out defined as a percentage of E is a very different instrument from one defined as a percentage of proceeds above the aggregate preference. The first pays in every outcome; the second pays only in outcomes where common was already getting paid. Read which one the plan says.", "Carve-outs are usually documented as a cash bonus plan rather than as equity, so they never appear on the cap table. A waterfall model built from the cap table alone will silently omit the most senior claim in the structure.", "The carve-out exists because a preference overhang leaves management with no reason to run a sale. Investors approve it for the same reason they resist it: it is a transfer from them to the people who have to execute the transaction." ] }, { "id": "preference-overhang-breakeven", "term": "Preference overhang and the first dollar to common", "definition": "With non-participating preferred, common receives nothing until proceeds clear the entire aggregate preference plus anything senior to it. That break-even, not the last post-money valuation, is the number that determines whether employee equity is worth anything.", "formula": "With no carve-out, common's break-even is E = T. With a carve-out set at a fraction c of E, it is E = T/(1 - c)", "fields": [ { "label": "Worked", "value": "T = 30,000,000 and no carve-out: common receives its first dollar above E = 30,000,000", "mono": true }, { "label": "Worked, with a carve-out", "value": "T = 30,000,000 with c = 0.10: E = 30,000,000/0.90 = 33,333,333. Check: 0.10*33,333,333 = 3,333,333 to management, leaving exactly 30,000,000 for the stack", "mono": true } ], "notes": [ "T is a sum of historical cheques and is therefore known exactly. The break-even is one of the few numbers in venture financing that involves no estimate at all, and it is rarely the number quoted to employees.", "Adding a participating series does not move the break-even, but it does reduce common's share of every dollar above it. The break-even and the slope are separate questions.", "If T has grown past the exit values the company can realistically reach, no amount of further operating progress makes common valuable. That is the condition a recapitalisation exists to reset." ] }, { "id": "pari-passu-mixed-multiples", "term": "Pari passu with unequal multiples", "definition": "Pari passu ranking is often described as equal treatment. It is equal ranking of preference amounts, which is not the same thing: a series carrying a multiple recovers a larger fraction of its invested capital than a 1x series ranking alongside it.", "formula": "R_j = x_j*I_j * E/T for E < T, so the recovery multiple on invested capital is R_j/I_j = x_j * E/T", "fields": [ { "label": "Setup", "value": "Seed 2,000,000 at 2x (preference 4,000,000); Series A 8,000,000 at 1x; Series B 20,000,000 at 1x. T = 32,000,000", "mono": true }, { "label": "Worked", "value": "At E = 16,000,000 the ratio is 0.50: Seed 2,000,000, Series A 4,000,000, Series B 10,000,000", "mono": true }, { "label": "Recovery on capital", "value": "Seed 2,000,000/2,000,000 = 1.00x; Series B 10,000,000/20,000,000 = 0.50x", "mono": true } ], "notes": [ "A multiple buys effective priority without asking for seniority, and it does so in a term that is negotiated as an economic point rather than a control point. The later investor conceding pari passu ranking to an earlier 2x has conceded more than the word suggests.", "The arithmetic generalises: within a pari passu group, relative recovery is set entirely by the ratio of preference amounts. Invested capital never enters the calculation.", "When a stack contains mixed multiples, check T against the aggregate of actual cheques written. A 32,000,000 preference on 30,000,000 of capital is a 6.7 percent invisible increase in the overhang." ] }, { "id": "four-series-flip-points", "term": "Where each series flips election in a four-series stack", "definition": "In a single-series cap table the conversion indifference point is x*I/p. In a stack it is not, because the preferences of every series that has not converted come off the top first and the residual is shared by a smaller group. Solved jointly, every flip point moves higher or stays where it is - here by 47,166,667 for Seed, 32,000,000 for Series A, 5,000,000 for Series B and nothing for Series C, which is the most senior and therefore has nothing ranking ahead of it.", "formula": "Series j converts when s_j/(s_common + sum of s_k over converting k) * (E - sum of preference amounts of non-converting series) > x_j*I_j", "fields": [ { "label": "Setup", "value": "Seed 2,000,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, all 1x non-participating. p = 9.600, 16.000, 16.000, 20.000 percent; common and pool 38.400 percent", "mono": true }, { "label": "Solved alone", "value": "Seed 20,833,333, Series A 50,000,000, Series B 125,000,000, Series C 150,000,000", "mono": true }, { "label": "Solved jointly", "value": "Seed 68,000,000, Series A 82,000,000, Series B 130,000,000, Series C 150,000,000", "mono": true }, { "label": "Understatement", "value": "Seed 47,166,667, Series A 32,000,000, Series B 5,000,000, Series C 0", "mono": true }, { "label": "Worked, the Seed flip", "value": "Seed converting alone shares the residual with common and pool: 9.600/(38.400 + 9.600) = 0.20000. Setting 0.20000*(E - 58,000,000) = 2,000,000 gives E = 68,000,000", "mono": true }, { "label": "Worked, the Series A flip", "value": "With Seed already converted, A's residual share is 16.000/(38.400 + 9.600 + 16.000) = 0.25000. Setting 0.25000*(E - 50,000,000) = 8,000,000 gives E = 82,000,000", "mono": true }, { "label": "Check at the top", "value": "At E = 150,000,000 the first three have converted and Series C is exactly indifferent: 0.20000 * 150,000,000 = 30,000,000 = its preference", "mono": true } ], "notes": [ "The naive figure x*I/p is not a conservative estimate, it is wrong in the founder-unfriendly direction. It tells the Seed investor it is protected only up to 20,833,333 when in fact it holds a cash claim all the way to 68,000,000, and it tells common that the preferred starts sharing far earlier than it does.", "The flip points are ordered, and the order is not the order of the naive points. Which series converts first is decided by the ratio of its preference to its ownership relative to every other series, not by vintage or by cheque size.", "Each flip point is a kink in the proceeds curve for every other holder. A model that samples exit values on a round grid can miss all four kinks and still look smooth. Sample at the flip points themselves.", "The practical use is negotiation timing: between 60,000,000 and 68,000,000 the Seed investor is a cash claimant with no interest in price, and above 82,000,000 both Seed and Series A are equity holders aligned with common. Those are different counterparties in the same room." ] }, { "id": "mixed-participation-stack", "term": "A stack with mixed multiples and mixed participation", "definition": "Replacing the clean Series C with a 1.5x participating preference capped at 2x changes three things at once: the aggregate preference, the residual sharing group, and every other series' flip point. The arithmetic is unchanged; only the inputs move.", "formula": "T = sum of x_j*I_j. A participating series joins the residual pool while still holding its preference, so the residual denominator includes its shares whether it has converted or not", "fields": [ { "label": "Setup", "value": "Seed, A and B unchanged at 1x non-participating. Series C 30,000,000 at 1.5x participating, capped at 2x invested", "mono": true }, { "label": "Aggregate preference", "value": "2,000,000 + 8,000,000 + 20,000,000 + 45,000,000 = 75,000,000, against 60,000,000 for the clean stack", "mono": true }, { "label": "Flip points", "value": "Seed 87,166,667, Series A 107,000,000, Series B 160,000,000, Series C 300,000,000", "mono": true }, { "label": "Worked at E = 150,000,000", "value": "Series C 60,000,000 (capped), Seed 10,500,000, Series A 17,500,000, Series B 20,000,000, common and pool 42,000,000", "mono": true }, { "label": "Same exit, clean stack", "value": "Series C 30,000,000, Seed 14,400,000, Series A 24,000,000, Series B 24,000,000, common and pool 57,600,000", "mono": true }, { "label": "Cost to common and pool", "value": "57,600,000 less 42,000,000 = 15,600,000 at this single exit value", "mono": true } ], "notes": [ "One structured series moves the whole stack. Series B's flip point rises from 130,000,000 to 160,000,000 and Series A's from 82,000,000 to 107,000,000 purely because Series C is taking more off the top and then sharing the remainder. The earlier investors pay for the later investor's structure before common does.", "A participating series is in the residual pool in every state of the world, so it can never be squeezed out the way a non-participating series can. That is what makes participation the more valuable term at the same headline multiple.", "When comparing a structured proposal with a clean one, compute the cost at three exit values, not one. Here the 15,600,000 at 150,000,000 becomes 0 below 75,000,000 - where common gets nothing under either structure - and shrinks again far above 300,000,000 once the cap forces conversion." ] }, { "id": "capped-participation-plateau-in-a-stack", "term": "The cap plateau widens when the series sits in a stack", "definition": "For a single capped participating series the flat band over which its proceeds do not move is I*(1 - p)/p. Inside a stack the band is wider, because the senior preferences of the other series delay the point at which participation reaches the cap while the conversion crossover stays where it was.", "formula": "Participation reaches the cap where x*I + p_resid*(E - sum of other non-converting preferences) = cap*I. Conversion overtakes the cap at E = cap*I/p", "fields": [ { "label": "Setup", "value": "Series C 30,000,000 at 1.5x participating, capped at 2x, p = 20.000 percent, in the four-series stack", "mono": true }, { "label": "Cap reached", "value": "With Seed and Series A converted the residual group is 84.000 percent of the ledger and C's share of it is 20.000/84.000 = 0.238095. 45,000,000 + 0.238095*(E - 65,000,000) = 60,000,000 gives E = 128,000,000", "mono": true }, { "label": "Conversion overtakes the cap", "value": "0.20000*E = 60,000,000 gives E = 300,000,000", "mono": true }, { "label": "Plateau width", "value": "300,000,000 - 128,000,000 = 172,000,000 of exit value over which Series C receives exactly 60,000,000", "mono": true }, { "label": "Same series standing alone", "value": "I*(1 - p)/p = 30,000,000*0.80/0.20 = 120,000,000", "mono": true } ], "notes": [ "The plateau is 52,000,000 wider inside the stack than it would be alone. Over that whole range the largest and most recent investor is economically indifferent to the sale price, and it is usually the investor with the board seats and the consent rights.", "The width is not a drafting artefact that can be negotiated away by moving the cap: the cap only decides where the band starts. Narrowing the band requires changing p or removing participation.", "Check where the band sits relative to the exits actually being discussed. A plateau that starts above every credible outcome costs common nothing; one that starts inside the likely range removes the pricing incentive of the holder who controls the process." ] }, { "id": "escrow-holdback-waterfall", "term": "Escrow and holdbacks - two allocation conventions that differ by millions", "definition": "An escrow defers part of the consideration. The waterfall can be run on cumulative proceeds as each tranche is released, or the escrow can be allocated in the same proportions as the closing distribution. Both appear in real allocation schedules and they are not close to each other.", "formula": "Cumulative: each holder's share of a release equals its allocation at (closing proceeds + release) less its allocation at closing proceeds. Pro rata to closing: each holder's share equals release * its closing allocation / closing proceeds", "fields": [ { "label": "Setup", "value": "E = 90,000,000 with a 9,000,000 escrow, so 81,000,000 is distributed at closing. All-1x four-series stack", "mono": true }, { "label": "At closing, 81,000,000", "value": "Seed 4,600,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 18,400,000", "mono": true }, { "label": "Cumulative at 90,000,000", "value": "Seed 6,000,000, Series A 10,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 24,000,000", "mono": true }, { "label": "Escrow allocated cumulatively", "value": "Seed 1,400,000, Series A 2,000,000, Series B 0, Series C 0, common and pool 5,600,000", "mono": true }, { "label": "Escrow allocated pro rata to closing", "value": "Seed 511,111, Series A 888,889, Series B 2,222,222, Series C 3,333,333, common and pool 2,044,444", "mono": true }, { "label": "Difference to common and pool", "value": "3,555,556 - the cumulative convention pays common 2.74 times as much", "mono": true } ], "notes": [ "The cumulative convention is the economically correct one: the escrow is part of the merger consideration and the charter allocates total consideration, not each instalment separately. The pro rata shortcut silently pays the preferences twice, once out of the closing tranche and again out of the escrow.", "The error only appears when the closing distribution and the total sit on opposite sides of a flip point. Here 81,000,000 is below Series A's 82,000,000 flip and 90,000,000 is above it, so the shortcut denies Series A and common the entire benefit of A's conversion.", "Ask for the allocation schedule to state which convention it uses, and ask for the escrow line specifically. It is a single sentence in an exhibit and it is worth more to common than most of the terms negotiated in the term sheet.", "A holdback for a working capital true-up behaves the same way, with the added feature that it can be reduced to nothing. Model the escrow at full release and at zero release; the two waterfalls are not proportional." ], "source": "Delaware General Corporation Law section 251 (merger consideration and its allocation)" }, { "id": "transaction-expenses-priority", "term": "Transaction expenses come off the top, so common pays most of them", "definition": "Banker fees, legal fees and other transaction expenses are paid before the waterfall runs. Every holder in the residual group bears them in proportion to its share of the residual, not its share of the company - and above the aggregate preference the residual group is much smaller than the company.", "formula": "With the residual shared among a group holding fraction f of the ledger, common's share of each expense dollar is p_common/f, not p_common", "fields": [ { "label": "Setup", "value": "E = 90,000,000, all-1x four-series stack, transaction expenses of 3,000,000", "mono": true }, { "label": "Without expenses", "value": "Seed 6,000,000, Series A 10,000,000, common and pool 24,000,000", "mono": true }, { "label": "With expenses, 87,000,000 reaches the stack", "value": "Seed 5,550,000, Series A 9,250,000, common and pool 22,200,000", "mono": true }, { "label": "Common's share of the expense", "value": "1,800,000 of 3,000,000 = 60.000 percent, against an ownership of 38.400 percent", "mono": true }, { "label": "Why", "value": "At this exit the residual group is Seed, Series A, common and pool, holding 38.400 + 9.600 + 16.000 = 64.000 percent. Common's share of the residual is 38.400/64.000 = 0.60000", "mono": true } ], "notes": [ "The multiplier is p_common divided by the residual group's total, and it rises as more of the stack sits outside the residual. Below the aggregate preference the multiplier is zero because common is receiving nothing; between the preference and the last flip point it is at its highest.", "This is why a fee negotiated as a percentage of proceeds is a transfer from common specifically. A 1 percent success fee on a 90,000,000 sale costs common 0.60 percent of the whole enterprise, not 0.38 percent.", "Expenses paid by the company before closing rather than out of the consideration produce the same result by a different route, because they reduce the price the buyer will pay. Where they sit in the documents changes the accounting and not the incidence." ] }, { "id": "carve-out-percentage-versus-fixed-pool", "term": "A percentage carve-out and a fixed carve-out pool are different instruments", "definition": "A carve-out sized as a percentage of proceeds grows with the exit; one sized as a fixed pool does not. They cross exactly once, and on either side of the crossing the two structures transfer value in opposite directions between management and everyone else.", "formula": "Percentage plan: carve-out = c*E. Fixed plan: carve-out = K. The two are equal at E = K/c, and the fixed plan is larger below it", "fields": [ { "label": "Setup", "value": "c = 0.10 against a fixed pool K = 6,000,000, so the crossing is at E = 60,000,000", "mono": true }, { "label": "At E = 20,000,000", "value": "Percentage pays management 2,000,000 and the fixed pool pays 6,000,000. Common receives nothing either way; the preferred absorb the whole difference", "mono": true }, { "label": "At E = 20,000,000, Series C proceeds", "value": "Percentage 9,000,000, fixed 7,000,000", "mono": true }, { "label": "At E = 150,000,000", "value": "Percentage pays 15,000,000 and the fixed pool pays 6,000,000", "mono": true }, { "label": "At E = 150,000,000, common and pool", "value": "Percentage 50,400,000, fixed 54,720,000 - a difference of 4,320,000", "mono": true } ], "notes": [ "Management prefers the fixed pool in bad outcomes and the percentage in good ones, which is exactly backwards from the incentive the plan is meant to create. A plan that pays 6,000,000 on a 20,000,000 sale rewards the outcome nobody wanted.", "The structure that aligns is a percentage of proceeds above a threshold, because it pays nothing in the outcomes where the sale destroys value and scales in the outcomes where the buyer had to be persuaded upward.", "Whichever plan is used, it is documented as a bonus plan and not on the cap table, so it will be missing from any waterfall built from the stock ledger. Ask for the plan document by name." ] }, { "id": "carve-out-above-the-preference", "term": "A carve-out on proceeds above the preference costs common much less", "definition": "The same headline percentage applied to proceeds above the aggregate preference rather than to total proceeds pays nothing in a shortfall and less in every outcome. It is a one-line drafting change with a large arithmetic consequence, and it is the version that actually rewards clearing the stack.", "formula": "Threshold plan: carve-out = c*max(0, E - T). Total plan: carve-out = c*E. The difference is c*min(E, T), which is c*T for any exit above the preference", "fields": [ { "label": "Setup", "value": "c = 0.10, T = 60,000,000, all-1x four-series stack", "mono": true }, { "label": "At E = 40,000,000", "value": "Threshold plan pays 0; total plan pays 4,000,000, taken entirely from the preferred", "mono": true }, { "label": "At E = 90,000,000", "value": "Threshold plan pays 3,000,000 against 9,000,000; common and pool receive 22,200,000 against 18,400,000", "mono": true }, { "label": "At E = 150,000,000", "value": "Threshold plan pays 9,000,000 against 15,000,000; common and pool receive 53,280,000 against 50,400,000", "mono": true }, { "label": "The general difference", "value": "c*T = 0.10*60,000,000 = 6,000,000 of consideration, in every outcome above the preference", "mono": true } ], "notes": [ "Read which base the plan uses before arguing about the percentage. Ten percent of proceeds above the preference and ten percent of proceeds are the same sentence with a six-million-dollar difference on this cap table.", "The threshold version is harder to sell to management precisely because it pays nothing in the outcome management is most worried about. That is the argument, not a drafting oversight: the reason to grant a carve-out at all is to create a reason to run a sale that clears the stack.", "A hybrid - a small fixed floor plus a percentage above the preference - is common and is easy to model as the sum of the two plans above." ] }, { "id": "carve-out-with-participating-stack", "term": "A carve-out against a participating stack leaks back to the preferred", "definition": "A carve-out is paid before the waterfall, so it reduces the residual as well as the preference payments. Where the senior series participates in the residual, part of every carve-out dollar is taken from that series and part of it is handed straight back, which changes who is really funding the plan.", "formula": "With a participating series holding residual share p_r, the carve-out reduces its proceeds by p_r*carve-out and reduces common's by p_common/f * carve-out, where f is the residual group's total", "fields": [ { "label": "Setup", "value": "Series C at 1.5x participating capped at 2x, E = 150,000,000, carve-out at 10 percent of E = 15,000,000", "mono": true }, { "label": "Without the carve-out", "value": "Series C 60,000,000, Seed 10,500,000, Series A 17,500,000, Series B 20,000,000, common and pool 42,000,000", "mono": true }, { "label": "With the carve-out", "value": "Series C 60,000,000, Seed 8,250,000, Series A 13,750,000, Series B 20,000,000, common and pool 33,000,000", "mono": true }, { "label": "Cost to common and pool", "value": "9,000,000 of the 15,000,000, or 60.000 percent", "mono": true }, { "label": "Cost to Series C", "value": "0 - Series C is at its cap in both cases, so it funds none of it", "mono": true } ], "notes": [ "A series sitting on its participation cap is immune to the carve-out: its proceeds are fixed at cap*I regardless of what is taken off the top. The plan is therefore funded entirely by the series below the cap and by common.", "That is the opposite of the usual framing, in which the carve-out is described as a concession by the most senior investor. Check whether that investor is inside or outside its cap at the exit values being discussed before crediting the concession.", "The same immunity applies to any non-converting series whose preference is fully covered. Only claimants in the residual pool fund a carve-out." ] }, { "id": "dividends-multiple-series-vintages", "term": "Accrued dividends across series of different vintages", "definition": "A cumulative dividend accrues from each round's closing, so the oldest series has the largest proportional accrual and the newest has none. Applied across a stack the effect is to reweight the preference towards the earliest money, which is the opposite of the seniority the later rounds negotiated.", "formula": "Preference of series j at the measurement date = x_j*I_j*(1 + d*t_j) for simple accrual, or x_j*I_j*(1 + d)^t_j if it compounds, with t_j measured from that series' closing", "fields": [ { "label": "Setup", "value": "d = 0.08 cumulative. Years accrued: Seed 6, Series A 4, Series B 2, Series C 0", "mono": true }, { "label": "Simple accrual", "value": "Seed 2,960,000, Series A 10,560,000, Series B 23,200,000, Series C 30,000,000", "mono": true }, { "label": "Compounding", "value": "Seed 3,173,749, Series A 10,883,912, Series B 23,328,000, Series C 30,000,000", "mono": true }, { "label": "Aggregate preference", "value": "60,000,000 with no dividend, 66,720,000 simple, 67,385,660 compounding", "mono": true }, { "label": "Effect on the common break-even", "value": "The first dollar to common moves from 60,000,000 to 66,720,000, a 11.200 percent increase from a clause that was never declared", "mono": true }, { "label": "Seed's recovery multiple at the aggregate", "value": "Seed's preference has grown 48.000 percent against Series B's 16.000 percent", "mono": true } ], "notes": [ "Compounding costs 665,660 more than simple accrual here on a 60,000,000 stack, and the whole difference is a single word in four charters. It is worth reading rather than assuming, because the two are drafted almost identically.", "The reweighting matters in a shortfall under pari passu ranking, where recovery is proportional to preference amounts. The dividend hands the oldest series a larger share of a small outcome without any change to seniority.", "Accrual usually runs to the date of the distribution, not to the signing date, so a slow closing keeps enlarging the preference. On a stack this size each additional month adds roughly 400,000 to the amount ahead of common.", "A non-cumulative dividend accrues nothing unless declared and is economically nil on these facts. Conceding it is free; conceding a cumulative dividend is the same as agreeing to a rising preference multiple." ] }, { "id": "recapitalisation-new-senior-series", "term": "Recapitalisation by inserting a new senior series", "definition": "A company can clear an overhang by converting the existing preferred to common, or it can leave the existing stack in place and put a new series senior to all of it. The second route is faster and does not need the old holders' economics to be renegotiated, and it makes the overhang worse rather than better.", "formula": "After a senior round of R_new at 1x, T' = T + R_new and the common break-even moves from T to T'. The new series' ownership is R_new/POST and every existing holder is multiplied by PRE/POST", "fields": [ { "label": "Setup", "value": "The four-series ledger, FD = 24,509,805, T = 60,000,000. A new Series D invests 20,000,000 at PRE = 30,000,000, senior to everything", "mono": true }, { "label": "Price and shares", "value": "Price = 30,000,000/24,509,805 = 1.2240; new shares 16,339,870; FD after = 40,849,675", "mono": true }, { "label": "Ownership after", "value": "Series D 40.000 percent, founders 19.584 percent, Seed 5.760 percent, Series A and B 9.600 percent each, Series C 12.000 percent", "mono": true }, { "label": "New aggregate preference", "value": "60,000,000 + 20,000,000 = 80,000,000, so the first dollar to common moves up 20,000,000", "mono": true }, { "label": "Worked at E = 100,000,000", "value": "Series D 24,418,605, Series C 30,000,000, Series B 20,000,000, Series A 8,000,000, Seed 3,516,279, common and pool 14,065,116", "mono": true } ], "notes": [ "A senior insertion is the cheapest financing to document and the most expensive to live with. It raises the preference by the full amount raised while diluting everyone, so common is worse off on both axes at once.", "Issuing a senior or pari passu security is on every standard list of protective provisions, so the existing series have to consent. In practice the consent is obtained by the same holders funding the new series, which is why this transaction is usually an inside round.", "The genuine reset is the other route: convert the whole stack to common and reduce the old cap table to a stated residual. That extinguishes the preference instead of adding to it, and it is the transaction a new management pool needs in order to be worth anything.", "Compare the two by computing the common break-even under each, not by comparing the headline valuations. Here the senior insertion sets it at 80,000,000 while a conversion recapitalisation sets it at the new money alone." ] }, { "id": "asset-sale-merger-ipo-treatment", "term": "Asset sale, merger and initial public offering are three different payouts", "definition": "The same enterprise value produces three different distributions depending on the legal form of the exit. A merger runs the charter waterfall on the consideration. An asset sale adds a corporate tax layer before anything is distributed. An initial public offering usually converts the whole preferred stack to common and pays no preference at all.", "formula": "Merger: distribute E through the waterfall. Asset sale by a C corporation: distribute E - corporate tax on the gain, then run the waterfall. Qualifying public offering: all preferred converts, so each holder receives p*E with no preference and no election", "fields": [ { "label": "Merger at 100,000,000", "value": "Seed 7,500,000, Series A 12,500,000, Series B 20,000,000, Series C 30,000,000, common and pool 30,000,000", "mono": true }, { "label": "Asset sale at 100,000,000", "value": "Tax basis 10,000,000, corporate rate 0.21: tax = 18,900,000, so 81,100,000 is distributable. Seed 4,620,000, common and pool 18,480,000", "mono": true }, { "label": "Common's share of the corporate tax", "value": "11,520,000 of 18,900,000 = 60.952 percent", "mono": true }, { "label": "Public offering at a 100,000,000 market value", "value": "Preferred converts: Seed 9,600,000, Series A 16,000,000, Series B 16,000,000, Series C 20,000,000, common and pool 38,400,000", "mono": true }, { "label": "What the conversion costs Series C", "value": "30,000,000 as a preference against 20,000,000 as common - the preference is worth 10,000,000 more at this value, which is why mandatory conversion is drafted with a price threshold", "mono": true } ], "notes": [ "The rate used above is the stated federal corporate rate; state tax, the character of the assets and the seller's attributes all move the real number. The structural point is that an asset sale interposes an entity-level tax that a stock sale or merger does not, and that common bears most of it.", "Buyers prefer asset purchases for liability and basis reasons and sellers prefer stock deals for exactly the tax reason above. The gap is negotiated as price, so the form of the deal is an economic term and not a legal detail.", "Mandatory conversion on a public offering is the one event that removes the entire preference at once. That is why its definition - a minimum price, a minimum size, sometimes a named exchange - is negotiated harder than the preference itself.", "A charter that treats an asset sale as a deemed liquidation but does not address a subsequent dissolution can strand proceeds at the company level. Read the deemed liquidation definition and the dissolution waterfall together." ], "source": "Delaware General Corporation Law section 271 (sale of substantially all assets) and section 251 (merger)" }, { "id": "deemed-liquidation-event", "term": "Deemed liquidation events", "definition": "A liquidation preference is written to apply on a liquidation, dissolution or winding up. Almost no venture-backed company reaches an exit that way, so the charter defines a list of transactions that are treated as if they were a liquidation. That definition, not the preference itself, decides whether the preference applies to the transaction that actually happens.", "formula": "No arithmetic. The definition is a list, and the preference applies if and only if the transaction is on it", "fields": [ { "label": "Commonly on the list", "value": "A merger or consolidation in which the company's stockholders cease to hold a majority of the voting power of the surviving entity; a sale, lease or exclusive licence of all or substantially all assets; a sale of a majority of the outstanding stock", "mono": false }, { "label": "Commonly excluded", "value": "A reincorporation or holding-company reorganisation; a financing; a transaction with a wholly owned subsidiary", "mono": false }, { "label": "Usually separate", "value": "A qualifying public offering, which triggers mandatory conversion instead of a preference payment", "mono": false }, { "label": "Where it bites", "value": "An exclusive licence of the core technology can be a disposal of substantially all assets in substance while not looking like a sale in form", "mono": false } ], "notes": [ "The gap that matters is between a transaction that transfers the business and a transaction that is on the list. A sale of a subsidiary holding the main product, a licence granting all commercial rights in every field, or a sale of a bare majority of the stock can each move the business without meeting a narrowly drafted definition.", "Whether the definition is drafted as automatic payment or as an option for the preferred to demand payment changes who controls the timing. An option gives the holder a second decision after the price is known.", "The definition also drives the acceleration triggers in equity grants and often the drag-along. A change of control defined one way in the charter and another way in the option plan produces a transaction that pays the preference without accelerating any employee equity.", "Ask for the definition as drafted rather than the summary. It is the single clause on which the entire liquidation section of the corpus depends for having any application at all." ], "source": "NVCA model Amended and Restated Certificate of Incorporation (definition of Deemed Liquidation Event)" }, { "id": "seniority-four-series-shortfall", "term": "Strict seniority across four series - the zero thresholds", "definition": "Under seniority-ordered payment each series has a single number that determines everything about its recovery in a shortfall: the aggregate preference ranking ahead of it. Below that number it receives nothing at all, and above it plus its own preference it is paid in full. The band between the two is narrow.", "formula": "R_j = min(x_j*I_j, max(0, E - S_j)), where S_j is the sum of preference amounts senior to j. Series j receives nothing for E <= S_j and is paid in full for E >= S_j + x_j*I_j", "fields": [ { "label": "Setup", "value": "Series C senior to B, B to A, A to Seed. All 1x, T = 60,000,000", "mono": true }, { "label": "Senior preference S_j", "value": "Seed 58,000,000; Series A 50,000,000; Series B 30,000,000; Series C 0", "mono": true }, { "label": "Full recovery from", "value": "Seed 60,000,000; Series A 58,000,000; Series B 50,000,000; Series C 30,000,000", "mono": true }, { "label": "Worked at E = 40,000,000", "value": "Series C 30,000,000; Series B 10,000,000; Series A 0; Seed 0", "mono": true }, { "label": "Same exit, pari passu", "value": "Series C 20,000,000; Series B 13,333,333; Series A 5,333,333; Seed 1,333,333", "mono": true }, { "label": "Seed's band", "value": "Nothing below 58,000,000 and 2,000,000 above 60,000,000 - a 2,000,000-wide transition over a 60,000,000 range", "mono": true } ], "notes": [ "Seniority converts every junior series into a binary claim. That is materially worse than it sounds, because the exit values at which seniority matters are exactly the ones the company is most likely to reach when it is being sold under pressure.", "The comparison to run is not stacked against pari passu at the last post-money, which is above T and where the two are identical. Run it at half of T and at three quarters of T.", "A four-series stack has three seniority decisions, taken in three separate negotiations, years apart, each time by an investor comparing outcomes that already look poor. That is how a stack ends up strictly ordered without anyone having agreed to order it.", "Tiered ranking - groups of series equal within the group and ordered between groups - is the usual landing point and is modelled by treating each tier as a single series with the tier's aggregate preference." ] }, { "id": "common-breakeven-four-series", "term": "The first dollar to common, with every deduction added", "definition": "The common break-even is the aggregate preference, adjusted for anything paid ahead of it. Each item is small and each is documented somewhere other than the charter, and together they move the number by half again.", "formula": "Break-even = (T + accrued dividends + net debt + transaction expenses + fixed carve-out) / (1 - c), where c is any carve-out expressed as a fraction of proceeds", "fields": [ { "label": "Aggregate preference T", "value": "2,000,000 + 8,000,000 + 20,000,000 + 30,000,000 = 60,000,000", "mono": true }, { "label": "Plus accrued dividends at d = 0.08 by vintage", "value": "66,720,000 - an increase of 6,720,000", "mono": true }, { "label": "Plus net debt of 6,000,000 and expenses of 3,000,000", "value": "75,720,000", "mono": true }, { "label": "Divided by (1 - c) for a 10 percent carve-out", "value": "84,133,333", "mono": true }, { "label": "Check", "value": "0.10 * 84,133,333 = 8,413,333 to management, leaving exactly 75,720,000 for debt, expenses and the stack", "mono": true }, { "label": "Against the last post-money", "value": "150,000,000. The break-even is 56.089 percent of it", "mono": true } ], "notes": [ "Every input here is a known historical number or a stated contract term. The break-even involves no forecast at all, which makes it the one number in a venture cap table that can be computed to the dollar and audited.", "It is also the number that is never quoted. Employees are told the last post-money valuation, which is a price paid for a senior instrument, and are left to infer that their common is worth the same fraction of it.", "Recompute it after every financing and after every year of accrual, not just at the point of a sale. The whole purpose of the figure is to know in advance which outcomes are worth pursuing.", "If the break-even has passed the exit values the business can credibly reach, further operating progress cannot make common valuable. That is the specific condition a recapitalisation exists to reset, and the break-even is how it is diagnosed." ] } ], "reviewed": "2026-08-27" }, { "slug": "dilution", "title": "Dilution and cap table math", "lede": "Pre-money, post-money, option pools, and the arithmetic of who absorbs the dilution.", "intro": "Ownership after a round is a function of three numbers: the pre-money valuation, the amount raised, and the fully diluted share count used as the denominator. Disputes about dilution are almost always disputes about the third one.", "tables": [ { "title": "Core identities", "columns": [ "Quantity", "Formula" ], "rows": [ [ "Post-money valuation", "POST = PRE + amount raised" ], [ "Investor ownership", "amount raised / POST" ], [ "Price per share", "PRE / fully diluted shares outstanding pre-round" ], [ "New shares issued", "amount raised / price per share" ], [ "Existing holder ownership after", "prior ownership * (PRE / POST)" ], [ "Dilution to existing holders", "1 - (PRE / POST) = amount raised / POST" ] ] }, { "title": "Option pool placement - 8,000,000 pre-money, 2,000,000 raised, 15 percent pool", "intro": "The same headline terms produce different founder ownership depending on whether the pool is created before or after the money. Figures are ownership percentages after the round.", "columns": [ "Stakeholder", "Pool created pre-money", "Pool created post-money" ], "rows": [ [ "New investor", "20.00", "20.00" ], [ "Option pool", "15.00", "15.00" ], [ "Founders and existing", "65.00", "65.00" ], [ "Effective pre-money paid", "8,000,000 less the pool value", "8,000,000" ], [ "Founder ownership impact", "Founders absorb the full pool dilution", "Pool dilution shared with the new investor" ] ] }, { "title": "Anti-dilution - three formulations on the same down round", "intro": "CP1 = 1.0000. Fully diluted shares before the new issue A = 10,000,000, of which outstanding preferred on an as-converted basis is 4,000,000 (the narrow base). The down round raises 3,000,000 at 0.5000 per share, so C = 6,000,000 shares issued and B = 3,000,000/1.0000 = 3,000,000. The protected holder owns 4,000,000 preferred shares. Ownership figures are after the round, including the adjustment shares.", "columns": [ "Formulation", "CP2", "Shares on conversion", "Protected holder", "Common and pool", "New investor" ], "rows": [ [ "No adjustment", "1.0000", "4,000,000", "25.000 percent", "37.500 percent", "37.500 percent" ], [ "Broad-based weighted average", "0.8125", "4,923,077", "29.091 percent", "35.455 percent", "35.455 percent" ], [ "Narrow-based weighted average", "0.7000", "5,714,286", "32.258 percent", "33.871 percent", "33.871 percent" ], [ "Full ratchet", "0.5000", "8,000,000", "40.000 percent", "30.000 percent", "30.000 percent" ] ] }, { "title": "Option pool placement - share-level detail", "intro": "8,000,000 founder shares before the round. PRE = 8,000,000 and R = 2,000,000, so POST = 10,000,000. A pool of 1,411,765 shares is created either way - the only difference is whether those shares are inside the pre-money count used to set the price. 