How a cap table shows who controls your company

A cap table tells you who owns a company, what kind of equity each person holds, and what everyone could own after outstanding rights convert or get exercised. It also exposes decisions that founders often make one document at a time: a SAFE here, an advisor grant there, a larger hiring pool in the term sheet. Each decision can look small while the combined result changes both ownership and control.

The dangerous version is not a cap table with an obvious arithmetic error. It is a polished table built on the wrong denominator, missing a convertible instrument, or disconnected from signed approvals. A founder can quote a technically correct percentage and still misunderstand what she will own after the next financing. Treat the cap table as a live model tied to legal records, and make it show the transaction before you approve the transaction.

A cap table records claims on the company

A capitalization table is a structured record of the equity claims on a company. It names holders, identifies their securities, records share or unit counts, and calculates ownership percentages on a stated basis. For a US corporation, founders usually hold common stock, venture investors usually buy preferred stock, employees may receive options or restricted stock, and early investors may hold SAFEs or convertible notes that become shares later.

The SEC makes a useful distinction in its Common Startup Securities guide: stock is an ownership interest, while SAFEs and convertible notes are separate securities with conversion terms. That distinction matters. A SAFE holder may not be a stockholder today, but ignoring the SAFE until a priced round hides equity that the company has already promised to issue.

A practical cap table should let a reader trace each number rather than merely display a percentage. At minimum, it should include:

• The holder's legal name and stakeholder category
• The security type, class or series, and certificate or grant identifier
• Shares issued, options granted, warrants, and the unallocated option reserve
• Issue or grant date, price per share, vesting status, and exercise price where relevant
• Conversion terms for each SAFE or note, plus a reference to the approval and signed document

The table should also separate current capitalization from a pro forma view. The current view reports what legally exists now. The pro forma view applies an assumed financing, conversion, option-pool increase, or other event and reports what would exist afterward. Label the assumptions and the effective date. A percentage without those two labels invites a bad comparison.

Do not mistake the cap table for the stock ledger. Delaware law gives the stock ledger a specific role: it records stockholders of record, their addresses and registered shares, plus stock issuances and transfers. Section 219 of the Delaware General Corporation Law says that ledger is the evidence used to determine who may inspect the stockholder list or vote at a stockholder meeting. Your planning spreadsheet may be more readable, but it does not replace the ledger, board consents, stock purchase agreements, option grant notices, or financing documents.

The denominator decides what a percentage means

Every ownership percentage is a fraction, and founders lose track of equity when the numerator looks precise but the denominator stays unnamed. Ten thousand options might be 1%, 0.5%, or far less. The answer depends on which securities the calculation includes and when you take the measurement.

Four share counts often appear in the same conversation:

• Authorized shares are the maximum shares the charter currently permits the corporation to issue.
• Issued shares are shares the corporation has actually issued.
• Outstanding shares are issued shares that remain held and have not returned to the corporation.
• Fully diluted shares usually add outstanding shares to options, warrants, the unallocated option pool, and shares expected from convertible securities under stated assumptions.

Authorized shares are capacity, not ownership. If the charter authorizes 20 million shares and the company has issued 8 million, a founder with 4 million shares does not own 20%. She owns 50% of the issued and outstanding shares before accounting for other claims. The 12 million unissued authorized shares do not sit in an invisible owner's account.

The phrase "fully diluted" also needs a definition. Some tables include granted options and the remaining pool but show SAFEs separately because the conversion price is not known. Others model each SAFE against an assumed financing. Either approach can work if the table states the basis. Mixing them cannot. When an investor asks for a percentage on a fully diluted, post-money basis, confirm whether that denominator includes the new financing shares, converting SAFEs and notes, warrants, granted equity awards, and the entire post-closing pool.

This is the core calculation:

ownership percentage = holder's as-converted shares / total shares on the same basis

"As-converted" means you calculate a preferred holder's position as though preferred stock converted into common stock according to its terms. That convention helps compare economic ownership, but it does not erase the preferred stock's contractual rights. A fully diluted percentage is not a prediction of sale proceeds, either. Liquidation preferences, participation rights, accrued dividends, debt, transaction costs, and the sale price can produce a different payout.