1,411,765 is 15 percent of the enlarged pre-round count, computed as 8,000,000 * 0.15/0.85.", "columns": [ "Quantity", "Pool inside the pre-money", "Pool outside the pre-money" ], "rows": [ [ "Founder shares before the round", "8,000,000", "8,000,000" ], [ "Pool shares created", "1,411,765", "1,411,765" ], [ "Pre-money count used to set the price", "9,411,765", "8,000,000" ], [ "Price per share", "0.8500", "1.0000" ], [ "Shares to the new investor for 2,000,000", "2,352,941", "2,000,000" ], [ "Total shares after the round", "11,764,706", "11,411,765" ], [ "Founder ownership after", "68.000 percent", "70.103 percent" ], [ "New investor ownership after", "20.000 percent", "17.526 percent" ], [ "Pool as a share of the total", "12.000 percent", "12.371 percent" ], [ "Founder shares valued at the round price", "6,800,000", "8,000,000" ] ] }, { "title": "Full share ledger, seed through Series C", "intro": "One company, four priced rounds, no secondary and no pool refresh after the seed. The seed round is the one already worked elsewhere in this corpus: 8,000,000 founder shares, a 1,411,765-share pool created inside the pre-money count, and 2,000,000 raised at a pre-money of 8,000,000. Series A raises 8,000,000 at 24,000,000 pre, Series B raises 20,000,000 at 80,000,000 pre, and Series C raises 30,000,000 at 120,000,000 pre. Every figure below is derived from those inputs and the share counts carry through to the liquidation, valuation and employee equity sections.", "columns": [ "Stage", "Founders", "Option pool", "Seed", "Series A", "Series B", "Series C", "Total shares", "Price per share", "Founder ownership" ], "rows": [ [ "At incorporation", "8,000,000", "-", "-", "-", "-", "-", "8,000,000", "-", "100.000" ], [ "Seed option pool created inside the pre-money", "8,000,000", "1,411,765", "-", "-", "-", "-", "9,411,765", "-", "85.000" ], [ "Seed closes - 2,000,000 at PRE 8,000,000", "8,000,000", "1,411,765", "2,352,941", "-", "-", "-", "11,764,706", "0.8500", "68.000" ], [ "Series A closes - 8,000,000 at PRE 24,000,000", "8,000,000", "1,411,765", "2,352,941", "3,921,569", "-", "-", "15,686,275", "2.0400", "51.000" ], [ "Series B closes - 20,000,000 at PRE 80,000,000", "8,000,000", "1,411,765", "2,352,941", "3,921,569", "3,921,569", "-", "19,607,844", "5.1000", "40.800" ], [ "Series C closes - 30,000,000 at PRE 120,000,000", "8,000,000", "1,411,765", "2,352,941", "3,921,569", "3,921,569", "4,901,961", "24,509,805", "6.1200", "32.640" ], [ "Ownership at close, percent", "32.640", "5.760", "9.600", "16.000", "16.000", "20.000", "100.000", "-", "-" ] ] }, { "title": "An option pool refresh at every round, and what it costs", "intro": "The same three rounds, with a pool refresh equal to 5 percent of the post-round share count created inside the pre-money count at each one. Because the refresh sits inside the pre-money count, the existing holders fund all of it and the new investor still takes exactly q. The founders' retention factor is therefore (1 - g - q) rather than (1 - q). The last two columns compare founder ownership with the ledger above, which has no refresh after the seed.", "columns": [ "Round", "Investor share q", "Refresh g", "Pool shares created", "Pre-money count used to price", "Price per share", "Shares to the investor", "Total shares after", "Founder ownership", "Without any refresh", "Cost in points" ], "rows": [ [ "Series A", "25.000", "5.000", "840,336", "12,605,042", "1.9040", "4,201,681", "16,806,723", "47.600", "51.000", "3.400" ], [ "Series B", "20.000", "5.000", "1,120,448", "17,927,171", "4.4625", "4,481,793", "22,408,964", "35.700", "40.800", "5.100" ], [ "Series C", "20.000", "5.000", "1,493,931", "23,902,895", "5.0203", "5,975,724", "29,878,619", "26.775", "32.640", "5.865" ] ] }, { "title": "Two successive triggering issuances against one adjustment for both", "intro": "CP1 = 1.0000 and the protected holder owns 4,000,000 preferred shares out of a fully diluted 10,000,000. The first down round raises 3,000,000 at 0.5000, issuing 6,000,000 shares. A second raises 2,000,000 at 0.4000, issuing 5,000,000. The broad-based weighted average is applied twice in sequence, then compared with a single adjustment computed as if both issuances had happened at once, and with full ratchet.", "columns": [ "Step", "Conversion price", "Shares on conversion", "Fully diluted after", "Protected holder ownership" ], "rows": [ [ "Before any down round", "1.0000", "4,000,000", "10,000,000", "40.000" ], [ "After issuance one, weighted average", "0.8125", "4,923,077", "16,923,077", "29.091" ], [ "After issuance two, weighted average", "0.7184", "5,567,766", "22,567,766", "24.671" ], [ "Both issuances as a single adjustment", "0.7143", "5,600,000", "22,600,000", "24.779" ], [ "Full ratchet, either sequence", "0.4000", "10,000,000", "27,000,000", "37.037" ] ] } ], "entries": [ { "id": "option-pool-shuffle", "term": "The option pool shuffle", "definition": "Placing a newly created option pool inside the pre-money share count means the pool is carved out of existing holders alone, not shared with the incoming investor. The headline pre-money valuation is unchanged while the effective price paid per existing share falls.", "formula": "Effective pre-money = stated PRE * (1 - pool percentage)", "fields": [ { "label": "Worked", "value": "Stated PRE 8,000,000 with a 15 percent pre-money pool: effective PRE = 6,800,000", "mono": true }, { "label": "Value transferred", "value": "1,200,000, from existing holders to the incoming investor", "mono": true } ], "notes": [ "This is not a hidden term - it appears plainly in the term sheet as the fully diluted definition. It is simply frequently not modelled.", "The negotiable variables are the size of the pool and whether it sits pre or post. Pool size should be driven by an actual hiring plan for the period to the next round, not a round number.", "A pool sized far above the hiring plan is economically identical to a lower pre-money valuation." ] }, { "id": "fully-diluted", "term": "Fully diluted share count", "definition": "The denominator used to compute ownership percentages. Which instruments are included is definitional and set by the documents, not by accounting convention.", "notes": [ "Commonly included: outstanding common, outstanding preferred on an as-converted basis, all granted options whether vested or not, and the unallocated option pool.", "Sometimes disputed: outstanding warrants, unconverted SAFEs and notes, and pool increases contemplated but not yet authorised.", "Every percentage in a term sheet is meaningless until the denominator is specified. Ask for the definition before modelling anything." ] }, { "id": "anti-dilution", "term": "Anti-dilution adjustment", "definition": "A mechanism that adjusts the conversion price of preferred stock downward when the company subsequently issues shares at a lower price, protecting the earlier investor from the down round.", "fields": [ { "label": "Full ratchet", "value": "New conversion price = the new lower issue price, regardless of how few shares are issued", "mono": false }, { "label": "Broad-based weighted average", "value": "CP2 = CP1 * (A + B) / (A + C)", "mono": true }, { "label": "A", "value": "Fully diluted shares outstanding before the new issue", "mono": false }, { "label": "B", "value": "Consideration received, divided by CP1 - the shares the money would have bought at the old price", "mono": false }, { "label": "C", "value": "Shares actually issued in the new round", "mono": false } ], "notes": [ "Narrow-based weighted average uses only outstanding preferred in A rather than all fully diluted shares, producing a larger adjustment for the investor.", "Full ratchet is severe: a single share issued at a low price resets the entire earlier round to that price. Broad-based weighted average is the common default.", "Anti-dilution adjusts the conversion ratio, not the share count held. The investor ends up converting into more common shares." ] }, { "id": "pool-percentage-base", "term": "Pool percentage - of which base?", "definition": "A term sheet that says a 15 percent pool has not said what the 15 percent is measured against. Measured on the pre-round count and measured on the post-round count give different pool sizes, different prices, and a different cost to founders, on identical headline terms.", "formula": "Pool as a fraction f of the enlarged pre-round count: pool = s_existing * f/(1 - f). Pool as a fraction g of the post-round count: pool = g * S_post, where S_post = s_existing/(1 - g - q) and q is the new investor's ownership", "fields": [ { "label": "Setup", "value": "s_existing = 8,000,000 shares, PRE = 8,000,000, R = 2,000,000, so q = 0.20", "mono": true }, { "label": "Worked, f = 0.15 pre-round base", "value": "pool = 8,000,000 * 0.15/0.85 = 1,411,765. Price = 8,000,000/9,411,765 = 0.8500. Founder shares worth 6,800,000", "mono": true }, { "label": "Worked, g = 0.15 post-round base", "value": "S_post = 8,000,000/(1 - 0.15 - 0.20) = 12,307,692. pool = 1,846,154. Price = 2,000,000/(0.20*12,307,692) = 0.8125. Founder shares worth 6,500,000", "mono": true }, { "label": "Difference", "value": "300,000 of value on a 2,000,000 round, from one undefined word", "mono": true } ], "notes": [ "The post-round base is the larger pool and the lower price, because the pool has to be big enough to survive the round's own dilution. Investors asking for a percentage of the post-round capitalisation are asking for a bigger pool than the same number implies pre-round.", "The check that resolves it in one line: ask for the closing cap table with the pool as its own row and the price per share stated. Any ambiguity about the base becomes a specific share count.", "Both conventions appear in real term sheets and both are described in this corpus. On the same 8,000,000 pre-money and 2,000,000 raise, a 15 percent post-round pool produces 20.000 percent to the investor, 15.000 percent pool and 65.000 percent to founders at a 0.8125 price; a 15 percent pre-round pool produces 20.000 / 12.000 / 68.000 at a 0.8500 price. Neither is wrong - they are different terms wearing the same number.", "The pool size argument is winnable on facts - a hiring plan through to the next round produces a defensible number - while the base argument is winnable only by reading the definition. Do the second one first." ] }, { "id": "weighted-average-sensitivity", "term": "What drives a weighted-average adjustment", "definition": "The weighted-average formula responds to the size of the down round as well as its price. A small issuance at a low price barely moves the conversion price; a large one at the same price moves it a great deal. Full ratchet ignores size entirely.", "formula": "CP2 = CP1 * (A + B)/(A + C), with B = consideration received / CP1 and C = shares actually issued", "fields": [ { "label": "Setup", "value": "CP1 = 1.0000, A = 10,000,000, new price 0.5000", "mono": true }, { "label": "Worked, raise 500,000", "value": "C = 1,000,000, B = 500,000: CP2 = 10,500,000/11,000,000 = 0.9545", "mono": true }, { "label": "Worked, raise 3,000,000", "value": "C = 6,000,000, B = 3,000,000: CP2 = 13,000,000/16,000,000 = 0.8125", "mono": true }, { "label": "Worked, raise 10,000,000", "value": "C = 20,000,000, B = 10,000,000: CP2 = 20,000,000/30,000,000 = 0.6667", "mono": true }, { "label": "Full ratchet, all three", "value": "CP2 = 0.5000 in every case", "mono": true } ], "notes": [ "This is the precise sense in which full ratchet is severe: the adjustment is unrelated to the harm. A 500,000 bridge at a low price resets the entire earlier round identically to a 10,000,000 recapitalisation.", "Because C appears in the denominator, the adjustment is self-limiting - a very large down round dilutes the protected holder anyway, adjustment or not. Anti-dilution protects against being repriced, not against being diluted.", "The narrow base uses only outstanding preferred in A. On the numbers above that changes CP2 from 0.8125 to 0.7000 - a 30.00 percent price reduction instead of 18.75 percent, and 16.1 percent more shares on conversion - from a single definitional choice about which shares count." ], "source": "NVCA model Amended and Restated Certificate of Incorporation (conversion price adjustment provisions)" }, { "id": "anti-dilution-incidence", "term": "Who actually pays for an anti-dilution adjustment", "definition": "The adjustment shares are new shares, so they dilute everyone outside the protected class - including the investor who priced the down round. That is why a waiver or reset from the protected series is usually a closing condition rather than a courtesy.", "formula": "New investor ownership = C/(A + C + adjustment shares) with the adjustment, and C/(A + C) with it waived", "fields": [ { "label": "Worked, broad-based", "value": "6,000,000/16,923,077 = 35.455 percent", "mono": true }, { "label": "Worked, waived", "value": "6,000,000/16,000,000 = 37.500 percent", "mono": true }, { "label": "Cost to the new investor", "value": "2.045 percentage points, for terms it did not negotiate", "mono": true }, { "label": "Under full ratchet", "value": "6,000,000/20,000,000 = 30.000 percent, a 7.500 point cost", "mono": true } ], "notes": [ "Founders often assume anti-dilution is a fight between them and the earlier investor. It is usually a fight between the earlier investor and the new one, with the founder's ownership as collateral damage in both directions.", "This alignment is useful. The new investor's insistence on a waiver does more to protect the common than any founder argument, so the productive move is to let the two investors resolve it and to negotiate the pool and the preference instead.", "A partial waiver - resetting to a stated conversion price rather than the formula result - is the common landing point and is easy to model: it is just a chosen CP2." ] }, { "id": "dilution-compounding", "term": "Dilution compounds, it does not add", "definition": "Ownership after a sequence of rounds is a product of retention factors, not a subtraction of percentages. Adding the round-by-round dilution figures overstates the total, and the error grows with the number of rounds.", "formula": "Ownership after n rounds = q_0 * product of (PRE_i/POST_i). A pool refresh of fraction g_i in round i multiplies in a further (1 - g_i)", "fields": [ { "label": "Setup", "value": "Seed 2,000,000 at PRE 8,000,000; Series A 8,000,000 at PRE 24,000,000; Series B 20,000,000 at PRE 80,000,000", "mono": true }, { "label": "Retention factors", "value": "8/10 = 0.80; 24/32 = 0.75; 80/100 = 0.80", "mono": true }, { "label": "Worked", "value": "1.00 * 0.80 * 0.75 * 0.80 = 0.4800, so founders hold 48.000 percent before any pool", "mono": true }, { "label": "With a 5 percent pool refresh each round", "value": "0.4800 * 0.95^3 = 0.4800 * 0.857375 = 41.154 percent", "mono": true }, { "label": "The additive error", "value": "20 + 25 + 20 = 65 percent of dilution added up, against 52.0 percent actual", "mono": true } ], "notes": [ "The product form makes the marginal cost of a round explicit: a round is a multiplier, so its cost in percentage points depends on how much is left, not on the round's own size. The fourth round of 20 percent costs less in points than the first.", "Pool refreshes are the term most often left out of a founder's own model, and three of them cost 6.8 points here on top of the priced rounds. Model each refresh as its own factor.", "The same product applies to every holder, so relative ownership between existing holders never changes through a priced round. Only new issuance moves relative positions." ] }, { "id": "denominator-price-effect", "term": "Every disputed inclusion is a price cut", "definition": "Price per share is the pre-money valuation divided by the pre-money fully diluted count. Each instrument added to that count reduces the price by exactly its share of the enlarged denominator, which converts an argument about definitions into an argument about a number.", "formula": "Price = PRE / FD_pre. Adding shares multiplies the price by FD_before/FD_after, a reduction of 1 - FD_before/FD_after", "fields": [ { "label": "PRE = 8,000,000 on 8,000,000 shares", "value": "1.0000", "mono": true }, { "label": "Worked, add a 1,411,765 pool", "value": "9,411,765 shares -> 0.8500, a 15.000 percent cut", "mono": true }, { "label": "Worked, also count 400,000 warrant shares", "value": "9,811,765 shares -> 0.8153, a further 4.077 percent", "mono": true }, { "label": "Worked, also count 1,000,000 SAFE conversion shares", "value": "10,811,765 shares -> 0.7399, a further 9.249 percent", "mono": true }, { "label": "Cumulative", "value": "1.0000 down to 0.7399 - a 26.007 percent reduction in the price paid for the same company", "mono": true } ], "notes": [ "The negotiation over whether the warrants count is a negotiation over 4.077 percent of the price. Price it before arguing it; the answer is frequently that the point is not worth the goodwill.", "Order does not matter to the final price but does matter to how the argument is framed. Ask for one fully diluted definition and one share count, then compute the price once, rather than debating instruments one at a time.", "The most commonly contested item is the unallocated pool, and it is also the largest. Anything after it is a rounding argument by comparison." ] }, { "id": "pay-to-play-arithmetic", "term": "What declining a pay-to-play round actually costs", "definition": "A pay-to-play provision converts a non-participating holder's preferred into common. The preference amount is extinguished and the resulting common sits behind the entire remaining preferred stack, so in any shortfall outcome the position goes to nothing.", "formula": "Before conversion the holder receives x*I * E/T in a pari passu shortfall. After conversion it receives nothing until E exceeds the remaining stack T', so the loss is x*I * E/T for all E < T'", "fields": [ { "label": "Setup", "value": "Series A preference 4,000,000 (the declining holder), Series B 20,000,000, new Series C 10,000,000. T = 34,000,000, pari passu", "mono": true }, { "label": "Worked, holder participates", "value": "At E = 20,000,000: A receives 4,000,000 * 20/34 = 2,352,941; B 11,764,706; C 5,882,353", "mono": true }, { "label": "Worked, holder converted to common", "value": "T' = 30,000,000 > 20,000,000, so B receives 13,333,333, C receives 6,666,667, and all common receives nothing", "mono": true }, { "label": "Cost of declining", "value": "The entire 2,352,941, plus the loss of protective provisions and anti-dilution", "mono": true } ], "notes": [ "The clause is written as a penalty but functions as a pricing mechanism: it sets the cost of not funding equal to the whole existing position, which is what makes an insider bridge happen when no outside price exists.", "It also concentrates the cap table exactly when the company can least afford a governance fight, because the holders who convert lose their consent rights at the same moment.", "From the holder's side, the decision is not the pro-rata cheque against the position's book value; it is the cheque against the position's value in the shortfall outcomes that triggered the round in the first place." ] }, { "id": "recapitalisation-arithmetic", "term": "Recapitalisation and cram-down arithmetic", "definition": "A recapitalisation clears an accumulated preference overhang by converting all preferred to common and reducing the existing cap table to a stated residual percentage. The reverse split that usually accompanies it is cosmetic; the residual percentage is the whole transaction.", "formula": "Post-recap ownership of the new money = R/(R + implied value of the residual). Setting a residual fraction v for the old cap table gives new money ownership 1 - v and an implied post-money of R/(1 - v)", "fields": [ { "label": "Setup", "value": "20,000,000 existing shares carrying 30,000,000 of aggregate preference. New money R = 5,000,000. Residual to the old cap table v = 0.10", "mono": true }, { "label": "Mechanics", "value": "All preferred converts to common; the old 20,000,000 shares are reverse split 20:1 to 1,000,000; the new investor buys 9,000,000 shares", "mono": true }, { "label": "Worked, price and valuation", "value": "Price = 5,000,000/9,000,000 = 0.5556. Post-money = 10,000,000 * 0.5556 = 5,555,556. Pre-money = 1,000,000 * 0.5556 = 555,556", "mono": true }, { "label": "What was extinguished", "value": "30,000,000 of preference, in exchange for 555,556 of implied residual value", "mono": true } ], "notes": [ "The purpose is usually not to punish existing holders but to make a new management pool worth something. A pool granted behind a 30,000,000 overhang has no value at any achievable exit, so no grant retains anyone.", "The reverse split ratio is chosen for share-count tidiness and is often mistaken for the economic term. Ask for the residual percentage; the ratio follows from it.", "A new pool is normally created inside the post-recap capitalisation, so the residual is smaller than the headline v. Confirm whether v is stated before or after the pool." ] }, { "id": "secondary-is-not-dilution", "term": "A secondary sale is not dilution", "definition": "A transfer of existing shares moves ownership between holders and leaves the fully diluted count unchanged, so no other holder's percentage moves. Only new issuance dilutes. The two are frequently conflated because both change who owns what.", "formula": "After a transfer of s shares, FD is unchanged; the seller holds (s_seller - s)/FD and the buyer s/FD. Every other holder's percentage is identical to before", "fields": [ { "label": "Setup", "value": "FD = 11,411,765 shares, of which founders hold 8,000,000 = 70.103 percent", "mono": true }, { "label": "Worked", "value": "A founder sells 1,000,000 shares in a secondary. Buyer holds 1,000,000/11,411,765 = 8.763 percent; founders hold 7,000,000/11,411,765 = 61.340 percent", "mono": true }, { "label": "Check", "value": "8.763 + 61.340 = 70.103 percent, unchanged. No other holder moves", "mono": true } ], "notes": [ "Secondary sales of common are an input to the fair market value of common, so a founder secondary can raise the strike price on every option granted afterwards. The transaction is free of dilution and not free of cost.", "A secondary priced above the last preferred round is harder to explain in a subsequent 409A analysis than a primary at the same price, because the buyer is paying that price for common rather than for preferred.", "Transfer restrictions, rights of first refusal and co-sale usually make the sale smaller than proposed. Model the co-sale cut-back before agreeing a size with the buyer." ] }, { "id": "share-ledger-seed-to-c", "term": "The share ledger is the only unambiguous statement of a cap table", "definition": "Percentages are derived quantities and depend on a denominator that is defined by the documents. Share counts are not. A ledger that carries every holder's share count at every stage answers every ownership, price and waterfall question without a single further assumption.", "formula": "Ownership of holder i after round n = s_i / FD_n, where FD_n = FD_(n-1) + new pool shares + shares issued in round n. Price in round n = PRE_n / (FD_(n-1) + new pool shares)", "fields": [ { "label": "Founder path", "value": "100.000 percent at incorporation, then 68.000, 51.000, 40.800, 32.640 percent", "mono": true }, { "label": "Retention factors", "value": "0.85000 for the pool, then 0.80000, 0.75000, 0.80000, 0.80000 for the four rounds", "mono": true }, { "label": "Check", "value": "1.00000 * 0.85000 * 0.80000 * 0.75000 * 0.80000 * 0.80000 = 0.32640, which is 32.640 percent", "mono": true }, { "label": "Total shares issued", "value": "24,509,805 fully diluted, of which 15,098,040 were sold for cash", "mono": true }, { "label": "Total cash raised", "value": "2,000,000 + 8,000,000 + 20,000,000 + 30,000,000 = 60,000,000", "mono": true }, { "label": "Blended price paid", "value": "60,000,000 / 15,098,040 = 3.9740 per preferred share", "mono": true } ], "notes": [ "The pool is the only line whose count never changes here, and that is what makes it fall from 12.000 percent to 5.760 percent of the company. An unrefreshed pool dilutes exactly like any other static holding, which is why a hiring plan two rounds out needs a refresh rather than an allocation.", "Because no secondary occurs, relative ownership between the pre-existing holders is unchanged by every round: each is multiplied by the same PRE/POST. All movement in relative position comes from new issuance.", "Ask for this table, not for percentages. Two people quoting different ownership figures for the same company are almost always using different denominators, and a share ledger makes the disagreement disappear in one line.", "The blended price paid across the four rounds is well below the last round price, which is the arithmetic behind an investor group's aggregate return being much better than the last round's implied multiple suggests." ] }, { "id": "pool-refresh-who-funds-it", "term": "Who funds an option pool refresh, and the two conventions", "definition": "A refresh created inside the pre-money share count is funded by the existing holders alone and the new investor still receives exactly q. A refresh created after the round is shared by everyone including the new investor. The first is the standard convention and it changes the founders' retention factor from (1 - q) to (1 - g - q).", "formula": "Pool inside the pre-money: S_post = s_existing/(1 - g - q), so existing holders retain (1 - g - q). Pool outside: existing holders retain (1 - q)*(1 - g)", "fields": [ { "label": "Series A, refresh inside the pre-money", "value": "q = 25.000 percent, g = 5.000 percent: retention = 1 - 0.05 - 0.25 = 0.70000, so founders go from 68.000 to 47.600 percent", "mono": true }, { "label": "Series A, refresh outside", "value": "Retention = 0.75000 * 0.95000 = 0.71250, so founders go to 48.450 percent", "mono": true }, { "label": "Difference on one round", "value": "0.850 of a percentage point, and the new investor's ownership moves from 25.000 to 23.750 percent", "mono": true }, { "label": "Three refreshes inside the pre-money", "value": "Founders 68.000 to 47.600 to 35.700 to 26.775 percent", "mono": true }, { "label": "Same rounds, no refresh", "value": "68.000 to 51.000 to 40.800 to 32.640 percent", "mono": true }, { "label": "Total cost of the three refreshes", "value": "5.865 percentage points of the company", "mono": true } ], "notes": [ "The two conventions are not a matter of house style. The inside version is a price reduction: the pool shares enlarge the denominator used to set the price, so the same pre-money valuation buys the investor more shares. The outside version leaves the price alone and dilutes the investor with everyone else.", "Read the term as three separate questions: how large, measured against which base, and created inside or outside the pre-money count. A term sheet frequently answers only the first.", "Size the refresh from a hiring plan through to the next expected financing and no further. A pool sized to cover two rounds of hiring is a discount on this round's price handed over for grants that will be made at a higher strike.", "The existing corpus models a post-money refresh, where each round multiplies ownership by a further (1 - g). Both models appear in real deals and they differ by g*q per round - here 1.25 percentage points at Series A. Establish which one the closing cap table uses before reconciling to it." ] }, { "id": "anti-dilution-successive-issuances", "term": "Two triggering issuances give less protection than one adjustment for both", "definition": "Weighted-average anti-dilution is applied to each triggering issuance in turn, using the conversion price and the share count as they stand at that moment. Because the first adjustment has already enlarged the fully diluted count, the second adjustment is computed on a larger base and moves the price less. Sequential application is therefore weaker than a single combined adjustment.", "formula": "CP_(n+1) = CP_n * (A_n + B_n)/(A_n + C_n), with A_n the fully diluted count immediately before issuance n, B_n the consideration divided by CP_n, and C_n the shares issued", "fields": [ { "label": "First issuance", "value": "A = 10,000,000, raise 3,000,000 at 0.5000 so C = 6,000,000 and B = 3,000,000: CP2 = 13,000,000/16,000,000 = 0.8125", "mono": true }, { "label": "Fully diluted after it", "value": "16,923,077, including 4,923,077 shares on conversion of the protected series", "mono": true }, { "label": "Second issuance", "value": "A = 16,923,077, raise 2,000,000 at 0.4000 so C = 5,000,000 and B = 2,000,000/0.8125 = 2,461,538: CP3 = 0.7184", "mono": true }, { "label": "Both at once instead", "value": "A = 10,000,000, B = 5,000,000, C = 11,000,000: CP = 15,000,000/21,000,000 = 0.7143", "mono": true }, { "label": "Cost of sequencing", "value": "0.0041 of conversion price, or 32,234 fewer shares to the protected holder", "mono": true }, { "label": "Full ratchet, either way", "value": "0.4000 and 10,000,000 shares - sequencing is irrelevant because only the last price matters", "mono": true } ], "notes": [ "The direction is counter-intuitive and consistently misread: more triggering events do not accumulate into more protection. Each adjustment dilutes the base that the next one is measured against, so the protection decays as it is used.", "This is the mechanism by which a series that has been repriced twice still ends up holding less than a holder who negotiated a single reset at the final price. Where a series expects a sequence of small down rounds, a stated floor price is worth more than the formula.", "The B term uses the current conversion price, not the original one, so it also shrinks with each adjustment. Both moving parts push the same way.", "Full ratchet is path-independent, which is the only respect in which it is the simpler term. It is also why a ratchet plus a small bridge is a complete reset of the earlier round: see the sensitivity entry above." ], "source": "NVCA model Amended and Restated Certificate of Incorporation (conversion price adjustment provisions)" }, { "id": "broad-base-denominator-composition", "term": "What the broad base actually contains, line by line", "definition": "Broad-based weighted average is named for its denominator, and the denominator is a defined term rather than an accounting concept. Each instrument included or excluded moves the adjustment by a computable amount, so the definitional argument has an exact price.", "formula": "CP2 = CP1*(A + B)/(A + C). A is the defined capitalisation immediately before the issuance; every share added to A reduces the size of the adjustment", "fields": [ { "label": "Typically included in A", "value": "Outstanding common; outstanding preferred on an as-converted basis; options and other rights outstanding, whether vested or not; shares reserved and available under an existing plan", "mono": false }, { "label": "Typically disputed", "value": "Warrants; shares issuable on conversion of outstanding convertible notes and SAFEs; a plan increase approved in connection with the round itself", "mono": false }, { "label": "Worked, A = 10,000,000", "value": "CP2 = 0.8125 - the base case in this corpus", "mono": true }, { "label": "Worked, adding 400,000 warrant shares", "value": "A = 10,400,000: CP2 = 0.8171, a smaller adjustment worth 0.0046 of price to the company", "mono": true }, { "label": "Worked, adding 1,000,000 SAFE conversion shares as well", "value": "A = 11,400,000: CP2 = 0.8276", "mono": true }, { "label": "Narrow base, A = 4,000,000 outstanding preferred only", "value": "CP2 = 0.7000", "mono": true } ], "notes": [ "Every inclusion favours the company and every exclusion favours the protected holder, in a strictly monotone way. That makes the negotiation tractable: compute CP2 under each proposed definition and compare four numbers rather than four arguments.", "The single largest item is the unallocated pool, and it is usually included without discussion. Everything after it - warrants, unconverted instruments, a contemplated increase - moves the price by fractions of a cent on these facts.", "The broad base in the anti-dilution formula is often not the same defined term as the fully diluted count used to price the round. Two definitions in one document is normal; assuming they match is the error.", "The narrow base is not a slightly narrower version of the broad base. It moves CP2 from 0.8125 to 0.7000 here, which is a 30.00 percent price cut instead of an 18.75 percent one, from a single definitional choice." ] }, { "id": "pay-to-play-partial-participation", "term": "Pay-to-play with partial participation", "definition": "A pay-to-play provision is usually drafted proportionally: a holder that funds part of its pro rata keeps that fraction of its preferred and converts the rest to common. The arithmetic is simple and the consequence is not, because the surviving preference is reduced while the shares are not.", "formula": "Preferred retained = s_pref * (amount funded / pro rata requirement). Preference retained = x*I * the same fraction. Shares converted to common = s_pref * (1 - that fraction)", "fields": [ { "label": "Setup", "value": "A holder with 4,000,000 preferred shares carrying a 4,000,000 preference and a 25.000 percent pro rata right, in a round raising 3,000,000", "mono": true }, { "label": "Pro rata requirement", "value": "0.25000 * 3,000,000 = 750,000", "mono": true }, { "label": "Funds 300,000", "value": "Fraction = 300,000/750,000 = 40.000 percent", "mono": true }, { "label": "Result", "value": "Preferred retained 1,600,000 shares carrying 1,600,000 of preference; 2,400,000 shares converted to common", "mono": true }, { "label": "Preference given up", "value": "2,400,000 of a 4,000,000 claim, for 450,000 less than the full pro rata cheque", "mono": true }, { "label": "Marginal price of preference", "value": "450,000 of cash preserved 2,400,000 of preference, a ratio of 5.33 to 1 before any consideration of the shares bought", "mono": true } ], "notes": [ "Partial participation is priced better than it looks in exactly the outcomes where the preference is worth something, and worse than it looks in the outcomes where the company recovers, because the converted shares participate fully in the upside either way. The decision is a view on the distribution, not on the round.", "Check whether the protective provisions and anti-dilution survive partial participation. Some drafting strips them entirely below full participation, in which case the proportional preference is the smaller half of what is being given up.", "A holder that funds nothing loses the whole preference. On these numbers that is a 4,000,000 claim surrendered to avoid a 750,000 cheque, which is why the clause works.", "Also check the denominator of the pro rata requirement. Measured on the whole round it is 750,000; measured only on the amount offered to existing holders it is smaller, and the difference decides whether a holder is in default of the clause." ] }, { "id": "cram-down-worked-end-to-end", "term": "A cram-down recapitalisation, worked end to end", "definition": "A cram-down converts the entire preferred stack to common, reduces the old cap table to a stated residual, and creates a new pool inside the new capitalisation. The reverse split is cosmetic. The three numbers that define the transaction are the residual fraction, the new pool, and the new money.", "formula": "Choose residual v for the old cap table and pool fraction g. Then new money takes 1 - v - g, POST = R/(1 - v - g), the old holders' implied value is v*POST, and the price is R divided by the new money's share count", "fields": [ { "label": "Before", "value": "Fully diluted 24,509,805 shares carrying 60,000,000 of aggregate preference. New money R = 5,000,000. Residual v = 10.000 percent, new pool g = 15.000 percent", "mono": true }, { "label": "Step 1, convert", "value": "All preferred converts to common. The 60,000,000 preference is extinguished and the stack disappears", "mono": true }, { "label": "Step 2, reverse split 20 to 1", "value": "24,509,805 shares become 1,225,490", "mono": true }, { "label": "Step 3, size the new capitalisation", "value": "Old holders are to hold 10.000 percent, so total = 1,225,490/0.10000 = 12,254,900. Pool = 15.000 percent = 1,838,235. New money = 9,191,175 shares", "mono": true }, { "label": "Step 4, price it", "value": "Price = 5,000,000/9,191,175 = 0.5440. POST = 5,000,000/0.75000 = 6,666,667; PRE = 1,666,667", "mono": true }, { "label": "What was traded", "value": "60,000,000 of preference and 32.640 percent founder ownership, for 10.000 percent of a 6,666,667 company - an implied residual value of 666,667", "mono": true }, { "label": "Founders after", "value": "32.640 percent * 0.10000 = 3.264 percent, before any new grant from the 15.000 percent pool", "mono": true } ], "notes": [ "The reverse split ratio is chosen so the resulting share counts look tidy and carries no economic content. Anyone negotiating the ratio rather than the residual fraction is negotiating the wrong number.", "The point of the transaction is the new pool, not the new money. A pool granted behind a 60,000,000 overhang is worth nothing at any exit the company can reach, so no grant retains anyone; the recapitalisation exists to make 15.000 percent of the company worth something to the people who have to operate it.", "Confirm whether the residual is stated before or after the new pool. Ten percent before the pool and ten percent after are different transactions, and the difference here is the whole of the old holders' remaining value.", "Every step needs consents that the old holders control: a charter amendment, usually a protective provision waiver, and the conversion itself. That is why a cram-down is normally led by an existing investor rather than a new one." ] }, { "id": "esop-burn-rate-and-refresh-sizing", "term": "ESOP burn rate and how to size the next refresh", "definition": "A pool is a stock of shares being consumed at a rate. Expressing grants as a monthly burn in shares, and the remaining pool as months of runway, turns the refresh negotiation into the same kind of question as the cash runway - and it is the only sizing argument that survives contact with an investor.", "formula": "Burn = shares granted / months elapsed. Runway = unallocated shares / burn. A refresh covering n months at the current burn, as a fraction of the post-round count, is n*burn/FD_post", "fields": [ { "label": "Setup", "value": "Seed pool 1,411,765 shares = 12.000 percent of the 11,764,706-share ledger. 850,000 shares granted over 18 months", "mono": true }, { "label": "Burn", "value": "850,000/18 = 47,222 shares a month, or 0.4014 percent of the fully diluted count a month", "mono": true }, { "label": "Unallocated", "value": "561,765 shares = 4.775 percent of the ledger", "mono": true }, { "label": "Runway at the current burn", "value": "11.9 months", "mono": true }, { "label": "Same headcount plan at Series A scale", "value": "The plan needs 1,133,333 shares to buy the same percentage of the enlarged company, which is 7.225 percent of the post-Series A count", "mono": true }, { "label": "Refresh for 18 further months", "value": "Round the requirement up to the nearest sensible number and ask for it as a share count with the hiring plan attached, not as a percentage", "mono": false } ], "notes": [ "Burn measured in shares is the wrong unit for a plan, because the same hire costs a different number of shares at every valuation. Burn measured as a percentage of the fully diluted count is the stable figure and is what a refresh has to replace.", "A pool with less runway than the cash has is a real operating constraint, not a paperwork problem: the next senior hire cannot be made without either a board-approved increase or a grant that dilutes the plan's remaining candidates.", "The reverse case matters too. A pool sized far above the plan is economically identical to a lower pre-money valuation, so unspent pool at the next round is money the founders paid for and did not use.", "Grants made from an existing authorised pool need only board approval; an increase generally needs a stockholder vote and often a protective provision consent. The consent calendar, not the arithmetic, is usually what makes a refresh slow." ] }, { "id": "tender-offer-and-the-cap-table", "term": "A tender offer at scale, and what it does and does not change", "definition": "A company-facilitated tender for common shares moves ownership between holders without issuing anything, so the fully diluted count and every other holder's percentage are unchanged. What it does change is the evidence available about the value of common.", "formula": "After a tender for s shares at price P, FD is unchanged, the seller holds (s_seller - s)/FD, the buyer holds s/FD, and every other percentage is identical. Proceeds to sellers = s*P", "fields": [ { "label": "Setup", "value": "Post-Series B ledger, FD = 19,607,844. Founders hold 8,000,000 = 40.800 percent. The Series B price was 5.1000", "mono": true }, { "label": "Tender", "value": "2,000,000 founder shares at 4.5000 = 9,000,000 of proceeds", "mono": true }, { "label": "Discount to the last preferred price", "value": "11.765 percent", "mono": true }, { "label": "After", "value": "Founders 6,000,000 = 30.600 percent; buyer 2,000,000 = 10.200 percent; FD still 19,607,844", "mono": true }, { "label": "Check", "value": "30.600 + 10.200 = 40.800 percent. No other holder moves", "mono": true }, { "label": "Effect on option strikes", "value": "A large arm's length sale of common at 4.5000 is direct evidence of the fair market value of common and will be considered in the next 409A analysis", "mono": true } ], "notes": [ "A tender is the cheapest way to raise the strike price on every option granted afterwards. The transaction itself is free of dilution and not free of cost: employees hired after it pay the new price for the same upside.", "The discount to the preferred price is doing real work in the 409A analysis. A tender priced at the preferred price is much harder to reconcile with a common valuation well below it, which is one reason company-run tenders are usually priced below the last round.", "A tender is a purchase offer to a class of holders and carries securities law consequences that a one-off negotiated transfer does not. The mechanics are cheap; the process is not.", "Model the co-sale and right of first refusal cut-back before agreeing a size with the buyer. A tender is normally run by express waiver precisely because running the gates cannot deliver a fixed size." ] }, { "id": "down-round-with-a-simultaneous-pool-increase", "term": "A pool increase at the same closing shrinks the anti-dilution adjustment", "definition": "If the pool increase is authorised before the new issuance and counts in the anti-dilution base, it enlarges A and therefore reduces the adjustment the protected holder receives. If it is authorised after, it does not. The ordering of two board actions on the same day decides a conversion price.", "formula": "CP2 = CP1*(A + B)/(A + C). Including a pool increase of s_pool in A raises CP2 by CP1*[(A + s_pool + B)/(A + s_pool + C) - (A + B)/(A + C)]", "fields": [ { "label": "Setup", "value": "CP1 = 1.0000, A = 10,000,000, down round of 3,000,000 at 0.5000 so C = 6,000,000 and B = 3,000,000. A 1,000,000-share pool increase is approved at the same closing", "mono": true }, { "label": "Pool excluded from A", "value": "CP2 = 13,000,000/16,000,000 = 0.8125", "mono": true }, { "label": "Pool included in A", "value": "CP2 = 14,000,000/17,000,000 = 0.8235", "mono": true }, { "label": "Cost to the protected holder", "value": "65,934 fewer shares on conversion", "mono": true }, { "label": "Who benefits", "value": "The new investor and the common, in proportion to their post-round holdings. The pool increase is being used twice: once as a pool and once as anti-dilution relief", "mono": true } ], "notes": [ "This is not a drafting error but it is frequently an unnoticed consequence. A down round almost always comes with a new pool, so the interaction arises in nearly every case where anti-dilution is actually triggered.", "The protected holder's response is to ask that the base be measured immediately before the issuance and to define the pool increase as part of the financing rather than as a prior corporate action. One sentence, one conversion price.", "The company's response is the opposite ordering, and it is equally defensible on the drafting. Because both readings are available, this is decided by who raises it, which is a reason to compute it in advance.", "The same question arises for converting bridge notes closing alongside the down round. Whether their conversion shares sit in A moves CP2 by the same mechanism." ] } ], "reviewed": "2026-08-27" }, { "slug": "convertibles", "title": "Convertible instruments", "lede": "SAFEs and convertible notes - how the conversion arithmetic actually resolves.", "intro": "Convertible instruments defer the valuation decision. They do not defer the dilution, and the interaction between a valuation cap, a discount, and the pool can produce founder ownership meaningfully below what the headline terms suggest.", "tables": [ { "title": "Instrument comparison", "columns": [ "Attribute", "Convertible note", "Pre-money SAFE", "Post-money SAFE" ], "rows": [ [ "Legal form", "Debt", "Contractual right to future equity", "Contractual right to future equity" ], [ "Interest", "Typically accrues and converts", "None", "None" ], [ "Maturity", "Yes - creates a hard deadline", "None", "None" ], [ "Ownership certainty for holder", "Depends on the priced round and other converting instruments", "Diluted by other SAFEs converting at the same time", "Fixed percentage of post-money, before the new round" ], [ "Who absorbs dilution from other SAFEs", "Shared", "Shared among all converting holders", "Founders alone" ] ] }, { "title": "Conversion price rules", "columns": [ "Term", "Conversion price" ], "rows": [ [ "Discount only", "Round price * (1 - discount)" ], [ "Cap only", "min(round price, cap / fully diluted pre-money shares)" ], [ "Cap and discount, holder takes the better", "min(round price * (1 - discount), cap / fully diluted pre-money shares)" ], [ "Neither", "Round price - the instrument converts with no economic benefit" ] ] }, { "title": "Pre-money versus post-money SAFE on identical headline terms", "intro": "One SAFE of 1,000,000 at a valuation cap of 8,000,000. Existing fully diluted count before conversion is 8,000,000 shares. The priced round raises R = 3,000,000 at PRE = 12,000,000, so POST = 15,000,000. Nothing differs between the two columns except which capitalisation the cap is measured against.", "columns": [ "Quantity", "Pre-money SAFE", "Post-money SAFE" ], "rows": [ [ "Existing fully diluted shares", "8,000,000", "8,000,000" ], [ "Capitalisation the cap divides by", "8,000,000, excluding SAFE shares", "9,142,857, including SAFE shares" ], [ "SAFE conversion price CP", "1.0000", "0.8750" ], [ "SAFE shares issued", "1,000,000", "1,142,857" ], [ "Pre-money share count for pricing the round", "9,000,000", "9,142,857" ], [ "Round price per share", "1.3333", "1.3125" ], [ "New shares for 3,000,000", "2,250,000", "2,285,714" ], [ "Total shares after close", "11,250,000", "11,428,571" ], [ "SAFE holder ownership", "8.889 percent", "10.000 percent" ], [ "New investor ownership", "20.000 percent", "20.000 percent" ], [ "Founders and existing holders", "71.111 percent", "70.000 percent" ] ] }, { "title": "Convertible note interest and the conversion amount", "intro": "Principal 1,000,000, simple interest at r = 0.06 per annum, converting at CP = 1.0000. The rate is illustrative. The final column is the discount that would produce the same number of shares if the note carried no interest, computed as 1 - 1/(1 + r*t).", "columns": [ "Years to conversion t", "Accrued interest", "Amount converting", "Shares issued", "Equivalent discount" ], "rows": [ [ "0.0", "0", "1,000,000", "1,000,000", "0.00 percent" ], [ "0.5", "30,000", "1,030,000", "1,030,000", "2.91 percent" ], [ "1.0", "60,000", "1,060,000", "1,060,000", "5.66 percent" ], [ "1.5", "90,000", "1,090,000", "1,090,000", "8.26 percent" ], [ "2.0", "120,000", "1,120,000", "1,120,000", "10.71 percent" ], [ "3.0", "180,000", "1,180,000", "1,180,000", "15.25 percent" ] ] }, { "title": "A three-instrument stack solved simultaneously", "intro": "Pre-round fully diluted count FD = 8,000,000. Three instruments convert into the same priced round, which raises R = 3,000,000 at PRE = 12,000,000. A pre-money SAFE takes a fixed share count computed on a capitalisation that excludes conversion shares; a post-money SAFE takes a fixed fraction of the capitalisation after all instruments convert. The two rules are solved together as S = (FD + pre-money shares)/(1 - sum of post-money fractions).", "columns": [ "Instrument", "Amount", "Cap", "Rule", "Conversion price", "Shares", "Percent of the pre-money count", "Percent after the round" ], "rows": [ [ "Pre-money SAFE", "500,000", "10,000,000", "cap / FD", "1.2500", "400,000", "3.869", "3.095" ], [ "Post-money SAFE", "1,000,000", "8,000,000", "I / cap of S", "0.7738", "1,292,308", "12.500", "10.000" ], [ "Post-money SAFE", "750,000", "12,000,000", "I / cap of S", "1.1607", "646,154", "6.250", "5.000" ], [ "Existing holders", "-", "-", "-", "-", "8,000,000", "77.381", "61.905" ], [ "Pre-money count S", "-", "-", "-", "-", "10,338,462", "100.000", "80.000" ], [ "New investor", "3,000,000", "-", "R / round price", "1.1607", "2,584,615", "-", "20.000" ], [ "Total after the round", "-", "-", "-", "-", "12,923,077", "-", "100.000" ] ] }, { "title": "Interest conventions on the same note", "intro": "Principal 1,000,000 at a stated rate of 6 percent per annum, converting eighteen months after issue at a conversion price of 1.0000. Only the accrual convention differs. The equivalent discount is the discount on price that would buy the same number of shares with no interest, computed as 1 - principal divided by the amount converting.", "columns": [ "Convention", "Accrued interest", "Amount converting", "Shares issued", "Equivalent discount" ], "rows": [ [ "Simple, 30/360, 1.500 years", "90,000", "1,090,000", "1,090,000", "8.257" ], [ "Simple, actual/365, 548 days", "90,082", "1,090,082", "1,090,082", "8.264" ], [ "Simple, actual/360, 548 days", "91,333", "1,091,333", "1,091,333", "8.369" ], [ "Compounded quarterly, 6 periods", "93,443", "1,093,443", "1,093,443", "8.546" ], [ "Compounded monthly, 18 periods", "93,929", "1,093,929", "1,093,929", "8.586" ] ] }, { "title": "SAFE, convertible note and priced seed on identical terms", "intro": "One investor puts in 1,000,000 against a pre-round fully diluted count of 8,000,000. The priced seed is done at a pre-money of 8,000,000. The convertible instruments carry a valuation cap of 8,000,000 and convert eighteen months later into a priced round raising 3,000,000 at a pre-money of 12,000,000; the note also carries 6 percent simple interest. Ownership is measured immediately after that priced round, except for the priced seed, which is measured at its own closing and then diluted by the same round.", "columns": [ "Instrument", "Fixed at signing", "Conversion price", "Shares", "Ownership after the priced round", "Downside claim" ], "rows": [ [ "Priced seed preferred", "Price, share count, ownership and a 1x preference", "1.0000", "1,000,000", "8.889", "1,000,000 liquidation preference from day one" ], [ "Pre-money SAFE, 8,000,000 cap", "Nothing - shares depend on the capitalisation at conversion", "1.0000", "1,000,000", "8.889 percent", "Purchase amount, junior to debt, senior to common on dissolution" ], [ "Post-money SAFE, 8,000,000 cap", "Ownership of the pre-new-money capitalisation, 12.500 percent", "0.8750", "1,142,857", "10.000 percent", "Purchase amount, junior to debt, senior to common on dissolution" ], [ "Convertible note, 8,000,000 cap, 6 percent", "A debt claim and a maturity date", "1.0000", "1,090,000", "9.593", "Principal plus interest as a creditor, ahead of all equity" ] ] }, { "title": "The three conversion paths and what each one pays", "intro": "A convertible instrument does not have one exit; it has a list, and each entry pays by a different rule. The figures use the post-money SAFE from the comparison above: 1,000,000 at an 8,000,000 post-money cap on a pre-round count of 8,000,000, giving a fixed 12.500 percent of the capitalisation before new money.", "columns": [ "Path", "Trigger", "What the holder receives", "Worked" ], "rows": [ [ "Equity financing", "A priced round, sometimes only one above a stated size", "Shares at the instrument's conversion price", "Converts at 0.8750 into 1,142,857 shares" ], [ "Liquidity event", "A change of control or an initial public offering before any priced round", "The greater of the purchase amount and the as-converted share of proceeds", "At E = 20,000,000: 0.12500 * 20,000,000 = 2,500,000, taken instead of the 1,000,000 purchase amount" ], [ "Liquidity event, small", "The same trigger at a lower value", "The purchase amount, elected as a cash claim", "At E = 6,000,000: 0.12500 * 6,000,000 = 750,000, so the 1,000,000 purchase amount is elected instead" ], [ "Dissolution", "A wind-up or general assignment for the benefit of creditors", "The purchase amount out of remaining assets, junior to creditors and senior to common", "Indifference between the two elections in a liquidity event is at E equal to the cap, 8,000,000" ] ] } ], "entries": [ { "id": "post-money-safe-shift", "term": "Why the post-money SAFE moved the dilution", "definition": "Under a pre-money SAFE, multiple SAFEs converting in the same round dilute each other, so no holder knows their final percentage until the round closes. A post-money SAFE fixes each holder's percentage of the post-money capitalisation, so subsequent SAFEs no longer dilute earlier ones - founders absorb all of it.", "formula": "Post-money SAFE ownership = investment / valuation cap", "fields": [ { "label": "Worked", "value": "500,000 on an 8,000,000 post-money cap = 6.25 percent, fixed", "mono": true }, { "label": "Stacking effect", "value": "Four such SAFEs = 25.00 percent, all from founders and existing holders", "mono": true } ], "notes": [ "The instrument is clearer and easier to model. It is also strictly more dilutive to founders than the pre-money version at the same headline cap.", "The common failure is raising several post-money SAFEs at different caps over time without maintaining a running total of committed post-money ownership.", "Model the aggregate before signing the second one, not before the priced round." ] }, { "id": "note-maturity", "term": "Convertible note maturity", "definition": "The date at which the note becomes repayable if no qualified financing has occurred. Because a company at that stage usually cannot repay, maturity functions as a renegotiation trigger rather than a repayment date.", "notes": [ "Common outcomes at maturity: extension, conversion at a negotiated valuation, or conversion at the cap by prior agreement.", "The holder's leverage at maturity is the ability to declare default. This is rarely exercised but shapes the negotiation.", "SAFEs were designed specifically to remove this dynamic by having no maturity." ] }, { "id": "qualified-financing", "term": "Qualified financing threshold", "definition": "The minimum size of a priced equity round that triggers automatic conversion of the instrument. Rounds below the threshold do not force conversion.", "notes": [ "Set too high, a genuine priced round fails to trigger conversion and the instruments hang over the cap table.", "Set too low, a small insider round can force conversion at terms the holder would not have chosen.", "Check that the threshold is consistent across every outstanding instrument. Inconsistent thresholds across a stack of notes create a partial conversion that is difficult to model and worse to negotiate." ] }, { "id": "pre-vs-post-money-safe-math", "term": "Pre-money and post-money SAFE conversion arithmetic", "definition": "Both instruments use a valuation cap, and on identical headline terms they produce different share counts, different round prices, and different founder ownership. The difference is which capitalisation the cap is divided by.", "formula": "Pre-money SAFE: CP = cap / FD, where FD excludes the shares issuable on conversion of the SAFEs. Post-money SAFE: ownership = I / cap of the capitalisation after all SAFEs convert and before new money, so CP = cap / S where S = FD / (1 - sum of I_k/cap_k)", "fields": [ { "label": "Setup", "value": "I = 1,000,000, cap = 8,000,000, FD = 8,000,000 shares. Priced round R = 3,000,000 at PRE = 12,000,000", "mono": true }, { "label": "Worked, pre-money", "value": "CP = 8,000,000/8,000,000 = 1.0000 -> 1,000,000 shares. Round price = 12,000,000/9,000,000 = 1.3333 -> 2,250,000 new shares. Total 11,250,000. SAFE holds 8.889 percent", "mono": true }, { "label": "Worked, post-money", "value": "S = 8,000,000/(1 - 0.125) = 9,142,857; CP = 8,000,000/9,142,857 = 0.8750 -> 1,142,857 shares. Round price = 12,000,000/9,142,857 = 1.3125 -> 2,285,714 new shares. Total 11,428,571. SAFE holds 10.000 percent", "mono": true }, { "label": "Who pays the difference", "value": "The new investor holds 20.000 percent either way. Founders and existing holders move from 71.111 to 70.000 percent, so the entire 1.111 point difference comes out of them", "mono": true } ], "notes": [ "The post-money version is easier to model precisely because it fixes the answer in advance, and that is exactly why it costs founders more: certainty for the holder has to be funded by someone, and the only party left is the existing cap table.", "The two instruments are not interchangeable at the same cap. Converting a pre-money term sheet to a post-money document without moving the cap is a real price change, not a documentation cleanup. The cap that leaves the holder in the same position is the pre-money cap plus the amount raised on the instruments: a 9,000,000 post-money cap issues exactly the same 1,000,000 shares as an 8,000,000 pre-money cap on 1,000,000 invested.", "Y Combinator publishes the post-money SAFE in valuation-cap, discount, and most-favoured-nation variants. Check which template is actually attached before modelling anything.", "Both instruments define their own capitalisation term. Read that definition rather than assuming it matches the fully diluted count the term sheet uses elsewhere." ], "source": "Y Combinator SAFE templates (pre-money 2013 series; post-money 2018 series)" }, { "id": "cap-vs-discount-binding", "term": "Valuation cap versus discount - which one binds", "definition": "An instrument carrying both a cap and a discount converts at whichever produces the lower price. Which one that is depends only on the round's valuation, and the switchover point is a single number known at signing.", "formula": "The cap binds when cap / FD < (PRE / FD)*(1 - discount), which reduces to PRE > cap / (1 - discount). Below that valuation the discount binds", "fields": [ { "label": "Worked", "value": "cap = 8,000,000, discount = 0.20: the switchover is PRE = 8,000,000/0.80 = 10,000,000", "mono": true }, { "label": "Above it", "value": "At PRE = 12,000,000 on FD = 8,000,000: cap price 1.0000 against discount price 1.5000*0.80 = 1.2000. The cap binds", "mono": true }, { "label": "Below it", "value": "At PRE = 9,000,000 on the same FD: cap price 1.0000 against discount price 1.1250*0.80 = 0.9000. The discount binds", "mono": true } ], "notes": [ "The switchover is independent of the share count, because FD cancels out. It depends only on the cap and the discount, so it can be computed the day the instrument is signed.", "A discount is protection against a round priced near the cap; a cap is protection against a round priced far above it. An instrument with only a discount has no upside protection at all, and an instrument with only a cap has no protection in a flat or down round.", "Negotiating the discount when the expected round is well above cap / (1 - discount) is negotiating a term that will never apply. Compute the switchover first and spend the negotiation on whichever term will actually bind." ] }, { "id": "mfn-clause", "term": "Most favoured nation", "definition": "An MFN provision lets the holder elect the terms of any later instrument the company issues before conversion. It converts a signed price into a ceiling on price, and it applies retroactively to every MFN instrument outstanding.", "formula": "Effective cap for an MFN holder = min(own cap, caps of all later qualifying instruments issued before conversion). Shares = I / (effective cap / FD)", "fields": [ { "label": "Worked", "value": "A holder signs 1,000,000 at a cap of 8,000,000 on FD = 8,000,000, so CP = 1.0000 and 1,000,000 shares. The company later issues at a 6,000,000 cap", "mono": true }, { "label": "After election", "value": "Effective cap 6,000,000: CP = 6,000,000/8,000,000 = 0.7500, so 1,000,000/0.75 = 1,333,333 shares - a 33.3 percent increase for the same money", "mono": true } ], "notes": [ "A later cheap instrument is more expensive than its own headline suggests, because it reprices every MFN instrument behind it. Before signing a lower cap, total the MFN paper outstanding and reprice it at the new cap.", "MFN is usually one sentence and is often the only term in a small early instrument, since a holder with no cap and no discount has nothing else to rely on.", "The election is typically exercisable at a defined point rather than continuously, and it usually excludes the priced round itself. Whether a later instrument is a qualifying one is a drafting question with a large arithmetic consequence." ], "source": "Y Combinator SAFE (MFN variant)" }, { "id": "pro-rata-side-letter", "term": "Pro-rata rights and what maintaining ownership costs", "definition": "A pro-rata right entitles the holder to buy enough of a future round to keep its percentage flat. The arithmetic is simpler than it looks: buying its own percentage of the new money is exactly sufficient, whatever the round price.", "formula": "To hold ownership constant at q through a round raising R, the holder must invest q*R. Post-round ownership = (q*s_pre + q*s_new)/(s_pre + s_new) = q", "fields": [ { "label": "Worked", "value": "A SAFE holder at 10.000 percent after conversion. The next round raises R = 10,000,000: maintaining 10.000 percent costs 0.10*10,000,000 = 1,000,000", "mono": true }, { "label": "If the holder passes", "value": "Ownership falls to 10.000 percent * PRE/POST. At PRE = 40,000,000, POST = 50,000,000: 10.000 * 0.80 = 8.000 percent", "mono": true } ], "notes": [ "The right is an option struck at the round price, and like any option it is worth most when the round is priced well and worth nothing when it is priced badly. Its value is not the ownership it preserves but the ability to decline.", "The cost to the company is allocation, not cash: every dollar reserved for pro-rata is a dollar the new lead cannot take, which is why leads negotiate pro-rata waivers as a condition of a competitive round.", "The post-money SAFE moved pro-rata out of the instrument and into a separate side letter. If no side letter was signed, the right does not exist regardless of what was discussed." ], "source": "Y Combinator SAFE (post-money) and the accompanying pro-rata side letter" }, { "id": "note-plus-priced-round-interaction", "term": "How converting instruments interact with the priced round", "definition": "Converting instruments are normally placed inside the pre-money share count used to price the round. That single convention decides who absorbs the conversion: the new investor's percentage is fixed by its cheque and the post-money valuation, so every additional converting share dilutes only the existing holders.", "formula": "If converting instruments sit inside the pre-money count, new investor ownership = R/POST regardless of how much converts. Existing holder ownership = s_existing / (FD + converting shares + R/price)", "fields": [ { "label": "Worked", "value": "R = 3,000,000 at PRE = 12,000,000, so R/POST = 3,000,000/15,000,000 = 20.000 percent. The new investor holds 20.000 percent whether the SAFE converts into 1,000,000 shares or 1,142,857", "mono": true }, { "label": "Where it lands", "value": "The extra 142,857 shares move founders and existing holders from 71.111 to 70.000 percent - the whole difference", "mono": true }, { "label": "The general rule", "value": "Anything inside the pre-money count is paid for by the pre-money holders", "mono": true } ], "notes": [ "This is the same mechanism as the option pool shuffle, applied to convertible instruments instead of options. Once the principle is seen in one place it explains most cap table surprises: the pre-money share count is where the cost of everything gets allocated.", "It also means the new investor has no economic reason to police the conversion arithmetic, since its own percentage is protected either way. The founder is the only party with an incentive to check it.", "Ask for the closing cap table with converting instruments shown as a separate line, not folded into the pre-money total. The fold is where errors hide." ] }, { "id": "safe-stacking-overdilution", "term": "SAFE stacking and the over-dilution failure mode", "definition": "Post-money SAFE ownership is additive and fixed, so a series of instruments signed at different caps commits a total percentage that can be computed exactly at any time. The failure mode is not signing a bad instrument; it is signing four reasonable ones without ever adding them up.", "formula": "Committed ownership = sum of I_k/cap_k. Founders and existing holders retain 1 - that sum before new money, and (1 - sum)*(1 - q) after a round selling q", "fields": [ { "label": "Setup", "value": "500,000 at 6,000,000; 750,000 at 8,000,000; 1,000,000 at 12,000,000; 250,000 at 5,000,000. Total raised 2,500,000", "mono": true }, { "label": "Worked", "value": "8.333 + 9.375 + 8.333 + 5.000 = 31.042 percent committed before any priced round", "mono": true }, { "label": "Implied blended cap", "value": "2,500,000/0.31042 = 8,053,691 post-money - well below the 12,000,000 headline of the largest instrument", "mono": true }, { "label": "After a round selling 20 percent", "value": "(1 - 0.31042)*0.80 = 55.167 percent to founders and existing holders", "mono": true } ], "notes": [ "Caps do not average. Ownership is I/cap, which is convex in the cap, so the low-cap instruments dominate the blended result. Raising a small amount at a low cap is disproportionately expensive relative to its size.", "The single number worth maintaining is the running sum of I/cap. It requires no model, no assumptions about the next round, and it is exact. Recompute it before signing each instrument, not before the priced round.", "This clean additivity holds only for post-money instruments. Pre-money SAFEs dilute one another, so summing I/cap across them overstates the committed total - which is a different problem, not a smaller one, because the total is then unknowable until the round closes." ] }, { "id": "mixed-instrument-circularity", "term": "Mixed pre-money and post-money instruments create a circular calculation", "definition": "A post-money instrument takes a fixed percentage of a capitalisation that includes all converting shares. A pre-money instrument takes a fixed share count based on a capitalisation that excludes them. Put both in the same round and each depends on the other.", "formula": "Solve S = FD + sum of shares_j(S) by substitution. With one pre-money instrument issuing a fixed s_pre and one post-money instrument taking fraction w: S = (FD + s_pre)/(1 - w)", "fields": [ { "label": "Setup", "value": "FD = 8,000,000. Pre-money SAFE of 500,000 at a 10,000,000 cap. Post-money SAFE of 1,000,000 at an 8,000,000 cap (w = 0.125)", "mono": true }, { "label": "Worked, pre-money leg", "value": "CP = 10,000,000/8,000,000 = 1.2500 -> 500,000/1.25 = 400,000 shares, fixed", "mono": true }, { "label": "Worked, post-money leg", "value": "S = (8,000,000 + 400,000)/(1 - 0.125) = 9,600,000 -> 0.125*9,600,000 = 1,200,000 shares", "mono": true }, { "label": "Result before new money", "value": "Existing 8,000,000/9,600,000 = 83.333 percent; pre-money SAFE 4.167 percent; post-money SAFE 12.500 percent", "mono": true } ], "notes": [ "The post-money holder's share count rose from 1,142,857 alone to 1,200,000 here, purely because the pre-money instrument enlarged the base it takes a fixed percentage of. It is diluted by nothing and expanded by everything.", "The pre-money holder is diluted by the post-money holder and receives nothing in return. Where both instruments exist, the older paper is systematically worse off, which is the opposite of the usual assumption that earlier money is better priced.", "Add a discount, an MFN election, or a note accruing interest to the same stack and there is no closed form left at all. Solve it numerically and reconcile the share counts against the closing cap table line by line." ] }, { "id": "note-interest-as-discount", "term": "Note interest expressed as a discount", "definition": "Accrued interest buys extra shares at the same conversion price. That is arithmetically identical to a discount on the price, which makes the two terms directly comparable once the holding period is fixed.", "formula": "Equivalent discount = 1 - 1/(1 + r*t) for simple interest. With both interest and a stated discount, total shares = (I*(1 + r*t))/(CP*(1 - discount)), a multiplier of (1 + r*t)/(1 - discount) on the no-terms case", "fields": [ { "label": "Worked", "value": "r = 0.06, t = 1.5: 1 - 1/1.09 = 8.26 percent equivalent discount", "mono": true }, { "label": "Worked, longer and dearer", "value": "r = 0.08, t = 2.0: 1 - 1/1.16 = 13.79 percent", "mono": true }, { "label": "Both terms together", "value": "r = 0.06, t = 1.5, discount 0.20: multiplier = 1.09/0.80 = 1.3625, so 36.25 percent more shares than principal at the round price", "mono": true } ], "notes": [ "Interest and a discount stack multiplicatively, not additively. A note with 6 percent interest held eighteen months and a 20 percent discount is a 26.61 percent effective discount on price, computed as 1 - 1/1.3625, not the 28.26 percent that adding the two terms suggests.", "Interest keeps accruing while the round is being negotiated, so every week of delay is a small transfer from the cap table to the noteholder. On a large note stack that is a real reason to close rather than optimise the last term.", "Some notes convert principal only and pay or waive interest in cash. That is a materially cheaper instrument at the same headline rate, and it is a one-line change to ask for." ] }, { "id": "conversion-below-the-cap", "term": "When the priced round comes in below the cap", "definition": "A valuation cap sets the conversion price by dividing the cap by a capitalisation. It is not a most-favoured price and it is not automatically the lower of the cap price and the round price. If the round prices below the cap, the standard documents still convert the instrument at the cap price.", "formula": "Cap price = cap / S. Round price = PRE / S. The instrument is worse off than new money whenever PRE < cap, by a factor of cap/PRE", "fields": [ { "label": "Setup", "value": "Post-money SAFE of 1,000,000 at an 8,000,000 cap on FD = 8,000,000, so S = 9,142,857 and the cap price is 0.8750", "mono": true }, { "label": "Worked", "value": "The round closes at PRE = 5,000,000: round price = 5,000,000/9,142,857 = 0.5469. The SAFE converts at 0.8750, which is 8,000,000/5,000,000 = 1.60x the price the new investor pays", "mono": true }, { "label": "The general result", "value": "The holder overpays by cap/PRE whenever the round prices below the cap", "mono": true } ], "notes": [ "A cap is a ceiling on price, not a floor on ownership. The instrument protects against the company becoming more valuable and offers nothing at all against it becoming less valuable - which is the outcome the holder was actually exposed to.", "This is the strongest practical argument for an MFN term or an explicit price floor. Both are short to draft and neither costs anything in the outcome where the round prices above the cap.", "Check the specific document rather than assuming. Conversion price definitions vary between templates and between amended instruments, and the difference between 'the SAFE price' and 'the lesser of the SAFE price and the price paid in the equity financing' is the entire question." ] }, { "id": "safe-stack-mixed-caps-simultaneous", "term": "Solving a mixed pre-money and post-money stack in one step", "definition": "Pre-money instruments fix a share count and post-money instruments fix a fraction of a total that includes those shares. With several of each the two rules are simultaneous, but the system is linear and collapses to a single division.", "formula": "S = (FD + sum of fixed pre-money share counts) / (1 - sum of post-money fractions w_k), where w_k = I_k/cap_k. Each post-money instrument then receives w_k*S shares", "fields": [ { "label": "Setup", "value": "FD = 8,000,000. Pre-money SAFE 500,000 at a 10,000,000 cap; post-money SAFEs of 1,000,000 at 8,000,000 and 750,000 at 12,000,000", "mono": true }, { "label": "Pre-money leg", "value": "CP = 10,000,000/8,000,000 = 1.2500, so 400,000 shares, fixed", "mono": true }, { "label": "Post-money fractions", "value": "w = 1,000,000/8,000,000 + 750,000/12,000,000 = 12.500 + 6.250 = 18.750 percent", "mono": true }, { "label": "Solve", "value": "S = (8,000,000 + 400,000)/(1 - 0.187500) = 10,338,462", "mono": true }, { "label": "Shares", "value": "Post-money instruments take 1,292,308 and 646,154. Check: 8,000,000 + 400,000 + 1,292,308 + 646,154 = 10,338,462", "mono": true }, { "label": "After a round of 3,000,000 at PRE 12,000,000", "value": "Round price 1.1607, new shares 2,584,615, total 12,923,077. New investor 20.000 percent, existing holders 61.905 percent", "mono": true } ], "notes": [ "The order of signature is irrelevant to the arithmetic and decisive to the outcome. Every post-money instrument takes its fraction of a base that the pre-money instruments have already enlarged, so