Whenever someone sends a percentage, ask for the numerator, denominator, date, and conversion assumptions. If the answer is "the software calculated it," the number still needs those inputs.

Dilution changes the fraction, not your certificate

Dilution usually happens because the company adds shares or rights to shares, so an existing holder owns a smaller fraction of a larger total. The company does not normally take shares out of a founder's account during a financing. Her share count can stay fixed while her percentage falls.

Consider two founders who own 6,000,000 and 4,000,000 common shares. Before any pool or financing, they own 60% and 40%. They then reserve a pool equal to 10% of the fully diluted pre-financing capitalization and raise $2 million at an $8 million pre-money valuation. Ignoring convertibles, the math looks like this:

Creating a 10% pool required 1,111,111 reserved shares, not 1,000,000. If the company merely added 1,000,000 to the original 10,000,000, the pool would equal about 9.09% of the new total. The equation is new pool shares / (existing shares + new pool shares) = target percentage.

At the seed round, post-money valuation is $10 million: the $8 million pre-money valuation plus $2 million of new cash. The new investor therefore owns 20% immediately after closing, before any later issuance. Every pre-financing holder and the pool gets multiplied by 80%. The next round in the table assumes another investor group buys 20% with no pool increase or convertible conversion, so every existing percentage gets multiplied by 80% again.

The founders' combined stake moved from 100% to 72%, then 57.6%, without either founder transferring a share. That is ordinary financing dilution. It can be a sensible trade if the capital raises the company's value enough. The mistake is evaluating a round only by its valuation headline and discovering the cumulative ownership result later.

Run this table with your actual share counts after every proposed SAFE, note, option-pool change, priced round, warrant, and secondary sale. Model all signed but unconverted instruments together. A single transaction rarely explains a founder's surprise; the stack does.

The option pool comes out of a negotiated pocket

An option pool is a reserve of shares that the company may grant to employees, advisors, or other service providers under an approved equity plan. No employee owns the unallocated portion. Yet the full reserve commonly sits in the fully diluted denominator, which means creating or enlarging it dilutes existing holders before the company grants the options.

The phrase "we need a 10% pool" is incomplete. Does 10% mean the total pool, or the amount still available after subtracting grants already made? Does it mean 10% before the financing or after it? Who absorbs the increase?

Use a clean example. A company has 10,000,000 founder shares and plans to raise $2 million at an $8 million pre-money valuation. The investor would own 20% after the round. If the term sheet also requires an unallocated pool equal to 10% after closing and puts that increase inside the pre-money capitalization, the company must reserve about 1,428,571 shares and issue about 2,857,143 investor shares. The post-closing ownership is 70% founders, 10% pool, and 20% investor.

If the same company closes the investment first and creates a 10% post-closing pool afterward, the investor shares the pool dilution. Founders end near 72%, the investor near 18%, and the pool at 10%. The two percentage points did not disappear; the sequence assigned them to a different holder. Cooley calls attention to this negotiation in its option-pool guidance, and practitioners often call the pre-money version the option-pool shuffle.

Investors ask for a pre-money top-up because they want enough equity available for hiring without reducing their negotiated post-closing stake. That request is understandable. Accepting an arbitrary pool size is not. Build a hiring plan through the next expected financing, estimate grants by role and level, subtract the unused reserve already available, and propose the pool that the plan supports.

One wording trap causes needless dilution: a term sheet may require the pool to equal a target percentage "available" after closing. If the company already granted 4% and needs 8% unallocated, the fully diluted plan may total roughly 12%, subject to the document's definitions. Reading "8% pool" when the term requires 8% remaining can materially understate the top-up.

Board approval of the plan does not grant options to a recruit. The board generally must approve individual grants as well, following the plan and company documents. An offer letter that mentions an intended grant is not the same record as an approved grant notice. Reconcile promised, approved, granted, vested, exercised, canceled, and available amounts instead of storing all of them in one "options" column.

SAFEs can hide the ownership already sold

A SAFE delays the share count, but it does not delay the economic bargain. Founders often collect several SAFE signature pages, record the cash, and wait for company counsel to calculate conversion at the priced round. By then, they have negotiated against a cap table that omitted part of the deal.