the pre-money holders subsidise the post-money holders and receive nothing for it.", "Add a discount, an MFN election or accruing interest and the system stops being linear, because the conversion price then depends on the round price which depends on S. Solve those numerically and reconcile share counts against the closing cap table line by line.", "The single number worth tracking while the instruments are being signed is the running sum of w. It is exact, it needs no forecast of the next round, and it is the fraction of the company already committed.", "Existing holders here end at 61.905 percent having sold 2,250,000 of instruments and 3,000,000 of equity against an 8,000,000-share starting position. Compute that figure before the second instrument, not before the round." ], "source": "Y Combinator SAFE templates (pre-money 2013 series; post-money 2018 series)" }, { "id": "discount-only-in-a-capped-stack", "term": "A discount-only instrument inside a capped stack", "definition": "A discount with no cap prices off the round, so its share count depends on the round price, which depends on the total pre-money count, which includes its own shares. It is the one common instrument whose conversion is genuinely circular even on its own.", "formula": "Shares = I / (discount factor * PRE / S) = I*S/((1 - discount)*PRE), which is a fixed multiple k of S. The stack then solves as S = (FD + fixed pre-money shares)/(1 - sum of w_k - k)", "fields": [ { "label": "Setup", "value": "The three-instrument stack above, plus 250,000 at a 20 percent discount and no cap", "mono": true }, { "label": "The multiple k", "value": "250,000/(0.80000 * 12,000,000) = 0.02604167 shares per share of S", "mono": true }, { "label": "Solve", "value": "S = (8,000,000 + 400,000)/(1 - 0.187500 - 0.026042) = 10,680,795", "mono": true }, { "label": "Discount instrument's shares", "value": "278,146 at a conversion price of 0.8988 against a round price of 1.1235", "mono": true }, { "label": "Effect on the round price", "value": "The round price falls from 1.1607 to 1.1235 because the pre-money count is larger", "mono": true }, { "label": "Effect on existing holders", "value": "From 61.905 to 59.921 percent", "mono": true } ], "notes": [ "A discount is protection against a round priced near where the instrument was signed. It gives nothing at all against a round priced far above, which is precisely the case in which a cap would have been worth something. An instrument with only a discount is unprotected on the upside.", "Because the discount is applied to the round price and the round price is depressed by every converting share, a discount instrument in a large stack converts at a lower absolute price than it would alone. It is the only instrument in the stack that benefits from the others being there.", "The circularity is a real modelling trap: solving for the round price with the discount shares omitted and then adding them produces a price that is too high and a share count that is too low. The closed form above removes the iteration.", "Where the same instrument carries both a cap and a discount, only one binds, and which one is known at signing from PRE against cap/(1 - discount). Compute that before negotiating either term." ] }, { "id": "note-interest-conventions", "term": "Interest conventions change the share count without changing the rate", "definition": "A note's stated rate is not sufficient to compute what converts. The day-count basis, the compounding frequency and whether interest converts or is paid in cash each move the amount, and on an eighteen-month hold the spread between the cheapest and dearest of the common conventions is 3,929 on 90,000 of interest, or 4.365 percent of it.", "formula": "Amount converting = principal * (1 + f), where f is the accrual factor: r*t for simple accrual with t on the stated basis, or (1 + r/n)^(n*t) - 1 for compounding n times a year", "fields": [ { "label": "Simple, 30/360, 1.500 years", "value": "Interest 90,000, amount 1,090,000, equivalent discount 8.257", "mono": true }, { "label": "Simple, actual/360, 548 days", "value": "Interest 91,333, amount 1,091,333", "mono": true }, { "label": "Compounded monthly", "value": "Interest 93,929, amount 1,093,929", "mono": true }, { "label": "Spread across the conventions", "value": "3,929 of extra shares at a conversion price of 1.0000, on the same stated 6 percent", "mono": true }, { "label": "Interest plus a 20 percent discount", "value": "Multiplier = 1.09000/0.80000 = 1.3625, so 36.250 percent more shares than principal at the round price", "mono": true } ], "notes": [ "Actual/360 is the convention that quietly pays the most, because it divides a real number of days by a 360-day year. On a stated 6 percent it delivers an effective 6.083 percent over a 365-day year, computed as 0.06 * 365/360, and nobody negotiating the rate mentions it.", "Interest accrues while the round is being documented, so the conversion amount is a function of the closing date. On a large note stack that makes speed worth more than the last term being argued over.", "Notes that pay or waive interest in cash rather than converting it are materially cheaper instruments at the same headline rate, and asking for that is a one-line change. It converts a dilution cost into a cash cost at a moment when the company has just raised cash.", "Reconcile the conversion amount with the noteholder in writing before the closing cap table is built. A one-day disagreement about the accrual end date is a share count disagreement, and share counts are what get filed." ] }, { "id": "maturity-and-conversion-triggers", "term": "The trigger set, not the cap, decides when an instrument converts", "definition": "A convertible instrument specifies the events on which it converts and the rule that applies to each. The valuation cap is only the price used on one of those paths. Two instruments with the same cap and different trigger sets are different securities.", "formula": "No single formula. Each path has its own rule: an equity financing converts at the instrument's conversion price, a liquidity event pays the greater of the purchase amount and the as-converted proceeds, a dissolution pays the purchase amount out of remaining assets", "fields": [ { "label": "Qualified financing threshold", "value": "A minimum round size above which conversion is automatic. Below it, conversion is usually optional or does not occur", "mono": false }, { "label": "Maturity", "value": "Present in notes and absent from SAFEs. At maturity the note is repayable, which in practice makes it a renegotiation date", "mono": false }, { "label": "Liquidity event", "value": "A change of control or public offering before any priced round. The holder elects the better of a cash claim and conversion", "mono": false }, { "label": "Dissolution", "value": "A wind-up. The holder ranks behind creditors and ahead of common for the purchase amount", "mono": false }, { "label": "Worked, the liquidity election", "value": "A 1,000,000 post-money SAFE at an 8,000,000 cap holds 12.500 percent. At E = 20,000,000 conversion pays 2,500,000; at E = 6,000,000 the 1,000,000 cash claim is better. Indifference is at E equal to the cap", "mono": true } ], "notes": [ "The indifference point in a liquidity event is exactly the cap, and the algebra is one line: conversion pays (I/cap)*E, which equals I when E equals cap. That gives the holder a clean rule and gives the founder a clean warning about what a small early sale actually costs.", "An inconsistent threshold across a stack of instruments produces a partial conversion, where some instruments convert and others hang over the new cap table with their caps intact. That is difficult to model and worse to negotiate, and it is avoidable by using one threshold.", "Maturity is the only trigger that arrives on a date rather than on an event, which is why it is the only one that reliably forces a conversation. SAFEs removed it deliberately; that is a real transfer of leverage to the company and the main structural difference from a note.", "Read whether conversion on an equity financing is automatic or at the holder's election. An election is worth something to the holder in any round priced below the cap, which is exactly the round in which the company most wants the instrument gone." ] }, { "id": "safe-note-priced-seed-comparison", "term": "The same 1,000,000 as a priced round, a SAFE and a note", "definition": "On identical headline terms the three instruments produce three ownership figures, three downside claims and three sets of rights. The differences are not a matter of documentation cost; they are economic, and they are computable at signing.", "formula": "Priced round: shares = I/(PRE/FD), and ownership is then diluted by later rounds. Pre-money SAFE: shares = I/(cap/FD). Post-money SAFE: ownership = I/cap of the pre-new-money capitalisation. Note: shares = I*(1 + r*t)/(cap/FD)", "fields": [ { "label": "Priced seed", "value": "PRE 8,000,000 on FD 8,000,000: price 1.0000, 1,000,000 shares, 11.111 percent at closing, 8.889 percent after the next round", "mono": true }, { "label": "Pre-money SAFE at an 8,000,000 cap", "value": "CP 1.0000, 1,000,000 shares, 8.889 percent after the priced round", "mono": true }, { "label": "Post-money SAFE at an 8,000,000 cap", "value": "CP 0.8750, 1,142,857 shares, 10.000 percent after the priced round", "mono": true }, { "label": "Note at an 8,000,000 cap, 6 percent simple, 1.5 years", "value": "Converting amount 1,090,000 at CP 1.0000, 1,090,000 shares, 9.593 percent after the priced round", "mono": true }, { "label": "Downside", "value": "Only the note is a creditor claim. Both SAFEs rank behind debt and ahead of common. The priced preferred carries a 1,000,000 liquidation preference from day one", "mono": true }, { "label": "Rights", "value": "Only the priced round delivers protective provisions, a board seat if negotiated, information rights and anti-dilution at the time the money goes in", "mono": true } ], "notes": [ "The priced seed is the least dilutive of the four on these numbers and the most expensive to document. That trade - legal cost and a real negotiation now against a larger share of the company later - is the actual decision, and it is usually made on the wrong axis.", "The note's interest is the smallest term in the comparison and the maturity date is the largest. A note that matures before the next round is likely to be renegotiated at a moment of weakness, and no cap protects against that.", "The post-money SAFE is the most dilutive at the same cap because it fixes the holder's percentage, which someone has to fund. The equivalent pre-money cap is the post-money cap less the amount raised on the instruments.", "None of these figures is a market observation. They are the arithmetic consequences of one set of stated terms, chosen so the four instruments can be compared on the same facts." ], "source": "Y Combinator SAFE templates and the NVCA model financing documents" }, { "id": "safe-overhang-at-the-priced-round", "term": "The overhang that only becomes visible at the priced round", "definition": "Instruments signed one at a time commit a total percentage that is never reported anywhere until the round closes and the closing cap table is built. At that point the conversion shares appear inside the pre-money count and the founders discover the arithmetic all at once.", "formula": "Conversion shares as a fraction of the pre-money count = (S - FD)/S, where S is the solved pre-money count. Founders' ownership after the round = FD_founders/S * (1 - q)", "fields": [ { "label": "Setup", "value": "The three-instrument stack: 2,250,000 raised on instruments against an 8,000,000-share starting count", "mono": true }, { "label": "Conversion shares", "value": "2,338,462 shares, which is 22.619 percent of the 10,338,462-share pre-money count", "mono": true }, { "label": "What the founders modelled", "value": "2,250,000 raised against an 8,000,000 pre-money-equivalent, which reads as roughly 22 percent", "mono": true }, { "label": "What the round delivers", "value": "Existing holders 61.905 percent after a round selling 20.000 percent", "mono": true }, { "label": "Where the difference goes", "value": "Nothing is lost - the instruments hold 18.095 percent between them, and the new investor still holds exactly R/POST", "mono": true } ], "notes": [ "The new investor's percentage is fixed by its cheque and the post-money valuation, so it is indifferent to how much converts. That leaves the founder as the only party with an economic reason to check the conversion arithmetic, and usually the only party without a model.", "The overhang is not a surprise in the sense of being hidden. Every input is in signed documents. It is a surprise in the sense that no single document states the total, and nobody is required to compute it before the round.", "The running total to maintain is the sum of I/cap for post-money instruments plus the fixed share counts of pre-money instruments. Two numbers, updated on each signature, and the whole thing is knowable in advance.", "Where the total is already uncomfortable, the fix is to raise the priced round earlier rather than to raise another instrument. Each additional instrument is priced against a cap set before the last one, so the cheapest capital available is also the most dilutive." ] }, { "id": "mfn-cascade-three-instruments", "term": "An MFN cascade across three instruments", "definition": "Most favoured nation clauses apply to every MFN instrument outstanding, so a single cheap instrument reprices all of them at once. With three instruments the committed ownership is not the sum of the signed terms; it is the sum computed at the lowest cap any of them can elect.", "formula": "Effective cap for every MFN holder = min over all qualifying instruments issued before conversion. Committed ownership = sum of I_k / effective cap", "fields": [ { "label": "Setup", "value": "Instrument one 300,000, MFN only with no cap. Instrument two 500,000 at a 10,000,000 post-money cap. Instrument three 400,000 at a 6,000,000 post-money cap. All three carry MFN", "mono": true }, { "label": "On their own terms", "value": "Instrument two 5.000 percent, instrument three 6.667 percent, instrument one has no fixed percentage at all. Committed 11.667 percent", "mono": true }, { "label": "After the elections", "value": "All three elect the 6,000,000 cap: 5.000 + 8.333 + 6.667 = 20.000 percent", "mono": true }, { "label": "What the last 400,000 cost", "value": "It committed 6.667 percent on its own terms and 8.333 further percentage points by repricing the paper behind it", "mono": true }, { "label": "Effective blended cap", "value": "1,200,000 committed for 20.000 percent is a blended post-money cap of 6,000,000. The two instruments that did carry a cap average 8,222,222 weighted by amount, computed as (500,000*10,000,000 + 400,000*6,000,000)/900,000, and the third carried no cap at all to average in - so no weighted average of the headline terms produces the committed figure", "mono": true } ], "notes": [ "Total the MFN paper outstanding and reprice it at the proposed new cap before signing a lower one. The marginal cost of a small cheap instrument is its own dilution plus the repricing of everything behind it, and the second term is usually the larger.", "MFN is often the only economic term in a small early instrument, because a holder with no cap and no discount has nothing else. That makes it easy to concede and easy to forget, and it is exactly the term that compounds.", "Whether a later instrument is a qualifying one, and whether the election is exercisable continuously or only at a defined point, are drafting questions with large arithmetic consequences. The priced round itself is usually excluded.", "An uncapped MFN instrument is the strangest object in the stack: it has no committed percentage until someone else sets one for it. Its cost is entirely a function of the cheapest thing signed afterwards." ], "source": "Y Combinator SAFE (MFN variant)" }, { "id": "safe-in-a-liquidity-event", "term": "A SAFE in a sale before any priced round", "definition": "If the company is sold before a priced round, the instrument never converts at a cap. The holder elects between a cash claim for its money back and its as-converted share of the proceeds. The switchover is at an exit value exactly equal to the cap.", "formula": "Holder receives max(I, (I/cap)*E) for a post-money instrument. The two are equal when E = cap", "fields": [ { "label": "Setup", "value": "1,000,000 post-money SAFE at an 8,000,000 cap on a pre-round count of 8,000,000, so S = 9,142,857 and the holder's fixed share is 12.500 percent", "mono": true }, { "label": "At E = 4,000,000", "value": "Conversion pays 500,000; the cash claim pays 1,000,000. Cash claim elected", "mono": true }, { "label": "At E = 8,000,000", "value": "Both pay 1,000,000. Indifference, at E equal to the cap", "mono": true }, { "label": "At E = 20,000,000", "value": "Conversion pays 2,500,000; the cash claim pays 1,000,000. Conversion elected", "mono": true }, { "label": "What common receives at E = 4,000,000", "value": "4,000,000 less the 1,000,000 cash claim = 3,000,000 across 8,000,000 shares, or 0.3750 a share", "mono": true }, { "label": "Ranking", "value": "Behind all indebtedness, ahead of common. The instrument is not stock and does not vote", "mono": true } ], "notes": [ "The election is worth exactly max(0, I - (I/cap)*E), which is largest at the smallest exits. That makes an early acquisition offer below the cap a transfer from common to the instrument holders, and it is the outcome founders least often model.", "Some documents pay the purchase amount and some pay a multiple of it on a liquidity event. A 1x cash claim and a 1.5x cash claim are meaningfully different instruments in a small sale, and the term is short enough to be missed.", "The cash claim is not a liquidation preference in the charter sense, because no stock has been issued. It is a contractual payment obligation, which is why it ranks behind indebtedness rather than alongside the preferred.", "In a sale before any priced round the instruments are the whole capital structure other than debt. Run the waterfall on the instrument terms rather than assuming the common takes everything above the debt." ], "source": "Y Combinator SAFE (post-money), Liquidity Event provisions" }, { "id": "pre-money-safes-dilute-each-other", "term": "Three pre-money SAFEs at the same cap deliver less than the sum of their parts", "definition": "Post-money ownership is additive: the sum of I/cap is exactly what the holders end up with. Pre-money ownership is not, because each instrument's shares enlarge the denominator that the others are measured against. Summing I/cap across pre-money instruments overstates the result.", "formula": "Pre-money instruments each take a fixed s_k = I_k/(cap_k/FD). Combined ownership before new money = sum of s_k / (FD + sum of s_k), which is strictly less than sum of I_k/cap_k", "fields": [ { "label": "Setup", "value": "Three pre-money SAFEs of 500,000 each at an 8,000,000 cap, on FD = 8,000,000", "mono": true }, { "label": "Each instrument", "value": "CP = 1.0000, so 500,000 shares each and 1,500,000 in total", "mono": true }, { "label": "Combined ownership before new money", "value": "1,500,000/9,500,000 = 15.789 percent", "mono": true }, { "label": "Naive sum of I/cap", "value": "3 * 6.250 = 18.750 percent", "mono": true }, { "label": "Overstatement", "value": "2.961 percentage points, and it grows with the number of instruments", "mono": true }, { "label": "The same 1,500,000 as post-money instruments", "value": "Each takes exactly 6.250 percent, total 18.750 percent, and the founders fund the difference", "mono": true } ], "notes": [ "This is not a point in favour of pre-money instruments. It means the holders are worse off than their own arithmetic suggests, and that they cannot know their percentage until the round closes. Uncertainty for the holder is the price of the smaller total.", "It also means the sum of I/cap is the wrong running total for a pre-money stack. The right one is the sum of the fixed share counts, which is exact and is knowable on each signature.", "Mixing the two conventions produces the worst of both: the pre-money holders are diluted by the post-money holders and receive nothing in return, so the older paper is systematically worse off.", "The post-money cap that leaves a pre-money holder in the same position is the pre-money cap plus the total raised on the instruments. On these facts a 9,500,000 post-money cap issues the same 500,000 shares as an 8,000,000 pre-money cap." ] } ], "reviewed": "2026-08-27" }, { "slug": "valuation", "title": "Valuation and pricing", "lede": "Pre-money, post-money, price per share, and the two valuations that are never the same number.", "intro": "A priced round has exactly three independent numbers: the pre-money valuation, the amount raised, and the pre-round fully diluted share count. Everything else - post-money valuation, price per share, new shares issued, ownership percentages, dilution - is derived from those three by arithmetic with no judgement in it. Most disagreement about valuation is disagreement about which of the three is being held fixed, or about a fourth number that is not a valuation at all.", "tables": [ { "title": "Solving the round from any two inputs", "intro": "R is the amount raised, q the investor's target ownership as a decimal, and FD the pre-round fully diluted share count.", "columns": [ "Known", "Derived", "Formula" ], "rows": [ [ "PRE and R", "POST, investor ownership", "POST = PRE + R; q = R/POST" ], [ "POST and R", "PRE, investor ownership", "PRE = POST - R; q = R/POST" ], [ "q and R", "POST, PRE", "POST = R/q; PRE = R/q - R" ], [ "q and PRE", "R, POST", "R = PRE*q/(1 - q); POST = PRE/(1 - q)" ], [ "PRE and FD", "price per share", "price = PRE/FD" ], [ "price and R", "new shares issued", "new shares = R/price" ], [ "PRE and POST", "dilution to existing holders", "1 - PRE/POST = R/POST" ] ] }, { "title": "The same round expressed nine ways", "intro": "One round: 2,000,000 raised for 20.000 percent, with a 1,411,765-share option pool created inside the pre-money count on a base of 8,000,000 founder shares. Every line below is derived from PRE = 8,000,000, R = 2,000,000 and FD = 9,411,765.", "columns": [ "Expression", "Value" ], "rows": [ [ "Amount raised R", "2,000,000" ], [ "Investor ownership q", "20.000 percent" ], [ "Post-money valuation POST", "10,000,000" ], [ "Pre-money valuation PRE", "8,000,000" ], [ "Pre-round fully diluted shares FD", "9,411,765" ], [ "Price per share", "0.8500" ], [ "New shares issued", "2,352,941" ], [ "Post-round fully diluted shares", "11,764,706" ], [ "Dilution to existing holders", "20.000 percent" ], [ "Retained fraction for existing holders", "80.000 percent" ], [ "Founder shares valued at the round price", "6,800,000" ] ] }, { "title": "Round by round: valuation, price and step-up", "intro": "The four rounds of the share ledger. The step-up column is this round's pre-money against the previous round's post-money; the price step is this round's price per share against the previous round's. The two are identical here because nothing but the round itself is issued at each closing.", "columns": [ "Round", "PRE", "R", "POST", "Pre-round FD", "Price per share", "Shares issued", "Post-round FD", "Investor ownership", "Step-up on prior POST", "Price step" ], "rows": [ [ "Seed", "8,000,000", "2,000,000", "10,000,000", "9,411,765", "0.8500", "2,352,941", "11,764,706", "20.000", "-", "-" ], [ "Series A", "24,000,000", "8,000,000", "32,000,000", "11,764,706", "2.0400", "3,921,569", "15,686,275", "25.000", "2.40", "2.40" ], [ "Series B", "80,000,000", "20,000,000", "100,000,000", "15,686,275", "5.1000", "3,921,569", "19,607,844", "20.000", "2.50", "2.50" ], [ "Series C", "120,000,000", "30,000,000", "150,000,000", "19,607,844", "6.1200", "4,901,961", "24,509,805", "20.000", "1.20", "1.20" ] ] }, { "title": "Entry ownership, exit ownership and return multiple", "intro": "The same ledger, held to an exit at E = 250,000,000, above every conversion flip point so that all preferred converts and each holder receives p*E. Entry ownership is measured immediately after the holder's own round; exit ownership after Series C. The retained fraction is the product of the later rounds' PRE/POST ratios.", "columns": [ "Holder", "Invested I", "Ownership at entry", "Ownership at exit", "Retained fraction", "Proceeds at E = 250,000,000", "Multiple on invested capital" ], "rows": [ [ "Seed", "2,000,000", "20.000", "9.600", "0.4800", "24,000,000", "12.00" ], [ "Series A", "8,000,000", "25.000", "16.000", "0.6400", "40,000,000", "5.00" ], [ "Series B", "20,000,000", "20.000", "16.000", "0.8000", "40,000,000", "2.00" ], [ "Series C", "30,000,000", "20.000", "20.000", "1.0000", "50,000,000", "1.67" ], [ "Founders", "-", "100.000", "32.640", "0.3264", "81,600,000", "-" ], [ "Option pool", "-", "12.000", "5.760", "0.4800", "14,400,002", "-" ] ] } ], "entries": [ { "id": "post-money-identity", "term": "The post-money identity", "definition": "Post-money valuation is pre-money valuation plus the money. That single identity generates the investor's ownership, the existing holders' retention, and the dilution, and it holds exactly as long as nothing else is issued in the same transaction.", "formula": "POST = PRE + R; investor ownership q = R/POST; existing holders retain PRE/POST; dilution = 1 - PRE/POST = R/POST", "fields": [ { "label": "Worked", "value": "PRE = 8,000,000 and R = 2,000,000: POST = 10,000,000; q = 20.000 percent; existing holders retain 80.000 percent", "mono": true }, { "label": "Solving backwards", "value": "An investor targeting q = 20.000 percent with R = 2,000,000 is proposing POST = 10,000,000, hence PRE = 8,000,000", "mono": true } ], "notes": [ "The identity fails the moment anything else is issued at the same closing - a new option pool inside the pre-money count, or converting SAFEs and notes. In those rounds the investor still holds R/POST, but the existing holders retain materially less than PRE/POST.", "That is the single most useful diagnostic on a term sheet. Compute PRE/POST, compare it to the founder ownership in the closing cap table, and the gap is the total cost of everything else being issued in the round.", "Because the identity is exact, any two of PRE, POST and R determine the third. A negotiation conducted in two of them and then reopened on the third is a negotiation reopened on nothing." ] }, { "id": "ownership-algebra", "term": "Solving for the pre-money from an ownership target", "definition": "Investors with an ownership target and a cheque size are not proposing a pre-money valuation - they are solving for one. Reading the pre-money as the primary variable inverts the causality and makes the negotiation harder than it is.", "formula": "PRE = R*(1 - q)/q and POST = R/q. Equivalently, for a given PRE, R = PRE*q/(1 - q)", "fields": [ { "label": "Worked, q = 0.20", "value": "R = 2,000,000: POST = 2,000,000/0.20 = 10,000,000; PRE = 10,000,000 - 2,000,000 = 8,000,000", "mono": true }, { "label": "Worked, q = 0.15", "value": "Same cheque: POST = 2,000,000/0.15 = 13,333,333; PRE = 11,333,333", "mono": true }, { "label": "Cross-check", "value": "PRE = 2,000,000*0.85/0.15 = 11,333,333", "mono": true }, { "label": "Sensitivity", "value": "Moving the ownership target 5 points moves the pre-money by 3,333,333 on an unchanged cheque", "mono": true } ], "notes": [ "PRE is hyperbolic in q, so ownership concessions are much more expensive at low q than at high q. Going from 25 to 20 percent on a 2,000,000 cheque moves PRE by 2,000,000; going from 15 to 10 percent moves it by 6,666,667.", "The productive lever is usually the cheque size, not the valuation. Raising less at the same ownership target is the same economic outcome as raising the valuation, and it is a much shorter conversation.", "Any argument about pre-money valuation that does not name the fully diluted denominator is incomplete, because the same PRE at a larger FD is a lower price per share." ] }, { "id": "price-per-share-invariance", "term": "Why the absolute share count is arbitrary", "definition": "Multiplying every share count by the same factor changes no ownership percentage, no payout, and no preference amount. It divides the price per share by that factor and nothing else. Price per share therefore carries no information on its own.", "formula": "Multiplying all share counts by k leaves every ownership percentage and every waterfall payout unchanged, and sets price = PRE/(k*FD)", "fields": [ { "label": "Worked", "value": "PRE = 8,000,000 on FD = 9,411,765 gives 0.8500. After a 10:1 forward split, FD = 94,117,650 and the price is 0.0850", "mono": true }, { "label": "Unchanged", "value": "Ownership percentages, exit proceeds, preference amounts, and every formula in this corpus that is expressed in p rather than in shares", "mono": true } ], "notes": [ "The count is chosen for two practical reasons only: a strike price granular enough to grant small slices to employees, and a per-share number that reads as normal to them. Neither is an economic term.", "It follows that a comparison of price per share between two companies, or between two rounds with different share counts, is meaningless. Compare valuations and ownership.", "Splits do matter to thresholds expressed as absolute share numbers - major-holder definitions, board-appointment minimums, protective-provision minimums. Well-drafted documents restate those after any split; check that they were." ] }, { "id": "safe-cap-is-not-a-valuation", "term": "A SAFE cap is not a valuation", "definition": "A valuation cap determines how many shares a given amount will buy when a priced round eventually happens. It does not price a share, does not establish a post-money valuation, and does not make the company worth the cap at any point.", "formula": "A post-money cap fixes ownership = I/cap of the capitalisation before new money. The company's implied post-money at conversion is the round price multiplied by the post-round share count, which can be above or below the cap", "fields": [ { "label": "Worked", "value": "I = 1,000,000 at an 8,000,000 post-money cap fixes 12.500 percent of the pre-new-money capitalisation", "mono": true }, { "label": "What the round actually says", "value": "The priced round closes at PRE = 12,000,000 and R = 3,000,000, so POST = 15,000,000. The company was never valued at 8,000,000 on any date", "mono": true }, { "label": "What the holder ends with", "value": "1,142,857 shares at a round price of 1.3125 = 1,500,000 of value on 1,000,000 invested, a 1.50x", "mono": true } ], "notes": [ "No security is issued when the instrument is signed, so there is no price and no valuation event. Describing a capped instrument as a raise 'at an 8,000,000 valuation' is a category error that then propagates into every subsequent model.", "A SAFE is not a priced preferred issuance, so it does not supply the evidence of fair market value that a closed preferred round does. Expecting it to reset a 409A analysis on its own is a common and expensive misreading.", "The cap is a ceiling on the conversion price, not a floor on ownership and not a floor on price. In a round priced below the cap the instrument converts at a worse price than the new money pays." ] }, { "id": "common-vs-preferred-price-gap", "term": "The 409A gap between common and preferred", "definition": "Option strike prices are set to the fair market value of common stock, determined under section 409A. Preferred stock carries a liquidation preference, protective provisions and other rights that common does not, so the fair market value of common is set below the price paid for preferred in the same round.", "formula": "Discount to the preferred price = 1 - (common fair market value / preferred price). The level is an appraisal output, not a rule", "fields": [ { "label": "Worked", "value": "Preferred price 0.8500 with a common fair market value of 0.2550: discount = 1 - 0.30 = 70.0 percent", "mono": true }, { "label": "Worked, a narrower gap", "value": "Same preferred price with a common value of 0.5100: discount = 1 - 0.60 = 40.0 percent", "mono": true }, { "label": "What drives it", "value": "The size of the aggregate preference relative to enterprise value, the probability weighting across exit scenarios, and marketability", "mono": true } ], "notes": [ "The gap is a direct consequence of the preference overhang, so it narrows as the company grows into its stack and widens when a large new preference is added. It is the same arithmetic as the dead zone, expressed as a valuation input rather than as a payout.", "Timing matters more than most grantees realise: a grant made before a 409A refresh that follows a strong round carries a lower strike than the identical grant made after it. That is a scheduling decision, not a compensation decision.", "Two different numbers are correct at the same time - the preferred price and the common fair market value - and quoting either as 'the' share price will mislead somebody. Always say which one.", "The appraisal is an independent determination. Do not model a target discount and back into the value; the direction of the arithmetic is the point of the rule." ], "source": "IRC section 409A" }, { "id": "valuation-is-not-proceeds", "term": "Ownership percentage is not a claim on the exit price", "definition": "Multiplying ownership by exit value gives the right answer only above the conversion indifference point. Below it, the preference comes off the top first and the residual is shared among a smaller group, so the naive calculation overstates common proceeds.", "formula": "Founder proceeds = (E - preferences of non-converting series - carve-out) * s_founder/(s_common + s_converting), which equals p*E only for E >= E* = x*I/p", "fields": [ { "label": "Setup", "value": "Post-round FD 11,764,706 shares: founders 8,000,000 (68.000 percent), pool 1,411,765 (12.000 percent), preferred 2,352,941 (20.000 percent) carrying 1x on 2,000,000", "mono": true }, { "label": "Worked at E = 5,000,000", "value": "Preferred takes max(2,000,000, 0.20*5,000,000) = 2,000,000. Residual 3,000,000 to 9,411,765 common and pool shares: founders 8,000,000/9,411,765 * 3,000,000 = 2,550,000; pool 450,000", "mono": true }, { "label": "The naive figure", "value": "0.68 * 5,000,000 = 3,400,000, overstating founder proceeds by 850,000 - 25.0 percent too high", "mono": true }, { "label": "Where it becomes correct", "value": "E* = 2,000,000/0.20 = 10,000,000; at and above that the preferred converts and ownership times exit value is exact", "mono": true } ], "notes": [ "The error is largest in the exit range that is most likely, which is what makes it consequential. Above E* the naive calculation is exactly right and nobody needs it; below E* it is wrong and everybody uses it.", "The residual is shared with the option pool, including unallocated shares in many structures. Founders modelling their own outcome frequently divide by the common they know about rather than by the full residual group.", "Run any ownership figure through the waterfall at three exit values - below the aggregate preference, between it and E*, and above E* - before treating the percentage as a number that means anything." ] }, { "id": "implied-step-up-and-price-step", "term": "Step-up and price step are the same number only when nothing else is issued", "definition": "A round's step-up is usually quoted as this pre-money against the last post-money. The number that actually determines whether existing holders gained is the price per share. They agree only when no shares beyond the round itself were issued in between.", "formula": "Step-up = PRE_n / POST_(n-1). Price step = (PRE_n/FD_(n-1)) / (PRE_(n-1)/FD_(n-2)). The two are equal when FD_(n-1) = FD_(n-2) + shares issued in round n-1 and nothing else", "fields": [ { "label": "Series A", "value": "PRE 24,000,000 on a prior POST of 10,000,000: step-up 2.40. Price 0.8500 to 2.0400: price step 2.40", "mono": true }, { "label": "Series B", "value": "PRE 80,000,000 on a prior POST of 32,000,000: step-up 2.50. Price 2.0400 to 5.1000: 2.50", "mono": true }, { "label": "Series C", "value": "PRE 120,000,000 on a prior POST of 100,000,000: step-up 1.20. Price 5.1000 to 6.1200: 1.20", "mono": true }, { "label": "With a 5 percent pool refresh at Series A", "value": "The pre-money count used to price rises to 12,605,042, so the price is 1.9040 and the price step is 2.24 against an unchanged 2.40 step-up", "mono": true }, { "label": "The wedge", "value": "Any share issued between the two rounds - a pool refresh, a converting instrument, a warrant exercise - makes the price step smaller than the step-up, and only the price step is felt by existing holders", "mono": true } ], "notes": [ "A round can be announced as a 2.4x step-up while the price per share rises by less, and both statements are true. Ask for the price per share; it is the only figure that is directly comparable between rounds on the same cap table.", "The reverse is possible too. A buyback or a reverse split raises the price without raising the valuation, which is why price per share cannot be compared between companies at all.", "Step-ups reported across a portfolio are almost always pre-money against prior post-money, because that is the pair available from public announcements. Treat them as a valuation series and not as a return series.", "For an existing holder the only figures that matter are its own retention factor PRE/POST and the price at which any new shares were sold. Both are in the closing documents and neither is in the announcement." ] }, { "id": "post-money-cap-equivalent-across-a-stack", "term": "The pre-money cap that is equivalent to a post-money cap", "definition": "A pre-money cap and a post-money cap are the same term measured against different capitalisations, and the conversion between them is exact. For one instrument the equivalent pre-money cap is simply the post-money cap less the amount raised. For a stack it is a single expression in the committed fraction.", "formula": "For a stack of post-money instruments with committed fraction w = sum of I_k/cap_k and total raised sum I_k, the pre-money cap issuing the same total share count is cap_pre = (sum I_k)*(1 - w)/w", "fields": [ { "label": "One instrument", "value": "1,000,000 at an 8,000,000 post-money cap: cap_pre = 1,000,000*0.87500/0.12500 = 7,000,000, which is the post-money cap less the amount raised", "mono": true }, { "label": "A two-instrument stack", "value": "1,000,000 at 8,000,000 and 750,000 at 12,000,000: w = 18.750 percent on 1,750,000 raised", "mono": true }, { "label": "Equivalent pre-money cap", "value": "1,750,000 * 0.81250/0.18750 = 7,583,333", "mono": true }, { "label": "Check, post-money route", "value": "S = 8,000,000/(1 - 0.18750) = 9,846,154, and w*S = 1,846,154 conversion shares", "mono": true }, { "label": "Check, pre-money route", "value": "1,750,000 * 8,000,000/7,583,333 = 1,846,154 conversion shares", "mono": true } ], "notes": [ "Converting a pre-money term sheet onto a post-money template without moving the cap is a price change, not a documentation cleanup. The difference is the amount raised on the instruments, which is a number both sides already know.", "The formula shows why the gap widens as more is raised on instruments: cap_pre falls away from cap_post by the whole amount raised, so the same nominal cap becomes progressively more generous to the holder as the stack grows.", "It also gives a clean way to compare instruments signed on different templates. Restate every cap as a pre-money cap on the same FD, and the stack becomes a single ordered list of prices.", "The identity holds only for capped instruments with no discount and no interest. Anything that makes the conversion price depend on the round price breaks it, and there is no closed form left." ] }, { "id": "flat-and-down-round-price-algebra", "term": "A down round is a price cut, and the price cut is exact", "definition": "A round priced below the last one is described in valuation terms and felt in price terms. The ratio of the two prices is the number that drives anti-dilution, the 409A analysis and every existing holder's mark, and it is not the ratio of the two valuations.", "formula": "Price ratio = (PRE_n/FD_(n-1)) / (PRE_(n-1)/FD_(n-2)). Existing holders retain PRE_n/POST_n regardless of whether the round is up or down", "fields": [ { "label": "Setup", "value": "After Series C the ledger is 24,509,805 shares at a 6.1200 price. A Series D raises 20,000,000 at PRE = 90,000,000", "mono": true }, { "label": "Price", "value": "90,000,000/24,509,805 = 3.6720, against 6.1200 - a price ratio of 0.6000 and a cut of 40.000 percent", "mono": true }, { "label": "Valuation comparison", "value": "PRE 90,000,000 against the prior POST of 150,000,000 is a step-down of 0.60, a different number", "mono": true }, { "label": "Shares and ownership", "value": "20,000,000/3.6720 = 5,446,623 shares; POST = 110,000,000; the new investor holds 18.182 percent", "mono": true }, { "label": "Existing holders retain", "value": "90,000,000/110,000,000 = 81.818 percent, so founders move from 32.640 to 26.705 percent", "mono": true }, { "label": "Anti-dilution input", "value": "The 0.6000 price ratio is what a full ratchet resets Series C to; a weighted average uses it together with the share count issued", "mono": true } ], "notes": [ "The price ratio and the valuation ratio differ whenever the share count changed between the rounds, and after a round of financing it always has. Quote the price.", "A flat round - the same price, not the same valuation - is the boundary case. Priced at 6.1200 on the same ledger it would be a 150,000,000 pre-money, which is a step-up in valuation terms and a flat round in price terms.", "A down round is what triggers anti-dilution, and anti-dilution is the reason a down round takes longer to close than an up round of the same size: the adjustment has to be computed, and usually waived or reset, before anyone can sign.", "The 409A analysis moves too, in both directions. A lower preferred price lowers the ceiling on the common value, which lowers the strike on new grants - one of the few compensating effects available in a down round." ] }, { "id": "entry-to-exit-ownership-and-moic", "term": "Return on invested capital is entry ownership times the retained fraction", "definition": "An investor's multiple at exit is not its entry ownership multiplied by the exit value. It is its exit ownership multiplied by the exit value, and exit ownership is entry ownership multiplied by the product of the later rounds' retention factors.", "formula": "p_exit = p_entry * product of PRE_i/POST_i over all subsequent rounds. Multiple on invested capital = p_exit*E/I, provided E is above the holder's conversion flip point", "fields": [ { "label": "Seed", "value": "Entry 20.000 percent, retained 0.75000*0.80000*0.80000 = 0.48000, exit 9.600 percent", "mono": true }, { "label": "Series A", "value": "Entry 25.000 percent, retained 0.80000*0.80000 = 0.64000, exit 16.000 percent", "mono": true }, { "label": "At E = 250,000,000", "value": "Seed 24,000,000 on 2,000,000 = 12.00x; Series A 40,000,000 on 8,000,000 = 5.00x", "mono": true }, { "label": "Series B and C", "value": "Series B 2.00x; Series C 1.67x", "mono": true }, { "label": "Aggregate", "value": "60,000,000 invested returns 154,000,000 = 2.57x across the stack", "mono": true }, { "label": "Per-share check", "value": "250,000,000/24,509,805 = 10.2000 a share, against entry prices of 0.8500, 2.0400, 5.1000 and 6.1200", "mono": true } ], "notes": [ "The per-share check is the fastest way to compute every holder's multiple at once: divide the exit value by the fully diluted count and divide the result by each round's price. On these numbers that is 12.00x, 5.00x, 2.00x and 1.67x, with no ownership arithmetic at all.", "It also makes the dependence explicit. An investor's multiple is set by the price it paid and the exit price per share; its ownership percentage is a derived quantity that happens to be easier to talk about.", "The formula holds only above the holder's conversion flip point. Below it the holder takes a preference and the multiple is x*I/I, capped at the multiple in the charter, which is why a preference is a floor on return rather than a claim on value.", "Aggregate returns across a stack are dominated by the earliest cheque, which is the smallest. That asymmetry, not the size of the later rounds, is what makes early ownership worth negotiating for." ] } ], "reviewed": "2026-08-27" }, { "slug": "terms", "title": "Term sheet provisions", "lede": "The standard clause set, grouped by what each one actually controls.", "intro": "Term sheet provisions fall into three groups: economics, control, and process. Attention concentrates on the first group, while the second determines who can force or block an outcome, and the third determines how long everything takes.", "tables": [ { "title": "Economic provisions", "columns": [ "Provision", "What it controls" ], "rows": [ [ "Valuation and amount", "Ownership split and price per share" ], [ "Liquidation preference", "Payout order and amount in an exit" ], [ "Participation", "Whether preferred shares in proceeds after its preference" ], [ "Dividends", "Whether a return accrues before common is paid; cumulative dividends compound the preference" ], [ "Anti-dilution", "Repricing on a subsequent down round" ], [ "Pay-to-play", "Penalty for not participating in a later round, usually conversion to common" ] ] }, { "title": "Control provisions", "columns": [ "Provision", "What it controls" ], "rows": [ [ "Board composition", "Who holds board seats and therefore board votes" ], [ "Protective provisions", "The list of actions requiring preferred consent regardless of board or common vote" ], [ "Drag-along", "Ability to compel minority holders into an approved sale" ], [ "Voting agreement", "How shares are voted on board election and specified matters" ], [ "Redemption rights", "Ability to force the company to repurchase preferred after a period" ] ] }, { "title": "Process and information provisions", "columns": [ "Provision", "What it controls" ], "rows": [ [ "Pro rata rights", "Right to maintain ownership percentage in future rounds" ], [ "Right of first refusal", "Company or investor right to match a proposed transfer" ], [ "Co-sale / tag-along", "Right to join a founder's sale on the same terms" ], [ "Information rights", "Frequency and content of financial reporting to holders" ], [ "Registration rights", "Ability to require registration of shares in a public offering" ], [ "No-shop / exclusivity", "Period during which the company cannot solicit competing offers" ] ] }, { "title": "Board composition arithmetic", "intro": "A simple majority of a board of n seats requires floor(n/2) + 1 votes. The last column assumes every director votes with their appointing group.", "columns": [ "Structure", "Seats", "Votes for a simple majority", "Largest bloc", "Who decides a contested vote" ], "rows": [ [ "2 founder, 1 investor", "3", "2", "Founders, 2", "Founders alone" ], [ "1 founder, 1 investor, 1 independent", "3", "2", "None, 1 each", "The independent, with either side" ], [ "2 founder, 2 investor, 1 independent", "5", "3", "Tied at 2", "The independent, with either side" ], [ "2 founder, 3 investor", "5", "3", "Investors, 3", "Investors alone" ], [ "2 founder, 2 investor", "4", "3", "Tied at 2", "Nobody - a 2-2 vote fails" ], [ "2 founder, 3 investor, 2 independent", "7", "4", "Investors, 3", "Investors plus one independent" ] ] }, { "title": "Protective provisions, and the consent structures available for each", "intro": "A protective provision is a list entry plus a consent structure. The list is what requires consent; the structure decides who gives it. The four structures in the last column are, from most founder-friendly to least: all preferred voting together as a single class by simple majority; the same by supermajority; a majority of each series voting separately; and a named series' individual consent. No frequencies are given here - the point is the choice, not its prevalence.", "columns": [ "Action requiring consent", "Why it is on the list", "What it blocks in practice", "Structures the consent can take" ], "rows": [ [ "Liquidation, dissolution or a deemed liquidation event", "The preference is only worth what the exit pays", "A sale at a price the holder considers too low", "Class majority; supermajority; series-by-series" ], [ "Amending the charter or bylaws adversely to the preferred", "Every economic term lives in the charter", "Any change to preference, participation, dividends or conversion", "Class majority is the minimum; adverse-effect language often adds a series vote" ], [ "Authorising or issuing senior or pari passu stock", "A new senior series moves the holder down the waterfall", "The next round, if it is structured", "Class majority; frequently a series consent for the most recent series" ], [ "Increasing or decreasing the authorised preferred or common", "Share counts set the denominator and the pool", "A pool refresh or a large new authorisation", "Class majority" ], [ "Redeeming or repurchasing shares", "Cash leaving to a holder is cash not available to the preference", "Founder secondary and tender offers", "Class majority, usually with carve-outs for repurchases at cost from departing employees" ], [ "Declaring or paying a dividend", "Same reason", "Any distribution to common", "Class majority" ], [ "Incurring indebtedness above a stated amount", "Debt ranks ahead of the entire preferred stack", "Venture debt and any material facility", "Class majority; the threshold is the negotiated term" ], [ "Changing the size of the board", "Board control is an integer problem", "Adding a seat to break a deadlock", "Class majority; often a series consent where that series appoints a director" ], [ "Creating a subsidiary or transferring material assets to one", "Assets moved out of the company can leave the waterfall", "A joint venture or a licensing structure", "Class majority" ], [ "Changing the principal business", "The holder priced a specific business", "A pivot", "Class majority, where it appears at all" ] ] }, { "title": "Who can block a class vote of the preferred", "intro": "The four-series ledger. A consent that runs to all preferred voting together as a single class needs a majority of the class, which is 30.800 percent of the fully diluted company. A holder blocks that consent only by holding more than half of the class. The last column is the additional share of the class each series would need in order to block alone.", "columns": [ "Holder", "Percent of the fully diluted company", "Percent of the preferred class", "Can block a class majority alone", "Short of a blocking position by" ], "rows": [ [ "Seed", "9.600", "15.584", "No", "34.416" ], [ "Series A", "16.000", "25.974", "No", "24.026" ], [ "Series B", "16.000", "25.974", "No", "24.026" ], [ "Series C", "20.000", "32.468", "No", "17.532" ], [ "All preferred", "61.600", "100.000", "-", "-" ], [ "Common and the option pool", "38.400", "-", "No - not part of the preferred class vote", "-" ] ] }, { "title": "Registration rights and what triggers each one", "intro": "Registration rights govern when a holder can require its shares to be included in a public offering. All three are dormant until the company is public or is going public, which is why they are conceded early and read late. Thresholds and windows below are the drafting variables, stated as variables.", "columns": [ "Right", "What the holder can demand", "Usual gating conditions", "What limits it" ], "rows": [ [ "Demand registration", "That the company file a registration statement covering the holder's shares", "Available only after the earlier of a stated date and a period following the initial public offering; requires holders of a stated percentage of registrable securities to join; subject to a minimum aggregate offering size", "A stated maximum number of demands, and the company's right to defer once in any twelve-month period" ], [ "Piggyback registration", "Inclusion in a registration the company is already filing", "Triggered by the company filing for its own account or for another holder, with stated exclusions such as employee benefit plan and business combination filings", "Underwriter cut-back, applied pro rata among selling holders after the company's own shares" ], [ "Form S-3 registration", "A short-form shelf registration once the company is eligible", "Requires company eligibility to use the form and usually a minimum aggregate offering size; limited in number per twelve-month period", "Same deferral and cut-back mechanics as a demand" ], [ "Lock-up", "Nothing - this is the corresponding obligation", "Holders agree not to sell for a stated period following the offering, at the underwriters' request", "Usually applies to all holders above a threshold, with release provisions that are worth reading" ] ] } ], "entries": [ { "id": "protective-provisions", "term": "Protective provisions", "definition": "A defined list of corporate actions that require the consent of the preferred holders as a class, independent of board approval or common stockholder vote. Functionally a veto.", "notes": [ "Commonly covered: sale of the company, issuing senior or pari passu securities, changing the size of the board, amending the charter, incurring debt above a threshold, and declaring dividends.", "The negotiable dimensions are the length of the list, the thresholds within it, and whether consent runs by series or across all preferred voting together.", "Series-by-series consent gives every round a veto and grows harder to manage with each financing. A single combined preferred vote is simpler and is usually the founder-preferred structure." ] }, { "id": "drag-along", "term": "Drag-along", "definition": "A provision compelling holders who did not approve a sale to vote for it and sell on the same terms, preventing a small holder from blocking a transaction.", "notes": [ "The key variable is the trigger: which combination of board, preferred, and common approval is required before the drag can be exercised.", "A drag triggered by preferred alone can force a sale at a price that pays the preference in full and leaves common with nothing.", "Requiring common approval in the trigger is the standard counterweight." ] }, { "id": "pay-to-play", "term": "Pay-to-play", "definition": "A provision penalising existing preferred holders who do not participate pro rata in a subsequent financing, typically by converting their preferred to common and stripping the preference and protective provisions.", "notes": [ "Rare in favourable financing conditions and common in difficult ones, because it is the mechanism by which a down round is made to happen.", "Variants range from full conversion to common through to partial conversion or loss of anti-dilution only.", "From the company's side it is the single most effective clause for compelling insiders to fund a bridge." ] }, { "id": "no-shop", "term": "No-shop and exclusivity", "definition": "A binding provision in an otherwise non-binding term sheet, prohibiting the company from soliciting or negotiating competing offers for a defined period.", "notes": [ "Most of a term sheet is non-binding. The no-shop, confidentiality, and expense provisions typically are binding.", "The period is the negotiable term. A long exclusivity with no deadline for the investor to complete diligence transfers all timing leverage to the investor.", "A reciprocal commitment - exclusivity in exchange for a defined closing timeline - is the standard counterweight." ] }, { "id": "board-composition-math", "term": "Board composition math", "definition": "Board control is an integer problem, and the seat count decides it before any individual is named. The two structures that look like compromises - an even board, and an odd board with a swing independent - behave very differently.", "formula": "Votes required for a simple majority = floor(seats/2) + 1. A bloc controls the board when its seats are greater than or equal to that number", "fields": [ { "label": "Worked", "value": "5 seats require 3 votes. A 2-2-1 split gives no group control and makes the independent decisive on every contested matter", "mono": true }, { "label": "Worked, even board", "value": "4 seats require 3 votes, so a 2-2 split cannot pass anything. Deadlock is the default outcome, not a tie-break", "mono": true }, { "label": "Worked, investor control", "value": "3 of 5 seats is control outright; no founder or independent vote is needed for any board action", "mono": true } ], "notes": [ "An even board is not a compromise, it is a mutual veto. Whichever side benefits from the status quo wins every deadlock, which is usually not the side that needed the resolution passed.", "The right to appoint the independent is worth more than any single seat, and it is why that appointment is normally drafted as a mutual-consent term rather than allocated to a series.", "Board math is only half the control picture. Protective provisions run to the preferred as a class regardless of the board, so a founder-controlled board can still be unable to act. Read both together." ], "source": "NVCA model Voting Agreement (board composition provisions)" }, { "id": "drag-along-thresholds", "term": "Drag-along thresholds and how they drift", "definition": "A drag is exercisable when the stated approvals are obtained. Whether a holder can block therefore depends on whether it holds more than the gap between the requirement and the votes available without it - and that gap moves as the share register changes.", "formula": "A holder blocks a class-vote requirement of fraction m when its holding exceeds (1 - m) of the class. For a simple majority of a class, a holder blocks by holding more than 50 percent of it", "fields": [ { "label": "Setup", "value": "The drag requires a majority of the preferred and a majority of the common, voting as separate classes. Common outstanding 10,000,000 shares", "mono": true }, { "label": "Worked, at signing", "value": "Founders hold 5,500,000 of the common = 55.0 percent. No drag can be exercised without them", "mono": true }, { "label": "Worked, after a secondary", "value": "Founders sell 1,000,000 shares: 4,500,000/10,000,000 = 45.0 percent. The same clause is now exercisable against them", "mono": true } ], "notes": [ "The threshold is measured on a class whose composition changes every time an option is exercised. A drag that is unexercisable at signing can become exercisable through nothing more than employees exercising vested options, with no transaction and no amendment.", "A drag triggered by the preferred alone can force a sale at a price that clears the preference and leaves common with nothing. Requiring a common majority in the trigger is the standard counterweight and is worth more than any cap on the preference.", "Check whether the common-vote requirement is a majority of the common or a majority of the common held by the founders specifically. The second is a much weaker protection once founders have sold or left." ], "source": "NVCA model Voting Agreement (drag-along provisions)" }, { "id": "redemption-rights-arithmetic", "term": "Redemption rights and the number they produce", "definition": "A redemption right lets the holder require the company to repurchase its preferred after a stated period, at the preference amount plus accrued and unpaid dividends. It is usually unenforceable as a cash claim and valuable as a lever.", "formula": "Redemption price = x*I + accrued dividends = x*I*(1 + d*t) for a simple cumulative dividend, payable in stated instalments out of funds legally available for the purpose", "fields": [ { "label": "Worked", "value": "I = 10,000,000 at 1x with d = 0.08 cumulative simple, redeemable after t = 5 years: 10,000,000*(1 + 0.40) = 14,000,000", "mono": true }, { "label": "In instalments", "value": "Payable over three annual instalments: 14,000,000/3 = 4,666,667 each", "mono": true }, { "label": "The binding constraint", "value": "Redemption is limited to funds legally available; a company without surplus cannot lawfully pay", "mono": true } ], "notes": [ "The number matters more than the payment. A redeemable position carries a stated, dated claim that a board has to disclose and act on, which is what turns a redemption right into a sale process.", "This is also why a cumulative dividend matters even when no dividend is ever declared: it sets the redemption number, and it grows the preference amount in a liquidation at the same time.", "Check whether redemption is at the holder's election or automatic, and whether failure to redeem triggers a change in board composition. The consequence clause is the operative term, not the price." ] }, { "id": "cumulative-dividends-arithmetic", "term": "Cumulative dividends and where they land", "definition": "A cumulative dividend accrues whether or not it is declared and is added to the preference amount. Whether it compounds is a single drafting choice with a large arithmetic consequence over a long private hold.", "formula": "Preference at time t = x*I*(1 + d*t) for simple accrual, or x*I*(1 + d)^t if the dividend compounds. Non-cumulative dividends add nothing unless declared", "fields": [ { "label": "Worked, simple", "value": "10,000,000 at 1x with d = 0.08 over t = 5: 14,000,000", "mono": true }, { "label": "Worked, compounding", "value": "10,000,000 * 1.08^5 = 14,693,281", "mono": true }, { "label": "Effect on the indifference point", "value": "At p = 0.20, E* moves from 10,000,000/0.20 = 50,000,000 to 14,000,000/0.20 = 70,000,000", "mono": true }, { "label": "Over seven years", "value": "15,600,000 simple against 17,138,243 compounding", "mono": true } ], "notes": [ "The dividend is rarely paid and almost always collected, through the liquidation preference or the redemption price. Treat it as a growing preference rather than as a yield.", "A 20,000,000 widening of the dead zone is a larger transfer from common than most of the terms that get negotiated harder, and it arrives through a clause that reads as a technicality.", "A non-cumulative dividend on preferred that is never declared is economically nil. It can be conceded without cost, and conceding it visibly is sometimes worth more than the term." ] }, { "id": "founder-vesting-acceleration", "term": "Founder vesting and acceleration arithmetic", "definition": "Founder shares are typically subject to repurchase or forfeiture on a vesting schedule with a cliff. Acceleration changes what happens to the unvested balance on a change of control, and single and double trigger are materially different instruments.", "formula": "Vested shares at month t on an n-month schedule with a c-month cliff = 0 for t < c, else s*t/n. Single trigger: the unvested balance accelerates on a change of control. Double trigger: acceleration requires both a change of control and a qualifying termination within a stated window", "fields": [ { "label": "Setup", "value": "s = 2,000,000 founder shares, n = 48 months, c = 12 months", "mono": true }, { "label": "Worked, at t = 11", "value": "0 shares vested - the cliff has not been reached", "mono": true }, { "label": "Worked, at t = 12", "value": "2,000,000 * 12/48 = 500,000", "mono": true }, { "label": "Worked, at t = 30", "value": "2,000,000 * 30/48 = 1,250,000, leaving 750,000 unvested", "mono": true }, { "label": "Worked, acquisition at t = 30", "value": "100 percent single trigger: 2,000,000. 50 percent acceleration: 1,250,000 + 375,000 = 1,625,000. Double trigger with no termination: 1,250,000", "mono": true } ], "notes": [ "An acquirer prices unvested founder equity as retention budget. Full single-trigger acceleration removes the retention, and acquirers routinely take the value back out of the purchase price - so the founder pays for it twice and receives it once.", "Double trigger is the structure that survives diligence, because it pays only in the case founders actually fear: the deal closes and they are removed.", "Credit for time already served before the vesting schedule is imposed is a separate negotiation from the schedule itself, and it is usually cheaper to win. Ask for the vesting start date, not just the length." ] }, { "id": "rofr-co-sale-mechanics", "term": "Right of first refusal and co-sale, in sequence", "definition": "A proposed transfer runs through three gates in order: the company's refusal right, the investors' secondary refusal right, and then co-sale. Each stage reduces what the seller can actually sell, and the arithmetic of the last one is the least anticipated.", "formula": "After the refusal rights leave n shares saleable, co-sale reduces the seller to n * s_seller/(s_seller + s_participating), with the balance sold by the participating holders", "fields": [ { "label": "Setup", "value": "A founder proposes to sell 1,000,000 shares and holds 8,000,000. Co-sale participants hold 4,000,000", "mono": true }, { "label": "Worked, gates one and two", "value": "The company declines; investors take 400,000 under the secondary right, leaving 600,000", "mono": true }, { "label": "Worked, co-sale", "value": "The founder sells 600,000 * 8,000,000/12,000,000 = 400,000; the participants sell 200,000", "mono": true }, { "label": "Result", "value": "A proposed 1,000,000-share sale becomes a 400,000-share sale", "mono": true } ], "notes": [ "Co-sale participation is measured on shares held, not on the number of participants, so a single large holder exercising has the same effect as all of them. Compute the worst case, which is full participation.", "In practice founder secondary is done by express waiver rather than by running the gates, because the buyer wants a fixed size and the process cannot deliver one. Negotiating the waiver is the transaction.", "The notice periods at each gate are additive and often total months. A sale that has to close by a date should start from the notice calendar, not from the price." ], "source": "NVCA model Right of First Refusal and Co-Sale Agreement" }, { "id": "information-rights-thresholds", "term": "Information and pro-rata thresholds - shares versus percentages", "definition": "Rights reserved for holders above a stated threshold behave completely differently depending on whether the threshold is an absolute share count or a percentage. Dilution cannot touch the first and steadily erodes the second.", "formula": "A percentage threshold q is lost once FD grows past s/q. An absolute share threshold is unaffected by dilution and is lost only on transfer", "fields": [ { "label": "Setup", "value": "A holder with s = 600,000 shares and a 5 percent threshold", "mono": true }, { "label": "Worked", "value": "The right survives while FD < 600,000/0.05 = 12,000,000 shares", "mono": true }, { "label": "At FD = 11,411,765", "value": "5.258 percent - the right holds", "mono": true }, { "label": "After one more round to FD = 14,000,000", "value": "600,000/14,000,000 = 4.286 percent - the right lapses with no amendment and no notice", "mono": true } ], "notes": [ "This is the quiet mechanism by which early investors lose information rights and pro-rata rights: nobody removes them, the denominator simply grows. The holder usually discovers it when it asks to exercise pro-rata.", "Well-drafted documents define the major-holder threshold as an absolute share number and restate it after any split, precisely to remove this drift. If a term sheet uses a percentage, that is a substantive choice worth raising.", "The same arithmetic applies to protective provisions that run to a series while a minimum number of its shares remain outstanding. Conversion or redemption of part of a series can extinguish the veto for the rest of it." ], "source": "NVCA model Investors' Rights Agreement (information and pro-rata provisions)" }, { "id": "protective-provisions-enumerated", "term": "The protective provision list, and the four structures a consent can take", "definition": "A protective provision is a veto over a specified corporate action. The list determines what is blocked; the consent structure determines who blocks it. Negotiating the list without negotiating the structure settles the smaller half of the question.", "formula": "Consent threshold as a fraction of the fully diluted company = m * (preferred as a fraction of the company), where m is the fraction of the class required. A holder blocks alone when its holding exceeds (1 - m) of the class", "fields": [ { "label": "Class majority, all preferred together", "value": "m = 0.50 on a class holding 61.600 percent of the company: the consent needs 30.800 percent of the company", "mono": true }, { "label": "Class supermajority", "value": "m = 0.6667: 41.067 percent of the company", "mono": true }, { "label": "Series by series", "value": "Each series consents separately, so a four-series stack has four independent vetoes", "mono": true }, { "label": "Named series consent", "value": "One series holds an unconditional veto over the listed action regardless of the rest of the class", "mono": true }, { "label": "What is usually on the list", "value": "Sale of the company; charter amendments adverse to the preferred; issuing senior or pari passu stock; changing authorised share counts; redemptions and repurchases; dividends; indebtedness above a threshold; changing the size of the board", "mono": false } ], "notes": [ "Series-by-series consent is the structure that ages worst. Each new round adds a veto, and by the fourth round a routine charter amendment needs four separate consents from holders with different incentives and different remaining stakes.", "A single combined class vote is simpler and is usually the founder-preferred structure, but it concentrates the veto in whichever holders can assemble a majority of the class - which after a large late round may be one investor plus one other.", "The list is negotiated as a governance point and prices like an economic one. The consent over issuing senior stock is a veto over the next round; the consent over indebtedness is a veto over venture debt; the consent over board size is a veto over resolving a deadlock.", "Read the adverse-effect language separately from the list. A provision requiring a series vote on any amendment that adversely affects that series is a general-purpose veto that does not appear as a list entry." ], "source": "NVCA model Amended and Restated Certificate of Incorporation (protective provisions) and model Certificate of Incorporation voting provisions" }, { "id": "protective-provision-vote-arithmetic", "term": "What a class majority actually requires on a real cap table", "definition": "Consent thresholds are stated as fractions of a class, and the class is a moving fraction of the company. Converting the threshold into a share of the fully diluted company shows who can grant it, who can block it, and which coalitions are decisive.", "formula": "Fraction of the company needed = m * p_class. Series j holds p_j/p_class of the class and blocks alone when that exceeds 1 - m", "fields": [ { "label": "Setup", "value": "Four-series ledger: Seed 9.600, Series A 16.000, Series B 16.000, Series C 20.000 percent of the company. The preferred class holds 61.600 percent", "mono": true }, { "label": "Class majority", "value": "0.50 * 61.600 = 30.800 percent of the company, or 7,549,020 shares", "mono": true }, { "label": "Each series' share of the class", "value": "Seed 15.584, Series A and B 25.974 each, Series C 32.468 percent", "mono": true }, { "label": "Who can block alone", "value": "Nobody - the largest holding is 32.468 percent of the class and blocking requires more than 50 percent", "mono": true }, { "label": "Decisive coalitions", "value": "Series B plus Series C = 58.442 percent of the class, a majority. Series A plus Series C = 58.442 percent. Seed plus Series A plus Series B = 67.532 percent", "mono": true }, { "label": "Under a two-thirds supermajority", "value": "Blocking requires more than 33.333 percent of the class, which Series C alone has at 32.468 percent", "mono": true } ], "notes": [ "The supermajority is the term that hands the newest investor a unilateral veto here, and it does so without naming it. Moving m from 0.50 to 0.6667 converts a coalition problem into a single consent, and it is presented as a stricter standard rather than as a transfer of control.", "Because the class is a moving fraction of the company, every subsequent round changes these figures. A holder that could block at the time it invested may not be able to two rounds later, and nothing in the documents will have changed.", "Compute the coalition table before conceding a threshold. The relevant question is never the percentage; it is which two holders can act together and whether the founder can get to either of them.", "Options and the unallocated pool are common stock and do not vote in a preferred class vote, so the pool's 5.760 percent is irrelevant to every figure above. It does matter to any consent requiring a common vote." ] }, { "id": "series-versus-class-consent-drift", "term": "Series-by-series consent multiplies with every round", "definition": "A consent that runs to all preferred as one class needs one negotiation regardless of how many series exist. A consent that runs series by series needs one negotiation per series, and the number of series only ever grows. The cost of the structure is invisible when it is agreed and compounding thereafter.", "formula": "Consents required = 1 for a class vote, or n for series-by-series with n series. A series-specific veto survives while that series has shares outstanding above any stated minimum", "fields": [ { "label": "At the seed round", "value": "One series, so the two structures are identical and the choice appears costless", "mono": true }, { "label": "After Series C", "value": "A class vote needs 30.800 percent of the company in one negotiation; series-by-series needs four separate majorities", "mono": true }, { "label": "The smallest veto", "value": "Seed holds 9.600 percent of the company and, under series-by-series consent, a full veto over every listed action", "mono": true }, { "label": "Minimum-shares conditions", "value": "A veto drafted to survive while a stated minimum of the series remains outstanding can be extinguished by conversion or repurchase of part of the series", "mono": true }, { "label": "Where it bites", "value": "A charter amendment for a routine pool increase, a bridge, or a sale can require consent from a holder whose remaining economic interest is a rounding error", "mono": true } ], "notes": [ "The asymmetry is what makes this worth raising at the seed round rather than the fourth one: the founder is negotiating the structure once, at the moment it costs nothing, on behalf of every future round.", "A holder whose stake has been diluted to a few percent has the least to lose from blocking and the least to gain from agreeing, which is exactly the wrong incentive to attach a veto to.", "The usual compromise is a class vote for the general list plus a series vote confined to amendments that adversely and specifically affect that series. That is a much narrower veto and it is defensible on its own terms.", "Where a series-by-series structure already exists, check the minimum-shares condition on each series. Extinguishing a stale veto by a partial conversion or repurchase is sometimes cheaper than obtaining the consent." ] }, { "id": "board-observer-mechanics", "term": "Board observers - what the right actually delivers", "definition": "An observer attends board meetings and receives board materials without voting and without a director's fiduciary duties. It is granted as a lesser alternative to a seat and is not a lesser version of the same thing: the information rights are nearly identical and the influence is entirely informal.", "formula": "No arithmetic. Observers are excluded from the board count, so they do not change the majority threshold of floor(seats/2) + 1", "fields": [ { "label": "What the observer gets", "value": "Notice of meetings, board packages, and the right to attend and speak", "mono": false }, { "label": "What the observer does not get", "value": "A vote, a fiduciary duty to the company, or the ability to be counted towards a quorum", "mono": false }, { "label": "Standard limitations", "value": "Exclusion from executive session; exclusion where attendance would waive attorney-client privilege or create a conflict; a confidentiality undertaking", "mono": false }, { "label": "Board arithmetic", "value": "A board of 5 seats needs 3 votes with any number of observers present. Observers change the room and not the count", "mono": true }, { "label": "Who asks for it", "value": "Holders too small for a seat, and holders whose own conflicts make a directorship awkward", "mono": false } ], "notes": [ "The exclusion for privilege and conflicts is the clause that matters, because it is the mechanism by which an observer is kept out of exactly the discussions it most wants to attend - a competing portfolio company, a sale process involving an affiliate, litigation.", "Observer rights accumulate. Four rounds of granting one observer produces a board meeting with four non-voting attendees, which changes what gets discussed at the meeting rather than what gets decided.", "Because the observer owes the company nothing, the confidentiality undertaking is the only protection. Ask for it in the document rather than relying on the investor's policy.", "An observer right is much easier to remove than a board seat, since it usually sits in a side letter or the investors' rights agreement rather than in the charter or the voting agreement. That asymmetry is a reason to prefer granting it." ], "source": "NVCA model Investors' Rights Agreement (board observer provisions)" }, { "id": "board-deadlock-resolution", "term": "Deadlock, and the four ways out of it", "definition": "An even board splits evenly, and a split vote fails. Because failure is a decision in favour of the status quo, a deadlock is not neutral: it resolves in favour of whichever side does not need the resolution passed.", "formula": "A simple majority of n seats is floor(n/2) + 1. On an even board an evenly split bloc structure cannot reach it, so no resolution passes", "fields": [ { "label": "The arithmetic", "value": "4 seats require 3 votes; a 2-2 split reaches 2. 6 seats require 4; a 3-3 split reaches 3", "mono": true }, { "label": "Route one, an independent seat", "value": "Move to 5 seats with a mutually appointed independent. Majority becomes 3 and the independent is decisive on every contested matter", "mono": true }, { "label": "Route two, a casting vote", "value": "Give the chair a second vote on a tie. This is control, relabelled, and it is usually resisted for that reason", "mono": false }, { "label": "Route three, escalation", "value": "Refer the deadlocked matter to a stockholder vote or to a defined dispute process, which relocates the deadlock to the share register", "mono": false }, { "label": "Route four, a deadlock trigger", "value": "Define a consequence - a buy-sell, a sale process, a change in board composition - that makes deadlock costly to both sides", "mono": false }, { "label": "The status quo bias", "value": "Whoever benefits from nothing happening wins. Identify which side that is on each foreseeable matter before agreeing an even board", "mono": false } ], "notes": [ "An even board is not a compromise, it is a mutual veto. It is agreed because it looks balanced on the page and because neither side wants to argue about the independent, and the cost arrives at the first contested decision.", "The right to appoint the independent is worth more than any single seat, which is why it is normally drafted as a mutual-consent term rather than allocated to a series. A seat allocated to a series is not independent whatever it is called.", "Board arithmetic is only half of control. Protective provisions run to the preferred as a class regardless of board composition, so a founder-majority board can still be unable to act. Read the two together and build a single matrix of who can do what.", "A deadlock trigger with a real consequence is the only route that changes behaviour before the deadlock happens. The other three change what happens afterwards." ], "source": "NVCA model Voting Agreement (board composition provisions)" }, { "id": "drag-along-protective-provision-interaction", "term": "A drag-along does not override a protective provision", "definition": "A drag compels holders to vote for and sell into an approved transaction. A protective provision requires the preferred's consent before the transaction can be approved at all. The two operate at different stages, so satisfying the drag does not satisfy the consent, and a holder can be dragged on a deal it was able to block.", "formula": "Sequence: obtain the protective provision consent (a class or series vote), then obtain the drag trigger approvals, then exercise the drag against the remaining holders. Each step has its own threshold", "fields": [ { "label": "Step one, the consent", "value": "A deemed liquidation event needs the preferred's consent: 30.800 percent of the company under a class majority", "mono": true }, { "label": "Step two, the drag trigger", "value": "Typically the board, plus a majority of the preferred, plus in the better-drafted version a majority of the common", "mono": true }, { "label": "Step three, the drag itself", "value": "Binds every holder party to the agreement, whether or not it voted for the deal", "mono": true }, { "label": "Where the gap opens", "value": "A holder that is part of a blocking coalition at step one, but is outvoted at step two, is dragged on a deal it could have prevented had it acted earlier", "mono": true }, { "label": "The common-vote counterweight", "value": "Requiring a majority of the common at step two is worth more to employees and founders than any cap on the preference, because it puts the residual claimants inside the trigger", "mono": true } ], "notes": [ "The practical lesson is about timing rather than thresholds. A holder's leverage is at the consent stage, and it is spent by the time the drag is being exercised. Raising an objection after the trigger has been met is raising it too late.", "A drag triggered by the preferred alone can force a sale at a price that pays the preference in full and leaves common with nothing. That is not a drafting accident - it is the term working as written - and the fix is the common vote in the trigger.", "Check whether the common-vote requirement is a majority of the common or a majority of the common held by the founders. The second is much weaker once founders have sold or left, and it degrades silently as options are exercised.", "Drags usually carry conditions protecting the dragged holder: consideration in the same form, representations limited to title, liability capped at proceeds and several rather than joint. Those conditions are where a dragged minority actually gets protected." ], "source": "NVCA model Voting Agreement (drag-along provisions) and model Amended and Restated Certificate of Incorporation (protective provisions)" }, { "id": "registration-rights-triggers", "term": "Registration rights are dormant until they are not", "definition": "Demand, piggyback and short-form registration rights all require the company to be public or to be going public. They cost nothing at the time they are granted and they constrain the underwriting of the offering that eventually happens, which is when they are read for the first time.", "formula": "No arithmetic in the grant. The negotiated variables are the number of demands, the earliest date they become exercisable, the minimum aggregate offering size, the deferral right, and the cut-back priority", "fields": [ { "label": "Demand", "value": "Requires the company to register the holder's shares. Gated by a date or a period after the initial public offering, a threshold of holders joining, and a minimum offering size", "mono": false }, { "label": "Piggyback", "value": "Inclusion in a registration the company is already making. Not gated by a date, which is what makes it the right that gets used", "mono": false }, { "label": "Short-form", "value": "A demand using the short registration form once the company is eligible, limited in number per twelve-month period", "mono": false }, { "label": "Company deferral", "value": "A right to postpone a demand once in any twelve-month period, usually for a stated number of days", "mono": false }, { "label": "Expiry", "value": "Rights normally terminate on a stated date after the offering, or when the holder's shares become freely saleable without registration", "mono": false } ], "notes": [ "The clause that does real work is the cut-back priority, because an offering is sized by the underwriters and not by the holders' requests. Priority decides who is cut and by how much when the requests exceed the size.", "Demand rights are almost never exercised against a healthy company, because a demand registration is an announcement that insiders want out. The right's value is as leverage over the timing and composition of a company-initiated offering.", "The lock-up is the mirror obligation and it applies to everyone above a threshold, including founders and often employees. Read the release language: a lock-up that releases early for one holder and not others is a real economic difference.", "These are the terms most often granted without negotiation on the reasoning that they only matter in a good outcome. That is true, and the good outcome is where the money is." ], "source": "NVCA model Investors' Rights Agreement (registration rights)" }, { "id": "registration-cutback-arithmetic", "term": "The underwriter cut-back, worked", "definition": "In a piggyback or a demand the underwriters set the size of the offering. If the shares requested exceed what the market will take, the requests are cut back - the company's own shares first in priority, then the selling holders pro rata to the shares each requested. The arithmetic is a single ratio applied to every request.", "formula": "Cut-back factor = shares the underwriters will include / total shares requested. Each holder sells its request multiplied by that factor", "fields": [ { "label": "Requests", "value": "Founders 2,000,000, Seed 2,352,941, Series A 3,921,569, Series B 3,921,569, total 12,196,079", "mono": true }, { "label": "Underwriters will include", "value": "5,000,000 shares from selling holders", "mono": true }, { "label": "Factor", "value": "5,000,000/12,196,079 = 40.997 percent", "mono": true }, { "label": "Result", "value": "Founders 819,936, Seed 964,630, Series A 1,607,717, Series B 1,607,717", "mono": true }, { "label": "Check", "value": "5,000,000 shares, equal to the allowance", "mono": true }, { "label": "If the company's own shares come first", "value": "An offering sized at 8,000,000 shares of which the company sells 3,000,000 leaves exactly this 5,000,000 for holders", "mono": true } ], "notes": [ "A holder that requests more than it wants to sell receives a larger share of the cut-back, because the factor is applied to the request. That is a well-known and entirely legitimate response, and it is why sophisticated holders over-request.", "Priority is the term to read, not the factor. Company shares ahead of holders is standard; holders ahead of the company is not; and a named series ahead of the other holders is the version that is worth objecting to.", "Founders are usually inside the pro rata group rather than behind it, which is worth confirming. A founder placed behind the investors in a cut-back can be cut to nothing in a small offering.", "The same mechanic appears in the co-sale agreement, where participation is measured on shares held rather than shares requested. Two similar-looking pro rata rules with different bases produce different answers, so check which one applies." ], "source": "NVCA model Investors' Rights Agreement (underwriter cut-back provisions)" }, { "id": "conversion-on-ipo", "term": "Mandatory conversion, and the threshold that decides whether the preference survives", "definition": "A qualifying public offering converts the whole preferred stack to common automatically, extinguishing every preference and every protective provision at once. Whether an offering qualifies is a definition with two numbers in it, and an offering that misses either one leaves the preference in place in a public company.", "formula": "Mandatory conversion occurs on an offering at a price per share of at least k times the original issue price with gross proceeds of at least a stated amount. Below either threshold conversion requires the ordinary class vote instead", "fields": [ { "label": "Setup", "value": "Series C issued at 6.1200. A threshold of 3 times the original issue price sets the qualifying price at 18.3600", "mono": true }, { "label": "An offering at 12.0000", "value": "Below 18.3600, so no mandatory conversion. Series C keeps a 30,000,000 preference and its consent rights", "mono": true }, { "label": "An offering at 20.0000", "value": "Above the threshold and above any stated minimum size, so the whole stack converts and every holder receives p of the company", "mono": true }, { "label": "Value of the preference at 12.0000", "value": "Series C's 20.000 percent of a 294,117,660 market value is 58,823,532, against a 30,000,000 preference - so at that price the preference is worth less than the stock and conversion is voluntary anyway", "mono": true }, { "label": "Where the threshold bites", "value": "At an offering price between the preference indifference point and the qualifying price, the holder prefers not to convert and cannot be forced to", "mono": true } ], "notes": [ "The threshold exists to stop a company going public at a price that leaves the preferred worse off than a private sale. Set high, it hands the preferred a veto over the offering; set low, it removes the protection the preference was bought for.", "A separate series consent for conversion, layered on top of the price threshold, is the version that gives the most recent investor a veto over the public offering itself. That is worth identifying as a control term rather than a mechanical one.", "Some charters allow the class to elect conversion by vote at any price. Where that exists, an offering below the threshold can still be done, but only with the consent the threshold was designed to make unnecessary.", "Preferred stock outstanding after an offering is unusual and expensive: the preference has to be disclosed, it complicates the share count used for earnings per share, and it means the underwriters are selling a junior security. Most offerings are structured to clear the threshold for exactly that reason." ], "source": "NVCA model Amended and Restated Certificate of Incorporation (mandatory conversion provisions)" }, { "id": "pro-rata-and-super-pro-rata", "term": "Pro rata, passing, and super pro rata, on the same round", "definition": "A pro rata right lets a holder invest enough to keep its percentage flat. Investing its own percentage of the new money is exactly sufficient, whatever the price. Investing more than that increases the percentage, and the arithmetic of how much more is straightforward once the price is known.", "formula": "To hold q constant through a round of R, invest q*R. To reach a target q', invest (q'*FD_post - s_held) * price, where FD_post is unchanged by who buys the round", "fields": [ { "label": "Setup", "value": "Series A holds 3,921,569 shares = 20.000 percent before a Series C that raises 30,000,000 at a price of 6.1200. Post-round FD is 24,509,805 whoever buys", "mono": true }, { "label": "Passes entirely", "value": "Ownership falls to 16.000 percent", "mono": true }, { "label": "Pro rata", "value": "0.20000 * 30,000,000 = 6,000,000, buying 980,392 shares for a total of 4,901,961 = 20.000 percent", "mono": true }, { "label": "Super pro rata at 1.5 times", "value": "9,000,000 buys 1,470,588 shares for a total of 5,392,157 = 22.000 percent", "mono": true }, { "label": "Cost of each point", "value": "Moving from 20.000 to 22.000 percent costs 3,000,000, so a point of ownership costs 1,500,000 at this price", "mono": true }, { "label": "What the company gives up", "value": "Allocation, not cash. Every dollar of pro rata is a dollar the new lead cannot take, which is why leads negotiate pro rata waivers in a competitive round", "mono": true } ], "notes": [ "The right is an option struck at the round price. Its value is not the ownership it preserves but the ability to decline, and it is worth most in exactly the rounds where the holder has the best information about whether to exercise.", "Super pro rata is the same right with a multiplier, and it is a much more aggressive ask because it lets an existing holder take allocation away from a new lead. Where it exists it is usually confined to a stated cap or a single round.", "A holder that passes is not diluted by the pro rata it declined; it is diluted by the round, at the same rate as everyone else. The pro rata right does not protect against dilution, it offers the chance to buy it back at the round price.", "The post-money SAFE moved pro rata out of the instrument and into a separate side letter. If no side letter was signed, the right does not exist regardless of what was discussed." ] }, { "id": "double-trigger-definition", "term": "Double trigger acceleration - the definitions are the whole term", "definition": "Double trigger acceleration pays only if two things happen: a change of control, and a qualifying termination within a stated window after it. The percentage accelerated is the headline. The window length, the definition of a qualifying termination and the definition of good reason are what determine whether it ever pays.", "formula": "Vested shares at month t on an n-month schedule = s*t/n after the cliff. Double trigger pays the unvested balance s*(1 - t/n) if a qualifying termination occurs within the window following the change of control", "fields": [ { "label": "Setup", "value": "2,000,000 founder shares, 48-month schedule, 12-month cliff. Change of control at month 30", "mono": true }, { "label": "At the closing", "value": "Vested 1,250,000, unvested 750,000", "mono": true }, { "label": "Terminated at month 38, 12-month window", "value": "Vested 1,583,333, so acceleration delivers the remaining 416,667 shares - worth 4,250,000 at a 10.2000 exit price", "mono": true }, { "label": "Same termination, 6-month window", "value": "Nothing accelerates. The window expired at month 36", "mono": true }, { "label": "Constructive termination", "value": "If good reason is not defined, an acquirer can move the role, the title or the location and let the founder resign, which is not a qualifying termination", "mono": true }, { "label": "Single trigger for comparison", "value": "The full 750,000 accelerates at the closing regardless of what happens afterwards", "mono": true } ], "notes": [ "The window is the term to negotiate and it is the one most often left at whatever the first draft says. A window shorter than the acquirer's normal integration period converts double trigger acceleration into no acceleration.", "Good reason has to be defined and the definition has to include the changes an acquirer actually makes: reporting line, scope of duties, base compensation, and required relocation beyond a stated distance. Without it, the second trigger is in the acquirer's gift.", "An acquirer prices unvested founder equity as retention budget. Full single-trigger acceleration removes the retention and acquirers routinely take the value back out of the purchase price, so the founder pays for it once and receives it once.", "Double trigger is the structure that survives diligence because it pays only in the case founders actually fear: the deal closes and they are removed. That is also why it is the structure an acquirer will accept." ] }, { "id": "exclusivity-and-closing-calendar", "term": "The no-shop is a deadline for one side only unless the calendar is in it", "definition": "Most of a term sheet is non-binding; the no-shop, confidentiality and expense provisions usually are not. An exclusivity period with no corresponding commitment on the investor's side transfers all timing leverage, because the company cannot run a process and the investor has no date to meet.", "formula": "No arithmetic in the clause. The test is whether the exclusivity period is longer than the sum of the steps that have to happen inside it. All day counts below are illustrative inputs, not standard periods", "fields": [ { "label": "The exclusivity period", "value": "A stated number of days from signing the term sheet, say 45", "mono": true }, { "label": "What has to happen inside it", "value": "Confirmatory diligence; documentation; charter amendment and stockholder consent; any protective provision consents; a pool increase approval; signature and funding", "mono": false }, { "label": "A worked calendar", "value": "Diligence 20 days, documents 20 days in parallel from day 10, consent solicitation 10 days, closing mechanics 5 days: 45 days with no slack at all", "mono": true }, { "label": "Where it goes wrong", "value": "A single consent that has to be chased, or a diligence request that arrives on day 30, pushes the closing past the expiry - at which point the company has spent its exclusivity and has no signed deal", "mono": false }, { "label": "The counterweight", "value": "Exclusivity in exchange for a defined closing date, with the period lapsing if the investor has not signed by it. Reciprocal, short to draft, and the standard answer", "mono": false } ], "notes": [ "A no-shop is the only clause in a non-binding document that a company can breach, so it is worth reading with the care given to a definitive agreement rather than the care given to a term sheet.", "The productive negotiation is not the length. It is a lapse provision, an obligation to complete diligence by a stated date, and an express right to respond to unsolicited inbound interest.", "Where a round has several investors, exclusivity signed with a lead that has not syndicated is exclusivity against a deal that may not exist. Ask what is still conditional on other parties before signing it.", "Expenses and confidentiality survive whether or not the deal closes. An expense reimbursement with no cap, in a deal that fails, is a real liability created by a document described as non-binding." ] } ], "reviewed": "2026-08-27" }, { "slug": "employee-equity", "title": "Employee equity and 409A", "lede": "Strike prices, tax events, and what a grant is actually worth behind a preference stack.", "intro": "Employee equity is priced off the fair market value of common stock and paid out of the residual left after the preferred stack. Those are two different numbers from two different parts of the documents, and almost every misunderstanding about a grant comes from using one where the other belongs. The arithmetic below runs on the same share ledger as the rest of this corpus: 8,000,000 founder shares, a 1,411,765-share pool, and four priced rounds at 0.8500, 2.0400, 5.1000 and 6.1200 a share, reaching a fully diluted count of 24,509,805. The fraction of the preferred price used as the common fair market value at each round is an input chosen for legibility, not a benchmark: a 409A valuation is an independent appraisal and its output cannot be derived from the round price.", "reviewed": "2026-08-27", "tables": [ { "title": "Four instruments, and when each one is taxed", "intro": "Restricted stock, restricted stock units, incentive stock options and non-qualified stock options are four different tax objects granted for the same purpose. The columns below are the events at which an amount is included in income; the amounts themselves are computed in the entries that follow. This is a structural map and not tax advice.", "columns": [ "Instrument", "At grant", "At vesting", "At exercise or settlement", "On sale", "Election available" ], "rows": [ [ "Restricted stock, purchased", "Nothing if the purchase price equals fair market value", "Ordinary income on the spread between fair market value and price at each vesting date, unless an 83(b) election was filed", "Not applicable - the shares are already outstanding", "Capital gain or loss from the price paid, with the holding period running from purchase if 83(b) was filed and from each vesting date if not", "Section 83(b), within 30 days of the transfer" ], [ "Restricted stock unit", "Nothing", "Ordinary income on the full fair market value at settlement", "Settlement is the taxable event and there is no exercise price", "Capital gain or loss from the value included at settlement", "None generally available; the timing is set by the plan" ], [ "Incentive stock option", "Nothing", "Nothing", "No ordinary income; the spread is an item of adjustment for alternative minimum tax purposes", "Capital gain from the strike if the statutory holding periods are met; otherwise the disposition is disqualifying and the spread becomes ordinary income", "Early exercise where the plan permits, combined with an 83(b) election on unvested shares" ], [ "Non-qualified stock option", "Nothing", "Nothing", "Ordinary income on the spread between fair market value and strike, with employment tax withholding", "Capital gain or loss from the value at exercise", "Early exercise where the plan permits, combined with an 83(b) election" ] ] }, { "title": "The 409A ladder against the share ledger", "intro": "The preferred price at each round comes from the share ledger. The common fraction is a stated input chosen so the arithmetic is legible - it is the assumption in this table and nothing in this corpus derives it. The strike is the common fair market value, and the last column is the value of that vintage at an exit of 250,000,000, which on a fully diluted count of 24,509,805 is 10.2000 a share.", "columns": [ "Round", "Preferred price", "Common fraction assumed", "Common fair market value and strike", "Discount to the preferred price", "Exit value per share", "Spread per share at exit" ], "rows": [ [ "Seed", "0.8500", "30", "0.2550", "70.0", "10.2000", "9.9450" ], [ "Series A", "2.0400", "35", "0.7140", "65.0", "10.2000", "9.4860" ], [ "Series B", "5.1000", "40", "2.0400", "60.0", "10.2000", "8.1600" ], [ "Series C", "6.1200", "45", "2.7540", "55.0", "10.2000", "7.4460" ] ] }, { "title": "A 200,000-option grant, priced four ways", "intro": "The same 200,000 options granted at each vintage, exercised at an exit price of 10.2000. Net exercise shares are n*(FMV - strike)/FMV, the share count whose value equals the spread. Sell-to-cover assumes an ordinary rate of 40 percent applied to the spread, an input chosen for legibility and not a rate recommendation; it sells enough shares at 10.2000 to fund the strike and the tax.", "columns": [ "Vintage", "Strike", "Cash to exercise 200,000", "Spread per share", "Total spread", "Shares delivered on a net exercise", "Cash needed to exercise and pay tax", "Shares sold to cover", "Shares retained" ], "rows": [ [ "Seed", "0.2550", "51,000", "9.9450", "1,989,000", "195,000", "846,600", "83,000", "117,000" ], [ "Series A", "0.7140", "142,800", "9.4860", "1,897,200", "186,000", "901,680", "88,400", "111,600" ], [ "Series B", "2.0400", "408,000", "8.1600", "1,632,000", "160,000", "1,060,800", "104,000", "96,000" ], [ "Series C", "2.7540", "550,800", "7.4460", "1,489,200", "146,000", "1,146,480", "112,400", "87,600" ] ] } ], "entries": [ { "id": "instrument-choice-iso-nso-rsu-restricted", "term": "Choosing between restricted stock, options and units", "definition": "The four instruments differ in when income is recognised and in what has to be paid to acquire the shares. At an early-stage company where the common value is low, restricted stock purchased outright is the cheapest to hold and the most exposed if the company fails. At a later stage the same instrument is unaffordable and units become the only workable form.", "formula": "Cash required at acquisition: restricted stock = shares * price; options = shares * strike; units = 0. Income recognised: restricted stock with an 83(b) election = spread at purchase; without = spread at each vesting date; options = spread at exercise for a non-qualified option, nothing for an incentive option; units = full value at settlement", "fields": [ { "label": "Restricted stock at the seed ladder", "value": "400,000 shares at 0.2550 costs 102,000 and, with a timely 83(b) election, recognises no income at all", "mono": true }, { "label": "The same 400,000 at the Series C ladder", "value": "1,101,600 of cash, which is why restricted stock stops being used", "mono": true }, { "label": "Options at the Series C ladder", "value": "No cash until exercise, and the exercise cost is the same 1,101,600", "mono": true }, { "label": "Units at any ladder", "value": "No cash ever, and ordinary income on the full value at settlement rather than on a spread", "mono": true }, { "label": "Why units arrive late", "value": "A unit is taxed on its whole value, so it is only workable when there is a market to sell into or a company willing to withhold in shares", "mono": false } ], "notes": [ "The instrument follows the common value, not the stage label. Restricted stock works while the strike is small enough to write a cheque for and stops working the moment it is not; nothing else about the company has to change.", "Units avoid the exercise problem and create a withholding problem: ordinary income arises with no liquidity to pay it. At a private company that means either a company loan, a tender, or a settlement date deferred until a liquidity event.", "Incentive options are the only instrument with no ordinary income at exercise, and the price of that is a statutory holding period, an annual limit, and alternative minimum tax exposure on the spread.", "Whatever the instrument, the payout comes out of the residual after the preferred stack. The instrument decides the tax and the cash; it does not change the position in the waterfall." ] }, { "id": "409a-ladder-and-the-strike", "term": "The 409A ladder - why the strike lags the round price", "definition": "Options must be granted at no less than the fair market value of common stock, determined under section 409A. Preferred stock carries a liquidation preference and other rights that common does not, so the appraised value of common sits below the price paid for preferred in the same round. The gap narrows as the company grows into its stack.", "formula": "Strike = common fair market value. Discount to the preferred price = 1 - (common fair market value / preferred price). The level is an appraisal output and cannot be derived from the round price", "fields": [ { "label": "Seed", "value": "Preferred 0.8500, assumed common fraction 30 percent, strike 0.2550, discount 70.0 percent", "mono": true }, { "label": "Series A", "value": "Preferred 2.0400, 35 percent, strike 0.7140, discount 65.0 percent", "mono": true }, { "label": "Series B", "value": "Preferred 5.1000, 40 percent, strike 2.0400, discount 60.0 percent", "mono": true }, { "label": "Series C", "value": "Preferred 6.1200, 45 percent, strike 2.7540, discount 55.0 percent", "mono": true }, { "label": "What drives the discount", "value": "The size of the aggregate preference relative to enterprise value, the probability weighting across exit scenarios, and marketability", "mono": false }, { "label": "Timing effect", "value": "A grant made before a refresh that follows a strong round carries a lower strike than the identical grant made after it. On this ladder that is a 1.3260 per share difference between the Series A and Series B strike", "mono": true } ], "notes": [ "The fractions in this ladder are inputs. They are chosen so the arithmetic can be followed, and any real strike is the output of an independent appraisal that considers the preference overhang, the scenario weighting and marketability. Modelling a target discount and backing into the value inverts the rule.", "The gap is the same arithmetic as the dead zone between the invested capital and the conversion indifference point, expressed as a valuation input rather than as a payout. A large new preference widens both at once.", "Two numbers are correct on the same day - the preferred price and the common fair market value - and quoting either as the share price will mislead somebody. Always say which one.", "Grant timing relative to a refresh is a scheduling decision with a compensation consequence, and it is one of the few levers a company has that costs it nothing. A refresh triggered by a closing should be planned around, not discovered." ], "source": "IRC section 409A" }, { "id": "early-exercise-and-83b", "term": "Early exercise and the 83(b) election, worked", "definition": "Buying unvested shares and filing an 83(b) election within thirty days fixes the taxable event at the purchase date, when the spread is zero. Without the election, income is recognised at each vesting date on that date's spread, which on a company that works is the whole of the appreciation.", "formula": "With an 83(b) election: ordinary income at purchase = shares * (fair market value - price paid), and all subsequent appreciation is capital. Without it: ordinary income at each vesting date = shares vesting * (fair market value at that date - price paid)", "fields": [ { "label": "Setup", "value": "400,000 shares purchased at the seed strike of 0.2550 for 102,000, vesting 100,000 a year over four years", "mono": true }, { "label": "With a timely 83(b) election", "value": "Ordinary income at purchase = 400,000 * (0.2550 - 0.2550) = 0. Everything after is capital gain", "mono": true }, { "label": "Without it, year 1 at a 0.7140 fair market value", "value": "100,000 * (0.7140 - 0.2550) = 45,900", "mono": true }, { "label": "Year 2 at 2.0400", "value": "178,500", "mono": true }, { "label": "Year 3 at 2.7540", "value": "249,900", "mono": true }, { "label": "Year 4 at 10.2000", "value": "994,500", "mono": true }, { "label": "Total ordinary income with no election", "value": "1,468,800, against zero with the election", "mono": true } ], "notes": [ "The thirty-day window is the whole term. It runs from the transfer of the property, it cannot be extended, and there is no cure. A filing made on day thirty-one produces the right-hand column above for the rest of the grant's life.", "The election accelerates risk as well as tax. Cash goes out at purchase, the shares can become worthless, and there is no deduction for the loss of an amount that was never included in income. It is the correct choice when the purchase price is small and a poor one when it is not.", "The income in years two to four is recognised with no liquidity to pay it, at a private company with no market in its stock. That is the failure mode: a tax bill on paper appreciation that cannot be sold.", "Early exercise combined with 83(b) also starts the capital gains holding period and, for stock that qualifies, the section 1202 holding period at the purchase date rather than at vesting. On a five-year test, three years of difference is the whole exclusion." ], "source": "IRC section 83(b)" }, { "id": "iso-100k-limit", "term": "The 100,000 limit on incentive stock options", "definition": "Incentive stock options are limited by the aggregate fair market value of the underlying stock, measured at grant, that first becomes exercisable in any one calendar year. Options above the limit are treated as non-qualified. The limit is a value, so the share count it permits falls as the strike rises.", "formula": "Shares that can be incentive options in one year = 100,000 / strike, using the fair market value at grant. Anything in excess is non-qualified", "fields": [ { "label": "At the Series A strike of 0.7140", "value": "100,000/0.7140 = 140,056 shares a year", "mono": true }, { "label": "A 200,000-share tranche vesting in one year", "value": "140,056 shares are incentive options and 59,944 are non-qualified", "mono": true }, { "label": "At the seed strike of 0.2550", "value": "100,000/0.2550 = 392,157 shares a year - the limit is not reached by a 200,000-share tranche", "mono": true }, { "label": "At the Series C strike of 2.7540", "value": "100,000/2.7540 = 36,311 shares a year", "mono": true }, { "label": "Measured at grant", "value": "The relevant value is the fair market value when the option was granted, so a later increase in value does not retroactively breach the limit", "mono": true }, { "label": "First becomes exercisable", "value": "A four-year monthly schedule spreads a grant across four calendar years, which is why standard vesting rarely breaches the limit and a cliff-heavy or accelerated schedule can", "mono": true } ], "notes": [ "The limit bites on acceleration, not on ordinary vesting. A change of control that accelerates several years of vesting into one calendar year can convert a large part of an incentive grant into non-qualified options at the worst possible moment.", "It also bites on late grants at a high strike. The same 200,000 options are entirely incentive at a 0.2550 strike and largely non-qualified at 0.7140, from nothing but the passage of two rounds.", "The split is determined at the plan and grant level and should appear on the grant documentation. A holder that assumes a grant is entirely incentive and exercises accordingly will find withholding due on the non-qualified part.", "Non-qualified treatment is not a disaster; it is ordinary income on the spread with withholding, and no alternative minimum tax exposure. In a year when the spread is small it is arguably the better outcome." ], "source": "IRC section 422(d)" }, { "id": "amt-on-iso-exercise", "term": "Alternative minimum tax on an incentive option exercise", "definition": "An incentive option produces no ordinary income at exercise, and the spread is an item of adjustment for alternative minimum tax. The result is a cash tax liability in the year of exercise, on stock that cannot be sold, computed on a value nobody received.", "formula": "Adjustment amount = shares exercised * (fair market value at exercise - strike). This enters the alternative minimum tax base in the year of exercise, whether or not the shares are sold", "fields": [ { "label": "Setup", "value": "Series A vintage options at a 0.7140 strike, exercised when the common fair market value is 2.0400", "mono": true }, { "label": "Exercising 140,056 shares", "value": "Spread = 140,056 * (2.0400 - 0.7140) = 185,714", "mono": true }, { "label": "Cash paid to exercise", "value": "100,000", "mono": true }, { "label": "Cash available from the shares", "value": "None - the stock is illiquid", "mono": true }, { "label": "Exercising 200,000 instead", "value": "Spread = 265,200, and the excess over the incentive limit is ordinary income rather than an adjustment", "mono": true }, { "label": "If the shares are sold in the same year", "value": "The disposition is disqualifying, the spread becomes ordinary income, and the adjustment does not arise", "mono": true } ], "notes": [ "The structural problem is that the adjustment is recognised on exercise and the liquidity arrives on sale, which can be years later or never. Exercising into a falling private valuation is the case that has bankrupted employees.", "Selling in the same calendar year as the exercise is a disqualifying disposition and converts the whole spread to ordinary income, which removes the adjustment. That is a real planning choice rather than a failure, and it is the only route available where a tender is the source of liquidity.", "The exposure scales with the spread, so exercising early - immediately after a grant, when the spread is zero - avoids it entirely. That is the main argument for early exercise on top of the 83(b) point.", "This corpus states the mechanism, not the computation of anyone's liability. Rates, exemptions, credits and the interaction with ordinary tax are outside its scope." ], "source": "IRC section 56(b)(3)" }, { "id": "qsbs-1202-structure", "term": "Qualified small business stock - the tests and the ceiling", "definition": "Section 1202 excludes gain on qualified small business stock from federal income tax, subject to a holding period, tests on the issuer, and a per-issuer ceiling. The structure is what matters here: the exclusion is capped, and the cap is the greater of a statutory dollar amount and ten times the holder's basis - which is nothing at all for a founder who paid par.", "formula": "Excluded gain = tier fraction * min(gain, ceiling), where ceiling = max(K, 10 * aggregate adjusted basis of the stock disposed of) and K, the tier fractions and the holding periods are statutory parameters", "fields": [ { "label": "Structural requirements", "value": "Stock acquired at original issue from a domestic C corporation; the corporation's gross assets below a statutory ceiling at and immediately after issuance; an active business requirement; a minimum holding period", "mono": false }, { "label": "Parameters used below", "value": "A ceiling K of 15,000,000 and a 100 percent tier at five years. These are inputs to the arithmetic. Read the current statute for the figures that apply to a given holding, which were amended for stock acquired after 4 July 2025", "mono": false }, { "label": "Worked, a founder position", "value": "8,000,000 founder shares with an aggregate basis of 8,000, sold at an exit of 250,000,000 where the shares are worth 10.2000 each", "mono": true }, { "label": "Proceeds and gain", "value": "81,600,000 of proceeds, 81,592,000 of gain", "mono": true }, { "label": "The ceiling", "value": "max(15,000,000, 10 * 8,000) = max(15,000,000, 80,000) = 15,000,000", "mono": true }, { "label": "Excluded and taxable", "value": "15,000,000 excluded, which is 18.384 percent of the gain; 66,592,000 remains taxable", "mono": true } ], "notes": [ "The ten-times-basis alternative is the part that surprises founders. It is designed for holders who paid real money for their stock, and a founder who paid par has a basis so small that the alternative is irrelevant. The dollar ceiling is the whole benefit.", "The gross assets test is measured at and immediately after issuance, so it is the early rounds that qualify and the later ones that can fail. Stock issued after the company has crossed the asset ceiling is not qualified stock, which means a single cap table can contain both.", "The holding period runs from issuance. Early exercise with an 83(b) election starts it years earlier than exercise at a liquidity event, which on a five-year test is frequently the difference between the full exclusion and none.", "A conversion of the company to a limited liability company or an S corporation, or an issuance by a non-corporate entity, breaks the test at the root. This is one of the few tax attributes that is destroyed by an entity choice rather than by a transaction." ], "source": "IRC section 1202" }, { "id": "option-repricing-and-exchange", "term": "Repricing and exchange offers for underwater options", "definition": "After a down round the outstanding grants have strikes above the new common value and no retention effect. The two remedies are to reprice the existing options to the new value, or to offer an exchange of old options for a smaller number of new ones. Both are a real cost and neither is arithmetic alone.", "formula": "A value-for-value exchange ratio is old options per new option, set so the model value of the new grant equals the model value surrendered. The ratio is an option-pricing output, not a spread calculation, because both grants are usually at or above the money", "fields": [ { "label": "Setup", "value": "200,000 options at a 2.0400 strike from the Series B ladder. A down round resets the common fair market value to 0.7140", "mono": true }, { "label": "Intrinsic value of the old grant", "value": "200,000 * max(0, 0.7140 - 2.0400) = 0 - the grant is underwater and its spread carries no information", "mono": true }, { "label": "A straight reprice", "value": "Strike moves to 0.7140, the share count is unchanged, and the accounting charge is the incremental fair value of the modification", "mono": true }, { "label": "An exchange at 2 old for 1 new", "value": "100,000 new options at 0.7140. Whether that is value-for-value depends on the volatility and term assumptions, not on the strikes alone", "mono": true }, { "label": "Value at a 10.2000 exit", "value": "Repriced: 200,000 * (10.2000 - 0.7140) = 1,897,200. Exchanged: 100,000 * (10.2000 - 0.7140) = 948,600", "mono": true }, { "label": "Incentive option consequences", "value": "A reprice is a new grant for incentive option purposes, which restarts the holding period and re-tests the annual limit at the new strike", "mono": true } ], "notes": [ "The exchange ratio is where the negotiation happens and it cannot be settled from the strikes. Two grants both at the money have the same intrinsic value and very different model values depending on term and volatility, so the ratio is an assumption dressed as arithmetic.", "A reprice is simpler, more generous, and harder to justify to investors who have just taken a markdown. An exchange at a ratio above one recovers pool shares, which is often the real reason it is chosen.", "Both need board approval, usually a plan amendment, and often stockholder approval; tender offer rules can apply to an exchange offered broadly to employees. The process is a larger obstacle than the arithmetic.", "The alternative that costs nothing in shares is to leave the old grants outstanding and make new grants at the new strike. It dilutes the pool faster and it avoids every process question above." ] }, { "id": "exercise-window-for-leavers", "term": "The post-termination exercise window, and what it costs to use", "definition": "A departing holder typically has a short window - commonly three months - to exercise vested options or lose them. Exercising means paying the strike and, for a non-qualified option, the tax on the spread, in cash, for stock that cannot be sold.", "formula": "Cash required = shares * strike + ordinary rate * shares * (fair market value - strike) for a non-qualified option. For an incentive option the second term is replaced by alternative minimum tax exposure on the same spread", "fields": [ { "label": "Setup", "value": "200,000 vested non-qualified options at the Series A strike of 0.7140. The holder leaves after the Series C round, when the common fair market value is 2.7540", "mono": true }, { "label": "Exercise cost", "value": "200,000 * 0.7140 = 142,800", "mono": true }, { "label": "Spread", "value": "200,000 * (2.7540 - 0.7140) = 408,000", "mono": true }, { "label": "Tax at an assumed 40 percent", "value": "163,200", "mono": true }, { "label": "Total cash within the window", "value": "306,000, for stock with no market", "mono": true }, { "label": "If the window lapses", "value": "The options are cancelled and the shares return to the pool, available for regrant to somebody else", "mono": true }, { "label": "Value forgone at a 10.2000 exit", "value": "200,000 * (10.2000 - 0.7140) = 1,897,200", "mono": true } ], "notes": [ "The window converts a grant into a purchase decision taken at the worst moment, with the least information and usually the least cash. That is the design: a short window transfers the value of unexercised grants back to the company and its remaining holders.", "Extended windows - measured in years rather than months - solve it and have a cost. An incentive option becomes non-qualified three months after termination as a matter of statute, so an extended window converts the instrument even if the plan says nothing.", "The shares returning to the pool are the quiet economics of the term. On a large workforce, forfeited grants are a material source of pool capacity, which is why the term survives.", "Where the window is short and the strike is high, the option is worth what the holder can borrow against it, which is usually nothing. Financing arrangements exist for exactly this and they are secured on an asset with no market." ] }, { "id": "net-exercise-and-sell-to-cover", "term": "Net exercise and sell-to-cover, worked", "definition": "Both mechanics let a holder exercise without writing a cheque, and they are different transactions. A net exercise withholds shares to pay the strike and delivers the balance. Sell-to-cover exercises in full and sells shares into the market to fund the strike and the tax.", "formula": "Net exercise delivers n*(FMV - strike)/FMV shares. Sell-to-cover delivers n - (n*strike + tax)/FMV shares, where tax is the ordinary rate applied to the full spread", "fields": [ { "label": "Setup", "value": "200,000 options at the Series A strike of 0.7140, exercised at 10.2000 a share. Assumed ordinary rate 40 percent", "mono": true }, { "label": "Net exercise", "value": "200,000 * (10.2000 - 0.7140)/10.2000 = 186,000 shares delivered", "mono": true }, { "label": "Value of those shares", "value": "1,897,200, exactly equal to the spread", "mono": true }, { "label": "Sell-to-cover, cash needed", "value": "Strike 142,800 plus tax 758,880 = 901,680", "mono": true }, { "label": "Shares sold", "value": "901,680/10.2000 = 88,400, leaving 111,600 shares", "mono": true }, { "label": "Comparison", "value": "Net exercise leaves 186,000 shares and an unpaid tax bill; sell-to-cover leaves 111,600 shares and nothing outstanding", "mono": true } ], "notes": [ "Net exercise does not pay the tax. It funds the strike out of shares and leaves the withholding to be settled separately, which is why it is usually combined with share withholding for taxes rather than used alone.", "The share count delivered on a net exercise is independent of the tax rate and depends only on the ratio of strike to fair market value. That makes it the cleaner mechanic to model and the one to use when the tax treatment is uncertain.", "Both mechanics reduce the shares outstanding relative to a cash exercise, so both reduce dilution. A company that permits net exercise across a large pool has a materially smaller fully diluted count at exit than the grant totals suggest.", "Sell-to-cover needs a market. At a private company the equivalent is a tender or a company-arranged sale, which means the mechanic is only available at the company's discretion and on its timetable." ] }, { "id": "one-percent-grant-arithmetic", "term": "What a one percent grant is worth after four rounds", "definition": "A grant quoted as a percentage is a share count computed on the fully diluted count on the day of the grant. The percentage then behaves exactly like any other static holding: it is multiplied by every subsequent round's retention factor, and nobody sends a notice when it changes.", "formula": "Shares = q * FD_grant. Ownership at exit = shares / FD_exit = q * product of PRE_i/POST_i over the rounds after the grant. Value = shares * exit price per share", "fields": [ { "label": "A 1.000 percent grant at the seed ledger", "value": "0.01000 * 11,764,706 = 117,647 shares", "mono": true }, { "label": "The same shares after Series C", "value": "117,647/24,509,805 = 0.480 percent", "mono": true }, { "label": "Check by retention factors", "value": "1.000 * 0.75000 * 0.80000 * 0.80000 = 0.480 percent", "mono": true }, { "label": "Value at a 250,000,000 exit", "value": "117,647 * 10.2000 = 1,200,000", "mono": true }, { "label": "Value at a 90,000,000 exit", "value": "Common and pool receive 24,000,000 across 9,411,765 shares, or 2.5500 a share, so the grant is worth 300,000", "mono": true }, { "label": "Value at a 60,000,000 exit", "value": "Zero - the aggregate preference is 60,000,000 and common receives nothing at or below it", "mono": true } ], "notes": [ "The percentage falls by more than half over four rounds with no pool refresh, and every round it falls is a round the company describes as good news. A grant negotiated as a percentage should be recorded as a share count, because the share count is the thing that was actually granted.", "The value at a 90,000,000 exit is a small fraction of the naive 0.480 percent of 90,000,000, because common shares only the residual after the preference. The naive calculation is correct only above the last conversion flip point, which here is 150,000,000.", "Three numbers make a grant legible: the share count, the fully diluted count on the grant date, and the aggregate preference. All three are known to the company and none of them is normally on the grant notice.", "Refresh grants exist to offset exactly this decay, and they are made at the current strike. A holder whose original grant has halved in percentage terms and whose strike has risen eightfold is not made whole by a top-up of the same percentage." ] }, { "id": "option-value-under-the-overhang", "term": "A grant is priced on common and paid from the residual", "definition": "The strike is set to the fair market value of common. The payout comes from whatever is left after the preferred stack. Those two facts are consistent - the appraisal reflects the overhang - and together they mean the exit value at which a grant becomes worth anything is the aggregate preference, not the strike.", "formula": "Payout per share = max(0, residual to common per share - strike). The residual per share is (E - non-converting preferences - carve-out) / (common shares + converting preferred shares), which is zero for E at or below T", "fields": [ { "label": "Setup", "value": "A 200,000-share grant at the Series A strike of 0.7140. Aggregate preference T = 60,000,000", "mono": true }, { "label": "At E = 60,000,000", "value": "Residual per share 0.0000, so the grant is worth nothing despite a 2.0400 common valuation at the time of the Series B", "mono": true }, { "label": "At E = 90,000,000", "value": "Common and pool receive 24,000,000 over 9,411,765 shares = 2.5500 a share; the grant is worth 200,000 * (2.5500 - 0.7140) = 367,200", "mono": true }, { "label": "At E = 150,000,000", "value": "Everything converts; the residual per share is 150,000,000/24,509,805 = 6.1200; the grant is worth 1,081,200", "mono": true }, { "label": "At E = 250,000,000", "value": "1,897,200", "mono": true }, { "label": "Break-even exit value", "value": "The grant is worth nothing below E = 60,000,000 and turns positive once the residual per share exceeds 0.7140", "mono": true } ], "notes": [ "The grant is an option on the residual, struck at the common fair market value. Describing it as an option on the company overstates it by the entire preference, which on this cap table is 60,000,000 of exit value.", "This is why the aggregate preference is the number to tell a grantee. The strike and the share count are on the grant notice and neither of them says at what outcome the grant pays anything.", "The residual is shared with the unallocated pool in many structures, so a grantee dividing by the common it knows about will overstate its own share. Divide by the full residual group.", "The company's own 409A appraisal already contains this analysis, because the scenario weighting is exactly a calculation of the residual across exit values. The information exists; it is simply not the part that gets communicated." ] }, { "id": "restricted-stock-versus-units-at-a-private-company", "term": "Why units are hard at a private company", "definition": "A restricted stock unit is taxed on its full value at settlement, with withholding due in cash. At a public company the shares are sold to fund it. At a private company there is nothing to sell, so the settlement date has to be engineered around the absence of a market.", "formula": "Tax at settlement = ordinary rate * shares * fair market value at settlement. Shares withheld to cover = shares * ordinary rate, independent of the price", "fields": [ { "label": "Setup", "value": "100,000 units settling when the common fair market value is 2.7540. Assumed ordinary rate 40 percent", "mono": true }, { "label": "Income at settlement", "value": "100,000 * 2.7540 = 275,400", "mono": true }, { "label": "Tax", "value": "110,160", "mono": true }, { "label": "Shares withheld to cover", "value": "100,000 * 0.40000 = 40,000 - the count depends only on the rate, not on the price", "mono": true }, { "label": "The company's problem", "value": "Withholding shares means the company remits cash it has not received, so a large settlement is a cash outflow for the company as well as a tax event for the holder", "mono": true }, { "label": "The usual answer", "value": "A double-trigger settlement condition: units vest on service and settle only on a liquidity event, so the tax and the liquidity arrive together", "mono": true } ], "notes": [ "The share count withheld to cover is the ordinary rate multiplied by the share count, at any price. That is the one genuinely simple piece of arithmetic in employee equity and it is worth knowing: a 40 percent rate costs 40 percent of the shares.", "Double-trigger settlement is the standard fix and it creates its own problem: the units are a deferred compensation arrangement and the settlement condition has to be drafted to fit within the applicable rules. This is where section 409A applies to units rather than to option strikes.", "Restricted stock avoids all of this by transferring the shares up front, which is why it persists at the earliest stage and disappears once the purchase price becomes real money.", "For the holder the difference is stark: an option can be left unexercised and a unit cannot be left unsettled. A unit is a certainty of tax and an option is a choice." ], "source": "IRC section 409A (deferred compensation) and section 83" } ] }, { "slug": "structures", "title": "Rounds, structures and bridges", "lede": "Bridges, inside rounds, structured terms, venture debt and the arithmetic of each.", "intro": "Not every financing is a clean priced round. A bridge prices off a round that has not happened; an extension prices off a round that already has; an inside round has no external price at all; and a structured round has a headline valuation that no longer describes the economics. Each of these has an arithmetic that can be written down, and in every case the useful question is the same one: what does this cost the common stock, measured at a stated exit value. All figures continue the share ledger used throughout this corpus, where the post-Series B count is 19,607,844 shares at a 5.1000 price and the Series C round raises 30,000,000 at a 120,000,000 pre-money.", "reviewed": "2026-08-27", "tables": [ { "title": "A 5,000,000 bridge at a 20 percent discount, across next-round prices", "intro": "The bridge converts into the next priced round at 80 percent of that round's price. Because its shares sit inside the pre-money count, the round price and the bridge's share count are simultaneous and solve as S = FD/(1 - amount/((1 - discount)*PRE)). The next round raises 30,000,000 in every column. The last column is the value of the bridge holder's shares at the round price, against the 5,000,000 advanced.", "columns": [ "Next round pre-money", "Round price", "Bridge conversion price", "Bridge shares", "Total shares after", "Bridge ownership", "New investor", "Common and pool", "Bridge value at the round price" ], "rows": [ [ "60,000,000", "2.7412", "2.1930", "2,279,982", "32,831,739", "6.944", "33.333", "28.667", "1.25" ], [ "90,000,000", "4.2712", "3.4170", "1,463,272", "28,094,821", "5.208", "25.000", "33.500", "1.25" ], [ "120,000,000", "5.8012", "4.6410", "1,077,354", "25,856,498", "4.167", "20.000", "36.400", "1.25" ], [ "180,000,000", "8.8612", "7.0890", "705,318", "23,698,689", "2.976", "14.286", "39.714", "1.25" ] ] }, { "title": "Structured terms on the same Series C, priced in common stock", "intro": "Series C invests 30,000,000 in every row and Seed, Series A and Series B stay at 1x non-participating. Only the Series C structure changes. T is the aggregate preference. The last column is the cost to common and the option pool at a 150,000,000 exit, measured against the clean 1x row.", "columns": [ "Series C structure", "Aggregate preference T", "Series C at E = 90,000,000", "Common and pool at E = 90,000,000", "Series C at E = 150,000,000", "Common and pool at E = 150,000,000", "Cost to common at 150,000,000" ], "rows": [ [ "Clean 1x non-participating", "60,000,000", "30,000,000", "24,000,000", "30,000,000", "57,600,000", "0" ], [ "2x non-participating", "90,000,000", "60,000,000", "0", "60,000,000", "42,000,000", "15,600,000" ], [ "3x non-participating", "120,000,000", "67,500,000", "0", "90,000,000", "24,000,000", "33,600,000" ], [ "1x full participating", "60,000,000", "39,411,765", "18,070,588", "53,809,524", "45,714,286", "11,885,714" ], [ "1.5x participating, capped at 2x", "75,000,000", "50,000,000", "9,600,000", "60,000,000", "42,000,000", "15,600,000" ], [ "2x participating, capped at 3x", "90,000,000", "60,000,000", "0", "76,666,667", "32,000,000", "25,600,000" ] ] }, { "title": "Venture debt against equity for the same 5,000,000", "intro": "A 5,000,000 facility with 20 percent warrant coverage, against 5,000,000 of equity sold at the Series B price of 5.1000. Warrant coverage is expressed as a percentage of the principal, so 20 percent coverage buys warrants over 1,000,000 of stock at the stated exercise price. Interest, fees and covenants are outside the dilution comparison and are the reason the comparison is not the whole decision.", "columns": [ "Quantity", "Venture debt with warrants", "Equity at 5.1000" ], "rows": [ [ "Cash received", "5,000,000", "5,000,000" ], [ "Shares issued now", "0", "980,392" ], [ "Shares issuable later", "196,078 on warrant exercise", "0" ], [ "Fully diluted count after", "19,803,922", "20,588,236" ], [ "Dilution to existing holders", "0.990", "4.762" ], [ "Ranking in an exit", "Principal ahead of the entire preferred stack", "Behind or alongside the existing preferred, per its terms" ], [ "Adds to the aggregate preference", "No - it is debt, and it is repaid before the waterfall runs", "Yes, by 5,000,000 at 1x" ], [ "Cost if the company fails", "Repayable, and secured", "Nothing further" ], [ "Value of the warrants at a 10.2000 exit", "1,000,000", "-" ] ] } ], "entries": [ { "id": "bridge-note-into-the-next-round", "term": "A bridge prices off a round that has not happened yet", "definition": "A bridge converts at a discount to the next round's price, and its own shares are inside the pre-money count used to set that price. The two facts make the calculation simultaneous, and they mean the bridge holder's return depends on the next round being priced well - which is the opposite of the protection a bridge investor thinks it is buying.", "formula": "S = FD / (1 - amount / ((1 - discount) * PRE)). Round price = PRE/S, conversion price = (1 - discount)*PRE/S, bridge shares = amount divided by that conversion price", "fields": [ { "label": "Setup", "value": "5,000,000 bridge at a 20 percent discount, on the post-Series B ledger of 19,607,844 shares. The next round raises 30,000,000", "mono": true }, { "label": "At PRE = 120,000,000", "value": "S = 20,685,198, round price 5.8012, conversion price 4.6410, bridge shares 1,077,354 = 4.167 percent after the round", "mono": true }, { "label": "At PRE = 60,000,000", "value": "Round price 2.7412, conversion price 2.1930, bridge shares 2,279,982 = 6.944 percent", "mono": true }, { "label": "At PRE = 180,000,000", "value": "Bridge shares 705,318 = 2.976 percent", "mono": true }, { "label": "Bridge value at the round price", "value": "At 120,000,000: 6,250,000 on 5,000,000 advanced, a multiple of 1.25 - which is 1/(1 - discount) at every price", "mono": true }, { "label": "Effect on the round price", "value": "Without the bridge the price would be 120,000,000/19,607,844 = 6.1200; with it, 5.8012", "mono": true } ], "notes": [ "The discount delivers a fixed multiple of 1/(1 - discount) regardless of the round price, so a 20 percent discount is a 1.25x on the money and nothing more. A bridge investor taking equity risk for a 1.25x is being paid a debt return for an equity position.", "That is why bridges are usually written with a cap as well as a discount, converting at the better of the two. The discount sets the floor return and the cap provides the upside if the round prices well above where the bridge was written.", "The bridge shares depress the round price, which the new investor is indifferent to - its percentage is R/POST either way. The cost lands on the pre-money holders, exactly as with a pool or a converting instrument.", "A bridge with no cap and no discount is a loan, and it should be documented and priced as one. Convertibility with no economic term attached gives the holder the downside of equity and the upside of debt." ] }, { "id": "extension-round", "term": "An extension round is a flat price at a later date", "definition": "An extension sells more of the same series at the same price, months after the original closing. In valuation terms it is a flat round; in economic terms it is a down round if the business is worth more than it was, and an up round if it is worth less. Nothing in the documents records which.", "formula": "Shares = amount / the original round price. Implied pre-money = original price * the current fully diluted count, which rises as the count rises even though the price has not moved", "fields": [ { "label": "Setup", "value": "5,000,000 added to the Series B at its 5.1000 price, on a count of 19,607,844 shares", "mono": true }, { "label": "Shares issued", "value": "980,392, taking the count to 20,588,236", "mono": true }, { "label": "Implied pre-money", "value": "5.1000 * 19,607,844 = 100,000,004, equal to the Series B post-money up to share-count rounding", "mono": true }, { "label": "Dilution to existing holders", "value": "4.762 percent; founders move from 40.800 to 38.857 percent", "mono": true }, { "label": "Aggregate preference", "value": "Rises from 30,000,000 to 35,000,000, so the first dollar to common moves up 5,000,000", "mono": true }, { "label": "Anti-dilution", "value": "Not triggered, because the price has not fallen. That is the main documentary attraction of an extension", "mono": true } ], "notes": [ "An extension is the cheapest financing to document because it uses the existing charter and the existing series. It needs no new anti-dilution analysis, no new protective provisions and often no new consents, which is why it is reached for under time pressure.", "It is also the financing that most reliably understates what happened. A company that has doubled its revenue and extends at the old price has taken a real down round in economic terms, and the cap table records a flat one.", "Check whether the extension shares carry the original series' terms including the original accrual start date for any cumulative dividend. Same series, later money, older dividend clock is a combination worth reading.", "Where an extension is being used to avoid triggering anti-dilution, the honest alternative is a priced down round with a negotiated reset. The extension defers the reset rather than avoiding it, and the deferral is paid for by the common." ] }, { "id": "inside-round-conflict-management", "term": "An inside round has no external price, so the price is the conflict", "definition": "When the existing investors fund the round, the people setting the price are on both sides of it. The arithmetic of the transfer is simple and computable, which is what makes the process protections around an inside round worth running rather than assuming.", "formula": "Extra shares to the insider from pricing at PRE_low rather than PRE_high = R*FD*(1/PRE_low - 1/PRE_high). Their value at the higher price is R*(PRE_high/PRE_low - 1)", "fields": [ { "label": "Setup", "value": "20,000,000 from existing holders on the post-Series C ledger of 24,509,805 shares", "mono": true }, { "label": "Priced at PRE = 120,000,000", "value": "Price 4.8960, shares 4,084,968, insider holds 14.286 percent of the enlarged company", "mono": true }, { "label": "Priced at PRE = 90,000,000", "value": "Price 3.6720, shares 5,446,623, insider holds 18.182 percent", "mono": true }, { "label": "Extra shares from the lower price", "value": "1,361,656, worth 6,666,667 at the higher price", "mono": true }, { "label": "Formula check", "value": "20,000,000 * (120,000,000/90,000,000 - 1) = 6,666,667", "mono": true }, { "label": "Process protections", "value": "Approval by directors without an interest in the financing; an independent valuation; and a rights offering giving every holder the chance to invest on the same terms", "mono": false } ], "notes": [ "The rights offering is the protection that does the most work, because it converts a price dispute into a choice. A holder that thinks the price is too low can buy at it, and a holder that declines has accepted the price by conduct.", "A disinterested director committee is the standard second protection, and it requires that such a director exists. On a board of two founders and three investors all of whom are funding the round, it does not.", "The arithmetic above is the entire transfer, and it is worth computing and circulating before the round rather than after. A 30,000,000 difference in pre-money on a 20,000,000 round is 6,666,667 of value moving between existing holders.", "Delaware review of a self-interested transaction is a matter of process and price together. Nothing in this corpus is legal advice, but the structural point is uncontroversial: the record of how the price was set is part of the transaction." ], "source": "Delaware General Corporation Law section 144 (interested director transactions)" }, { "id": "pay-to-play-bridge-sizing", "term": "Sizing a pay-to-play bridge across the stack", "definition": "A pay-to-play bridge is offered pro rata to the existing preferred, with conversion to common as the penalty for declining. Sizing it means allocating the amount across the series in proportion to their holdings, and the allocation is what determines who is being asked for what.", "formula": "Series j's pro rata share of a bridge of B = B * p_j / p_class, where p_class is the preferred's aggregate ownership. A holder that declines converts its preferred to common and gives up x_j*I_j of preference", "fields": [ { "label": "Setup", "value": "A 6,000,000 bridge offered pro rata across the four-series preferred, which holds 61.600 percent of the company", "mono": true }, { "label": "Allocation", "value": "Seed 935,065, Series A 1,558,442, Series B 1,558,442, Series C 1,948,052", "mono": true }, { "label": "Check", "value": "6,000,000 total", "mono": true }, { "label": "Cost of declining, Seed", "value": "A 935,065 cheque against a 2,000,000 preference, a ratio of 2.14 to 1", "mono": true }, { "label": "Cost of declining, Series C", "value": "A 1,948,052 cheque against a 30,000,000 preference, a ratio of 15.40 to 1", "mono": true }, { "label": "Preference protected per dollar of bridge", "value": "Seed 2.14 to 1, Series A 5.13 to 1, Series B 12.83 to 1, Series C 15.40 to 1", "mono": true }, { "label": "Why the ratios differ", "value": "The pro rata allocation follows ownership and the preference at risk follows invested capital. The later series has far more preference per dollar of ownership, so the clause bites hardest on it", "mono": true } ], "notes": [ "The clause is written as a penalty and functions as a pricing mechanism: it sets the cost of not funding equal to the whole existing position, which is what makes an insider bridge happen when no outside price exists.", "The ratios above are the real content. The Seed holder is being asked for a small cheque to protect a small preference and the Series C holder for a larger cheque to protect a much larger one, so the same clause produces very different decisions around the table.", "It also concentrates the cap table exactly when the company can least afford a governance fight, because holders that convert lose their consent rights at the same moment.", "From the holder's side the decision is not the cheque against the position's book value. It is the cheque against the position's value in the outcomes that made the bridge necessary." ] }, { "id": "structured-terms-priced-in-common", "term": "Every structured term can be quoted as a number of dollars taken from common", "definition": "Multiples, participation, caps and ratchets are argued as structures and settle as amounts. Fixing an exit value and running the waterfall converts each proposal into a single figure, which is the only form in which they can be compared with each other or with a lower price.", "formula": "Cost to common of a structure = common's proceeds under a clean 1x non-participating alternative, less its proceeds under the structure, both measured at the same E", "fields": [ { "label": "Reference case", "value": "Clean 1x Series C at 150,000,000: Series C 30,000,000, common and pool 57,600,000", "mono": true }, { "label": "2x non-participating", "value": "Series C 60,000,000, common and pool 42,000,000 - a cost of 15,600,000", "mono": true }, { "label": "1x full participating", "value": "Series C 53,809,524, common and pool 45,714,286 - a cost of 11,885,714", "mono": true }, { "label": "1.5x participating capped at 2x", "value": "Series C 60,000,000, common and pool 42,000,000 - the same 15,600,000, by a different route", "mono": true }, { "label": "3x non-participating", "value": "Series C 90,000,000, common and pool 24,000,000 - a cost of 33,600,000", "mono": true }, { "label": "At a lower exit of 90,000,000", "value": "The clean case pays common 24,000,000; a 2x pays it 0; a 3x pays it 0", "mono": true } ], "notes": [ "Two structures that look nothing alike can cost the same. A 2x non-participating and a 1.5x participating capped at 2x both deliver Series C exactly 60,000,000 at this exit, because both are binding at the cap. Which is worse depends entirely on the exit value chosen.", "That is the reason to price a structure at three exit values rather than one: below the aggregate preference every structure costs common nothing, and far above the caps they converge again. The differences live in the middle.", "Once the cost is a number, it can be traded against the pre-money. A structure that costs common 15,600,000 at the likely exit is worth conceding only if the alternative clean price costs more than that.", "Structured terms also carry a second cost that does not appear in the waterfall: they widen the gap between the preferred price and the common fair market value, which lowers option strikes and raises the 409A discount. That is a benefit to employees and a signal to everyone else." ] }, { "id": "ipo-ratchet", "term": "An IPO ratchet, and the shares it issues", "definition": "A ratchet guarantees a holder a stated value at a public offering by issuing additional shares if the offering prices below a floor. The additional shares come from everyone else, and the number rises hyperbolically as the offering price falls.", "formula": "Shares required = guaranteed value / offering price. Additional shares issued = max(0, that figure - shares already held). Every other holder is diluted by additional/(FD + additional)", "fields": [ { "label": "Setup", "value": "Series C holds 4,901,961 shares from a 30,000,000 investment, with a 1x ratchet at a floor equal to its 6.1200 issue price", "mono": true }, { "label": "Offering at 6.1200", "value": "Shares required 4,901,961, already held 4,901,961, additional issued 0", "mono": true }, { "label": "Offering at 5.0000", "value": "Required 6,000,000, additional 1,098,039, count rises to 25,607,844, common and pool falls to 36.753 percent", "mono": true }, { "label": "Offering at 4.0000", "value": "Required 7,500,000, additional 2,598,039, count rises to 27,107,844, common and pool falls to 34.720 percent", "mono": true }, { "label": "Offering at 2.0000", "value": "Required 15,000,000 shares, more than half the pre-offering count", "mono": true }, { "label": "The asymmetry", "value": "The holder's value is fixed at 30,000,000 across every price above zero, and the entire variance is transferred to the other holders", "mono": true } ], "notes": [ "A ratchet is a put option on the offering price, written by the common. Priced as an option it is expensive, and it is negotiated as a technical condition rather than as the transfer of variance that it is.", "Because the share count is inversely proportional to the price, the ratchet is worthless in the outcome everyone is planning for and enormous in the outcome nobody is. That shape is what makes it hard to argue about with numbers taken from the plan.", "The countermeasures are a floor on the number of additional shares, a cap on the total, a sunset date, and a requirement that the ratchet lapse if the company achieves stated milestones. All four are arithmetic and all four are drafting.", "A ratchet also complicates the offering itself, because underwriters have to describe a share count that depends on the price they are setting. That practical friction is often a better argument against it than the economics." ] }, { "id": "venture-debt-sizing-and-warrant-coverage", "term": "Venture debt, warrant coverage and the dilution comparison", "definition": "Venture debt is priced in interest and fees and dilutes through warrants. Warrant coverage is quoted as a percentage of the principal, which converts to a share count at the stated exercise price. Compared with equity for the same cash the dilution is far smaller, and the claim in an exit is far more senior.", "formula": "Warrant shares = coverage * principal / exercise price. Dilution = warrant shares / (FD + warrant shares). Equity for the same cash issues amount/price shares", "fields": [ { "label": "Setup", "value": "A 5,000,000 facility with 20 percent warrant coverage, exercise price at the Series B price of 5.1000, on a count of 19,607,844", "mono": true }, { "label": "Warrant shares", "value": "0.20000 * 5,000,000 / 5.1000 = 196,078 shares = 0.990 percent", "mono": true }, { "label": "Equity for the same 5,000,000", "value": "5,000,000/5.1000 = 980,392 shares = 4.762 percent", "mono": true }, { "label": "Ratio", "value": "Equity is 5.00 times as dilutive as the warrants for the same cash", "mono": true }, { "label": "Ranking", "value": "The principal is repaid before the waterfall runs, so it reduces E rather than adding to T. Debt is ahead of the entire preferred stack", "mono": true }, { "label": "Warrant value at a 10.2000 exit", "value": "1,000,000, against 1,000,000 of nominal coverage", "mono": true } ], "notes": [ "The dilution comparison is the part of the decision that is arithmetic, and it always favours debt. Everything that argues the other way - covenants, a material adverse change clause, an amortisation schedule that starts before revenue does, and a lender that can accelerate - is not in the comparison.", "Debt reduces the exit proceeds available to every equity holder, so it is senior to the entire stack including the most recent round. That makes an indebtedness threshold in the protective provisions one of the more consequential list entries.", "Facilities sized against a recent equity round rather than against cash flow are effectively lending against the next round happening. If the next round does not happen, the facility is what determines the timetable.", "Warrants are usually not in the fully diluted count used to price the next round unless someone asks, which is the same denominator question as everywhere else in this corpus - and worth roughly 1 percent of the price here." ] }, { "id": "revenue-based-financing-arithmetic", "term": "Revenue-based financing - the cap is not the rate", "definition": "A revenue-based facility advances cash against a fixed repayment cap, collected as a percentage of revenue. The cap looks like a multiple and behaves like a rate, because the same 1.4x repaid quickly is much more expensive than the same 1.4x repaid slowly. Speed is the whole cost.", "formula": "Total repayment = cap * advance. Monthly payment = share * monthly revenue. Duration = total repayment / monthly payment. The implied rate solves sum of payment/(1+i)^t over the duration = advance", "fields": [ { "label": "Setup", "value": "1,000,000 advanced, a 1.4x repayment cap, collected at 8 percent of a flat 500,000 of monthly revenue", "mono": true }, { "label": "Total repayment", "value": "1,400,000; monthly payment 0.08000 * 500,000 = 40,000", "mono": true }, { "label": "Duration", "value": "1,400,000/40,000 = 35.0 months", "mono": true }, { "label": "Implied rate", "value": "1.9996 percent a month, an effective annual rate of 26.82 percent", "mono": true }, { "label": "Same cap repaid over 18 months", "value": "Payment 77,778 a month: 3.8099 percent a month, 56.63 percent a year", "mono": true }, { "label": "The asymmetry", "value": "Growing faster repays sooner and costs more. A 1.4x cap on a company doubling revenue is roughly twice the annual rate of the same cap on a flat one", "mono": true } ], "notes": [ "The cap is the only term quoted and the duration is the term that sets the price, so the instrument is systematically cheaper-looking than it is. Compute the implied annual rate under the company's own revenue plan before comparing it with anything.", "The incentive is genuinely perverse: outperforming the plan accelerates collection and raises the effective rate, which is the opposite of every other financing instrument. A cap that steps down with duration fixes it and is rare.", "Collections scale with revenue, so the instrument is self-limiting in a downturn, which is its real attraction. It converts a fixed obligation into a variable one and prices the option accordingly.", "There is usually no dilution at all, which makes it directly comparable with the warrant coverage on a debt facility. On the numbers here, 1,400,000 of total cost against 1,000,000 advanced is a large price to avoid roughly 1 percent of the company." ] }, { "id": "spv-fee-stacking", "term": "SPV fee stacking - what two layers cost", "definition": "A special purpose vehicle charges a management fee and a carried interest on top of whatever the underlying investment returns. Stack two vehicles and both layers charge on the same gross return, which converts a strong gross multiple into an ordinary net one.", "formula": "Net multiple = [invested * gross - carry * (invested * gross - commitment)] / commitment, where invested = commitment * (1 - total management fee). Two layers apply the same transformation twice", "fields": [ { "label": "Assumptions", "value": "A commitment of 100,000, a 2 percent annual management fee for five years drawn from the commitment, 20 percent carried interest over the commitment, and a 3.00x gross return on invested capital. All four are stated inputs", "mono": true }, { "label": "One layer", "value": "Fees 10,000, invested 90,000, gross return 270,000, carry 34,000, net 236,000 = 2.36x", "mono": true }, { "label": "Two layers, capital deployed", "value": "Feeder invests 90,000 into the vehicle, which invests 81,000 into the company", "mono": true }, { "label": "Two layers, distributions", "value": "Gross 243,000, vehicle carry 30,600, back to the feeder 212,400, feeder carry 22,480, net 189,920 = 1.90x", "mono": true }, { "label": "Cost of the second layer", "value": "2.36x becomes 1.90x - a loss of 0.46x on an unchanged 3.00x gross", "mono": true }, { "label": "Break-even gross", "value": "The gross multiple needed for a two-layer structure to return the one-layer net is materially higher, and it rises with the fee period", "mono": true } ], "notes": [ "Two layers cost about half a turn of multiple here on a 3.00x gross. On a 1.5x gross the same structure can return less than the commitment, because the fees are drawn whatever happens and the carry is only charged when there are gains.", "The management fee is the term that does the damage in a poor outcome and the carry is the term that does it in a good one. Both are quoted as small percentages and both are charged twice.", "Ask which layer's carry is charged over what. Carry over the commitment, over invested capital, or over a preferred return are three different numbers, and the difference is largest exactly where fees have reduced invested capital below the commitment.", "None of the figures above is a market observation; they are the arithmetic of one stated fee structure. The purpose is to show that the transformation is mechanical and can be computed before committing rather than reconciled afterwards." ] }, { "id": "structured-round-versus-a-lower-price", "term": "The clean price that is equivalent to a structured round", "definition": "A structured round at a high headline valuation and a clean round at a low one can leave the common in exactly the same position. Solving for the equivalent clean price turns a structural argument into a valuation argument, which is the argument founders are equipped to have.", "formula": "Find PRE such that common's proceeds under a clean 1x round of R at PRE equal its proceeds under the structured round at a stated E. Solve numerically: lowering PRE raises the investor's share count and lowers common's ownership", "fields": [ { "label": "The structured proposal", "value": "30,000,000 at a 120,000,000 pre-money with a 2x non-participating preference. At E = 150,000,000 common and pool receive 42,000,000", "mono": true }, { "label": "The equivalent clean round", "value": "30,000,000 at a pre-money of 43,750,000, 1x non-participating", "mono": true }, { "label": "Its arithmetic", "value": "Price 2.2312, Series C shares 13,445,379, fully diluted 33,053,223, common and pool 28.475 percent", "mono": true }, { "label": "Common and pool at E = 150,000,000", "value": "42,000,032, against 42,000,000 under the structure", "mono": true }, { "label": "What that means", "value": "At this exit value the 2x preference is worth the same to common as a 63.542 percent reduction in the pre-money valuation", "mono": true }, { "label": "Caveat", "value": "The equivalence holds at one exit value. At a different E the equivalent clean price is different, because a preference and a share count scale differently", "mono": true } ], "notes": [ "This is the single most useful calculation available in a structured negotiation, and it is almost never done. It reframes 'we need a 2x' as 'we are offering 43,750,000 pre-money, not 120,000,000' at the exit value both sides say they are underwriting.", "The equivalence is exit-dependent by construction, which is a feature: it forces the exit value into the conversation. Two parties who cannot agree on the equivalent clean price have not agreed on the exit range, and that is the disagreement worth having.", "A clean round at a lower price is better for common at high exit values and worse at low ones, because a share count participates without limit while a preference does not. Solve at the top and bottom of the credible range, not just the middle.", "The clean round is also better for everything downstream: a lower preferred price lowers the option strike, a smaller preference lowers the common break-even, and a simpler charter is cheaper to amend at the next round." ] } ] }, { "slug": "exits", "title": "Exit mechanics", "lede": "From headline price to cash in a holder's account, with every deduction in between.", "intro": "The headline price in an acquisition is not the number the waterfall runs on. Net debt, a working capital adjustment, transaction expenses, an escrow and an earnout each stand between the announced consideration and the amount distributed, and each of them lands unevenly across the cap table because the waterfall is not linear. This section works one transaction end to end on the share ledger used throughout this corpus: four series holding 9.600, 16.000, 16.000 and 20.000 percent, common and the option pool holding 38.400 percent, and an aggregate 1x preference of 60,000,000.", "reviewed": "2026-08-27", "tables": [ { "title": "The consideration bridge, from enterprise value to a distribution", "intro": "One transaction, all deductions stated as inputs chosen for legibility. Each line is subtracted in turn; the waterfall is then run on the closing distribution and again on cumulative proceeds as the escrow is released. Every row of the waterfall sums to the amount available.", "columns": [ "Step", "Amount", "Running total" ], "rows": [ [ "Enterprise value agreed with the buyer", "100,000,000", "100,000,000" ], [ "Less net debt repaid at closing", "-6,000,000", "94,000,000" ], [ "Less working capital shortfall against the target", "-1,800,000", "92,200,000" ], [ "Less transaction expenses", "-3,000,000", "89,200,000" ], [ "Equity consideration available to the cap table", "-", "89,200,000" ], [ "Less escrow holdback at 10 percent, released later", "-8,920,000", "80,280,000" ], [ "Distributed at closing", "-", "80,280,000" ] ] }, { "title": "Escrow and earnout allocated on cumulative proceeds", "intro": "The same transaction. Each deferred payment is allocated as the difference between the waterfall at cumulative proceeds including it and the waterfall at cumulative proceeds without it. Series B and Series C are at their preference amounts throughout, so they receive none of any deferred payment.", "columns": [ "Distribution", "Seed", "Series A", "Series B", "Series C", "Common and pool" ], "rows": [ [ "Closing distribution of 80,280,000", "4,456,000", "8,000,000", "20,000,000", "30,000,000", "17,824,000" ], [ "Cumulative on full escrow release, 89,200,000", "5,880,000", "9,800,000", "20,000,000", "30,000,000", "23,520,000" ], [ "Cumulative if a 3,000,000 claim is allowed, 86,200,000", "5,430,000", "9,050,000", "20,000,000", "30,000,000", "21,720,000" ], [ "Cumulative with a 20,000,000 earnout achieved, 109,200,000", "8,880,000", "14,800,000", "20,000,000", "30,000,000", "35,520,000" ], [ "Increment from a full escrow release", "1,424,000", "1,800,000", "0", "0", "5,696,000" ], [ "Increment from the escrow net of the claim", "974,000", "1,050,000", "0", "0", "3,896,000" ], [ "Increment from the earnout", "3,000,000", "5,000,000", "0", "0", "12,000,000" ] ] }, { "title": "Asset sale against stock sale on the same 100,000,000", "intro": "A C corporation with a 10,000,000 tax basis in its assets. In an asset sale the corporation recognises the gain and pays entity-level tax before distributing; in a stock sale the consideration reaches the holders directly. A 21 percent rate is used, which is the stated federal corporate rate; state tax, asset character and the seller's attributes all move the real figure. Holder-level tax is outside the table.", "columns": [ "Quantity", "Asset sale", "Stock sale" ], "rows": [ [ "Headline consideration", "100,000,000", "100,000,000" ], [ "Corporate-level gain", "90,000,000", "None" ], [ "Corporate tax at 21 percent", "18,900,000", "0" ], [ "Distributable to the cap table", "81,100,000", "100,000,000" ], [ "Seed", "4,620,000", "7,500,000" ], [ "Series A", "8,000,000", "12,500,000" ], [ "Series B", "20,000,000", "20,000,000" ], [ "Series C", "30,000,000", "30,000,000" ], [ "Common and pool", "18,480,000", "30,000,000" ], [ "Common's share of the corporate tax", "11,520,000 of 18,900,000, or 60.952 percent", "-" ] ] } ], "entries": [ { "id": "consideration-bridge", "term": "The bridge from headline price to distributed consideration", "definition": "The waterfall runs on equity consideration, not on the price in the announcement. Four deductions stand between them, and because the waterfall has kinks at each conversion flip point the deductions do not reduce every holder proportionally.", "formula": "Equity consideration = enterprise value - net debt - working capital adjustment - transaction expenses. Closing distribution = equity consideration - escrow. Run the waterfall on each figure separately", "fields": [ { "label": "The bridge", "value": "100,000,000 less 6,000,000 of net debt, 1,800,000 of working capital shortfall and 3,000,000 of expenses = 89,200,000", "mono": true }, { "label": "Escrow", "value": "8,920,000 held back, so 80,280,000 is distributed at closing", "mono": true }, { "label": "Closing waterfall", "value": "Seed 4,456,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 17,824,000", "mono": true }, { "label": "If the whole 100,000,000 had reached the cap table", "value": "Seed 7,500,000, Series A 12,500,000, common and pool 30,000,000", "mono": true }, { "label": "Cost to common of the bridge", "value": "12,176,000 of the 19,720,000 of deductions, or 61.744 percent", "mono": true }, { "label": "Common's ownership", "value": "38.400 percent - so common absorbs a substantially higher share of every deduction than it owns of the company", "mono": true } ], "notes": [ "Common bears more than its ownership share of every deduction above the aggregate preference, because the deductions come out of the residual and the residual is shared by a smaller group than the whole cap table. The multiplier is common's ownership divided by the residual group's total.", "The working capital adjustment and the expenses are the two lines a seller can still influence at signing, and they are the two that are usually delegated to the finance team. Both are worth the founders' attention for the same reason the preference is.", "Net debt reduces the consideration and does not enter the preference, so a facility drawn to extend runway is repaid ahead of the entire preferred stack. That is the same arithmetic as the venture debt comparison, seen from the other end.", "Ask for the bridge as a schedule with each line named, alongside the allocation schedule. Two documents, and between them they answer every question about who receives what." ] }, { "id": "escrow-release-schedule", "term": "An escrow release schedule, with and without a claim", "definition": "An escrow is a deferred part of the consideration, released on a schedule and reduced by any allowed claim. Because it is allocated on cumulative proceeds, a claim does not reduce every holder in proportion - it reduces the residual claimants first and most.", "formula": "Release to holder i = allocation at (closing + release) - allocation at closing. A claim of K reduces the release to (closing + escrow - K) and the increments are recomputed", "fields": [ { "label": "Setup", "value": "8,920,000 escrow on a closing distribution of 80,280,000, releasable after a stated period", "mono": true }, { "label": "Full release", "value": "Seed 1,424,000, Series A 1,800,000, Series B 0, Series C 0, common and pool 5,696,000", "mono": true }, { "label": "Release after a 3,000,000 claim", "value": "Seed 974,000, Series A 1,050,000, common and pool 3,896,000", "mono": true }, { "label": "Common's share of the claim", "value": "1,800,000 of 3,000,000 = 60.000 percent", "mono": true }, { "label": "Why Series B and C receive nothing", "value": "Both are at their preference amounts at every figure between 80,280,000 and 109,200,000, so no deferred payment changes their allocation", "mono": true }, { "label": "Time value", "value": "The release is nominal, so a holder receiving 5,696,000 eighteen months later has received less than that in present value, with no adjustment in the documents", "mono": true } ], "notes": [ "The holders with the most to lose from a claim are the ones with no ability to defend it, because the residual claimants are the common and the converted series while the claim is usually negotiated by the buyer with a stockholder representative appointed by the preferred.", "That misalignment is the reason to read the indemnification provisions as an economic term. A cap set at the escrow amount, several rather than joint liability, and a materiality scrape are each worth a specific number to common.", "Representation and warranty insurance changes the arithmetic by moving the claim off the escrow and onto a policy, at a premium paid out of the consideration. The premium is a certain deduction against an uncertain one, and it is a computable trade.", "Present value is never adjusted for and it should at least be quantified. Eighteen months of deferral on a tenth of the consideration is a real cost borne almost entirely by the residual claimants." ] }, { "id": "earnout-across-the-stack", "term": "An earnout allocated across a preferred stack", "definition": "An earnout is contingent additional consideration. Allocated on cumulative proceeds it behaves as a second tranche of the same waterfall, which means it goes almost entirely to the residual claimants - the holders who have the least influence over whether the earnout is achieved.", "formula": "Earnout allocation to holder i = allocation at (closing consideration + earnout) - allocation at closing consideration", "fields": [ { "label": "Setup", "value": "A 20,000,000 earnout on top of the 89,200,000 of equity consideration, payable if stated targets are met", "mono": true }, { "label": "Allocation if achieved", "value": "Seed 3,000,000, Series A 5,000,000, Series B 0, Series C 0, common and pool 12,000,000", "mono": true }, { "label": "Check", "value": "20,000,000 allocated", "mono": true }, { "label": "Common's share", "value": "60.000 percent of the earnout, against a 38.400 percent ownership", "mono": true }, { "label": "Who controls achievement", "value": "The buyer, after closing, through the operating decisions of the acquired business", "mono": true }, { "label": "If not achieved", "value": "Every holder receives the closing allocation and nothing more; the entire variance sits with the residual claimants", "mono": true } ], "notes": [ "An earnout transfers variance to the residual claimants and control to the buyer. That is the structural objection to it, and it is stronger than the usual objection about measurement disputes.", "Because the preferred are already whole at these figures, the preferred holders negotiating the deal have no economic interest in the earnout at all. The parties who care are the founders and the employees, and they are usually not at the table for that clause.", "The protections that matter are operational rather than financial: a covenant to run the business consistently with past practice, a defined measurement methodology, an acceleration on a subsequent sale or on termination of key employees, and an offset restriction so the earnout cannot be used to satisfy indemnity claims.", "Model the transaction with the earnout at zero. If the deal is only acceptable with it achieved, the earnout is being used to bridge a valuation gap rather than to price genuine uncertainty." ] }, { "id": "allocation-schedule", "term": "The allocation schedule is the document that decides who gets paid", "definition": "The charter states the waterfall and the allocation schedule applies it. It is delivered late, prepared by the company's counsel, certified by an officer, and relied on by the buyer's paying agent. It is also the only document in the transaction that states each holder's number.", "formula": "For each holder: shares held, class and series, election where applicable, gross allocation, escrow portion, withholding, and net payment at closing. The gross allocations must sum to the consideration available", "fields": [ { "label": "What it must reconcile to", "value": "The stock ledger, the option ledger, the charter waterfall, the carve-out plan if any, and the consideration bridge", "mono": false }, { "label": "Where errors live", "value": "Conversion elections computed series by series instead of jointly; the carve-out omitted because it is not on the cap table; unvested and terminated options treated inconsistently; the escrow allocated pro rata to closing rather than cumulatively", "mono": false }, { "label": "The arithmetic check", "value": "Every column sums: gross allocations equal the consideration available, escrow portions equal the escrow, and each series' figure equals what the charter formula produces at that consideration", "mono": true }, { "label": "Worked check on this transaction", "value": "Seed 4,456,000 + Series A 8,000,000 + Series B 20,000,000 + Series C 30,000,000 + common and pool 17,824,000 = 80,280,000", "mono": true }, { "label": "Rely on it at your peril", "value": "The schedule is usually accompanied by an indemnity from the company to the buyer and the paying agent for errors in it, which means the risk of a mistake sits with the selling holders", "mono": false } ], "notes": [ "Ask for the schedule in draft, with the formula for each line shown rather than only the result. Every holder is entitled to check the arithmetic that produces its own number, and the arithmetic is not complicated - it is the four expressions in the liquidation section of this corpus.", "The most common material error is an election computed in charter order rather than solved jointly, because that is what a spreadsheet built one row at a time produces. It is wrong precisely in the exit range where the answer matters.", "The second most common is the carve-out, which is documented as a bonus plan and never appears on the cap table. A schedule built from the stock ledger alone omits the most senior claim in the structure.", "Options require their own treatment: vested in the money, vested out of the money, unvested, and terminated but within the exercise window are four different populations with four different outcomes, and the plan and the merger agreement have to agree on each." ], "source": "Delaware General Corporation Law section 251 and the NVCA model Amended and Restated Certificate of Incorporation (liquidation provisions)" }, { "id": "asset-sale-versus-stock-sale", "term": "An asset sale costs the common most", "definition": "In an asset sale by a C corporation the entity recognises the gain and pays tax before anything is distributed, so the same headline price produces a smaller distribution. Because the shortfall comes out of the residual, the common bears far more of it than its ownership share.", "formula": "Distributable = consideration - corporate tax on the gain. Then run the waterfall. Common's share of the tax = its proceeds at the full consideration less its proceeds at the reduced figure", "fields": [ { "label": "Setup", "value": "100,000,000 headline, 10,000,000 tax basis, a 21 percent corporate rate", "mono": true }, { "label": "Corporate tax", "value": "0.21 * 90,000,000 = 18,900,000, leaving 81,100,000 distributable", "mono": true }, { "label": "Stock sale", "value": "Seed 7,500,000, Series A 12,500,000, Series B 20,000,000, Series C 30,000,000, common and pool 30,000,000", "mono": true }, { "label": "Asset sale", "value": "Seed 4,620,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 18,480,000", "mono": true }, { "label": "Common's share of the tax", "value": "11,520,000 = 60.952 percent of it, on a 38.400 percent ownership", "mono": true }, { "label": "Series B and C", "value": "Unchanged - both are at their preference amounts under either structure", "mono": true } ], "notes": [ "The buyer's preference for an asset purchase is about liabilities and tax basis, and the seller's preference for a stock deal is about the entity-level tax. The gap is negotiated as price, so the form of the transaction is an economic term.", "The holders who bear the cost of an asset sale are not the holders who negotiate the form. The preferred are whole either way at these figures, and the founders and employees absorb almost two-thirds of the tax.", "Corporate attributes change the arithmetic materially. Loss carryforwards, the character of the assets, and state tax all move the figure, and a company with substantial losses may face little entity-level tax at all - which is worth establishing before conceding the form.", "The rate used here is the stated federal corporate rate and the example is arithmetic, not tax advice. The structural conclusion - that an entity-level tax is borne mostly by the residual claimants - does not depend on the rate." ], "source": "Delaware General Corporation Law section 271 (sale of substantially all assets)" }, { "id": "qsbs-on-exit", "term": "Applying the section 1202 exclusion at the exit", "definition": "The exclusion is computed per holder and per issuer, against a ceiling. On a founder position with near-zero basis the ceiling is the statutory dollar amount, and it is a small fraction of the gain on a successful exit.", "formula": "Excluded gain = tier fraction * min(gain, max(K, 10 * aggregate adjusted basis)). Taxable gain = total gain less excluded gain", "fields": [ { "label": "Setup", "value": "A founder holding 8,000,000 shares with an aggregate basis of 8,000, at an exit of 250,000,000 where all preferred converts and the per-share value is 10.2000", "mono": true }, { "label": "Proceeds and gain", "value": "81,600,000 of proceeds, 81,592,000 of gain", "mono": true }, { "label": "Ceiling", "value": "max(K, 10 * 8,000). With K taken as 15,000,000 the ceiling is 15,000,000", "mono": true }, { "label": "Excluded and taxable", "value": "15,000,000 excluded at a 100 percent tier, which is 18.384 percent of the gain. 66,592,000 remains taxable", "mono": true }, { "label": "Two founders holding the same position between them", "value": "Each has its own ceiling, so 30,000,000 of gain is excluded rather than 15,000,000", "mono": true }, { "label": "A holder that paid real money", "value": "An investor with a 20,000,000 basis has a ten-times-basis ceiling of 200,000,000, so the alternative test is the operative one for it and not for the founder", "mono": true } ], "notes": [ "The exclusion is per holder, which makes it additive across founders, family trusts and any other qualifying holder of separately issued stock. That is why the planning around it happens years before the exit and involves who holds the stock rather than what the company does.", "The ten-times-basis alternative is where the design shows: it is generous to holders who paid cash and irrelevant to holders who paid par. Founders are systematically in the second group.", "Every parameter above - the tier percentages, the holding periods, the per-issuer ceiling and the gross-asset test - is statutory and was amended for stock acquired after 4 July 2025. Read the current statute; this entry states the mechanism and uses one set of figures to show the arithmetic.", "The tests are applied at issuance and at disposition, and both can fail for reasons that have nothing to do with the exit: an entity conversion, a redemption within a lookback period, or gross assets crossing the ceiling before the stock was issued." ], "source": "IRC section 1202" }, { "id": "tender-offer-pricing-versus-last-round", "term": "Pricing a tender relative to the last round", "definition": "A secondary or company-run tender is a purchase of common stock, and the last round set a price for preferred stock. The two are different securities on the same cap table, so a tender priced at a discount to the round is not a markdown - it is a different instrument at a different price.", "formula": "Discount to the last preferred price = 1 - tender price / preferred price. The economically comparable figure is the tender price against the common fair market value, not against the preferred price", "fields": [ { "label": "Setup", "value": "The Series B price is 5.1000. A tender is offered for common at 4.5000", "mono": true }, { "label": "Discount to the preferred price", "value": "11.765 percent", "mono": true }, { "label": "Against the assumed common fair market value of 2.0400", "value": "The tender is at 2.21 times the appraised common value", "mono": true }, { "label": "Consequence for option strikes", "value": "A large arm's length purchase of common at 4.5000 is evidence of the fair market value of common and will be weighed in the next 409A analysis", "mono": true }, { "label": "Consequence for the seller", "value": "Proceeds are certain and the position is given up before any preference is cleared, which is worth most when the aggregate preference is large relative to plausible exits", "mono": true }, { "label": "Consequence for the buyer", "value": "Common with no preference, no protective provisions and no anti-dilution, at a price below the last preferred round", "mono": true } ], "notes": [ "The gap between the tender price and the preferred price is doing real work in the appraisal. A tender priced at or above the preferred price is very hard to reconcile with a common value well below it, which is one reason company-run tenders are priced below the round.", "A tender is the cheapest available mechanism for raising the strike on every option granted afterwards, and that consequence is rarely stated to the employees being offered liquidity in it.", "For a seller, the comparison that matters is the tender price against the expected value of the same shares through the waterfall, not against the last round price. Below the aggregate preference the shares are worth nothing, so certainty is worth a great deal.", "Tenders are purchase offers to a class of holders and carry process requirements that a single negotiated transfer does not. The pricing is the easy part." ] }, { "id": "working-capital-adjustment", "term": "The working capital adjustment, and why it is a price term", "definition": "The purchase price is agreed on the assumption that the business is delivered with a normal level of working capital. The adjustment settles the difference between that target and the actual figure at closing, and it reduces the equity consideration dollar for dollar.", "formula": "Adjustment = actual closing working capital - target working capital. A negative figure reduces the consideration; a positive figure increases it. The adjustment flows into the equity consideration before the waterfall runs", "fields": [ { "label": "Setup", "value": "A target of 5,000,000 against an actual closing figure of 3,200,000", "mono": true }, { "label": "Adjustment", "value": "3,200,000 - 5,000,000 = -1,800,000, reducing the equity consideration to 89,200,000", "mono": true }, { "label": "Effect on common", "value": "Common and pool receive 17,824,000 rather than the 19,248,000 they would have received without the shortfall - a cost of 1,424,000 from an 1,800,000 adjustment", "mono": true }, { "label": "Common's share", "value": "79.111 percent of the adjustment", "mono": true }, { "label": "Where the target comes from", "value": "A negotiated average of historical months, adjusted for seasonality and for items the parties agree to exclude. It is a negotiation, not a measurement", "mono": true }, { "label": "The estimate and the true-up", "value": "A closing estimate is used to fund the payment and a post-closing statement settles the difference, usually against a separate holdback", "mono": true } ], "notes": [ "The definition of working capital for this purpose is bespoke in every deal. Deferred revenue, accrued bonuses, capitalised software costs and transaction expenses are each argued in or out, and each decision moves the target.", "Because the target is negotiated and the actual is measured on the negotiated definition, the same balance sheet can produce very different adjustments. The definition is the term; the arithmetic is trivial once it is fixed.", "A company that manages cash tightly into a closing will collect receivables and delay payables, which raises the closing figure and is exactly what the adjustment mechanism is designed to neutralise. A well-drafted definition anticipates it.", "Common bears more than its ownership share of the adjustment, for the same reason it bears more of every other deduction: the reduction comes out of the residual." ] }, { "id": "appraisal-rights", "term": "Appraisal rights, and the risk they create for the consideration", "definition": "A stockholder that dissents from a merger and follows the statutory procedure can ask a court to determine the fair value of its shares and be paid that instead of the merger consideration. The exposure is a claim against the surviving company, which in a deal with an escrow and an indemnity becomes a claim against the selling holders.", "formula": "Exposure = dissenting shares * (determined fair value - merger consideration per share) plus statutory interest. Fair value is determined without reference to the preference where the court concludes the charter does not bind it", "fields": [ { "label": "Why it matters to a venture cap table", "value": "Fair value is a going-concern valuation of the shares, and a court is not obliged to accept a contractual waterfall as the measure of value for the dissenting class", "mono": true }, { "label": "Worked exposure", "value": "1,000,000 dissenting common shares against a merger consideration of 1.8938 a share. A determination of 3.0000 a share creates an exposure of 1,106,200 plus interest", "mono": true }, { "label": "Who bears it", "value": "Usually the selling holders, through the indemnity and the escrow, which means the residual claimants absorb most of it", "mono": true }, { "label": "Why it is usually avoided", "value": "A drag-along obliges holders to vote in favour, and a holder that votes in favour loses appraisal. Written consent from a large majority plus a drag removes most of the exposure", "mono": true }, { "label": "The procedural requirements", "value": "Strict: no vote in favour, a written demand before the vote, continuous holding, and a timely petition. Most attempts fail on procedure rather than on value", "mono": false }, { "label": "Closing condition", "value": "Buyers commonly require that dissenting shares be below a stated percentage as a condition to closing", "mono": true } ], "notes": [ "The drag-along is the practical answer to appraisal risk, which is why the drag's trigger is a closing mechanic and not only a governance term. A drag that cannot be exercised leaves the appraisal exposure open.", "Appraisal is one of the few routes by which a common holder can challenge a waterfall outcome, which is why it is worth understanding from both sides: as a risk to the consideration and as the residual claimants' only leverage.", "The exposure is asymmetric in a low-price deal. Where the consideration to common is near zero because the preference absorbs everything, a fair value determination has a great deal of room above it.", "Nothing here is legal advice and the procedural requirements are unforgiving. The arithmetic point is that an appraisal claim is a contingent liability of the consideration and belongs in the bridge alongside the escrow." ], "source": "Delaware General Corporation Law section 262 (appraisal rights)" } ] } ] }