Y Combinator's post-money SAFE primer says its main advantage is that ownership sold is transparent and calculable. Under a simple post-money valuation-cap example where the cap controls, an investor's implied ownership is the investment divided by the post-money cap. A $500,000 SAFE at a $5 million post-money cap points to 10%. Five separate checks do not dilute one another merely because they arrived on different days; add their modeled percentages and then model the priced round dilution that follows.

That shortcut has limits. A SAFE may have a discount, a most favored nation provision, pro rata rights, or another term that changes the result. The priced round valuation can sit below or too close to the cap. Different SAFE forms define company capitalization differently. A convertible note also accrues interest and has a maturity date, so its conversion amount may grow. The signed instrument, not the filename or a row labeled "SAFE," controls the calculation.

Maintain a separate line for every instrument. Record the purchase amount, form version, pre-money or post-money structure, valuation cap, discount, date, pro rata side letter, and any amendment. Then show at least two pro forma cases: conversion at the expected financing terms and conversion at a lower valuation where the cap or discount behaves differently.

The quiet failure looks like this. A founder raises several uncapped or differently capped instruments over 18 months. Each investor conversation refers to "about a few percent," but nobody consolidates the instruments. The priced-round term sheet adds a fresh pool requirement, the notes accrue interest, and the new investor prices the round on a fully diluted pre-money basis. The founder sees the complete conversion spreadsheet after exclusivity or late in legal diligence, when changing the economic deal is harder.

Do not wait for a priced round. After each instrument closes, update a live SAFE and note schedule, model it beside the cap table, and show the cumulative implied ownership range. If you cannot explain the range without opening every PDF again, the record is not ready.

Ownership and control are different questions

The cap table measures economic ownership, but corporate control also depends on voting rights, board composition, class votes, protective provisions, and signed voting agreements. A founder can own most of the common stock and still lack unilateral authority over a financing, sale, charter amendment, or board decision. Another founder can own less than half of the economics and still influence board decisions through agreed appointment rights.

Common and preferred stock may vote together on an as-converted basis for some matters, while the preferred class votes separately on others. A certificate of incorporation may require preferred approval before the company changes authorized shares, issues senior securities, sells the company, or takes other listed actions. The exact list comes from the charter and financing documents. The National Venture Capital Association publishes separate model documents for the certificate of incorporation, voting agreement, investor rights agreement, and purchase agreement because one ownership table cannot carry all those rights.

Board control is its own calculation. Write down every current seat, who elects or designates it, who can remove that director, whether a seat is vacant, and what approvals require a board majority. Then model the proposed round's board arrangement beside the equity model. "Three founder seats" means little if only one founder still works at the company, a voting agreement controls replacement, or a deadlock provision shifts practical power.

Voting control can also move without a financing. A co-founder departure may trigger repurchase rights over unvested shares. A voting proxy may continue. An option exercise can turn a service provider into a stockholder. A charter amendment can create a new class with separate rights. A secondary sale changes the holder of existing shares without bringing new cash into the company and usually does not dilute the total, but transfer restrictions and voting agreements still affect the result.

Do not reduce control to "stay above 50%." Ask two questions before signing: what percentage will I own on an outstanding and fully diluted basis, and which company actions can happen without my board or stockholder vote? Your lawyer should answer the second question from the actual documents. The cap table supplies only part of the evidence.

Small record errors become expensive transaction work

Most broken cap tables start with ordinary administrative shortcuts. Someone copies last quarter's spreadsheet, changes a total, and leaves an old formula in one stakeholder row. An advisor grant appears in an email but never reaches the board consent. A departed employee's unvested options remain outstanding in the model. A repurchase payment clears, but the ledger never records the returned shares. None of these mistakes feels urgent until financing counsel has to prove every line.

The most common denominator error is mixing outstanding and fully diluted percentages in the same column. Founder stock gets divided by outstanding shares, while an employee option grant gets divided by a total that includes the pool. The column still adds to 100% because someone plugs the remainder. It cannot answer any coherent ownership question.

Another error comes from treating the cap table as the approval. Entering a grant does not authorize it. A board consent without a matching grant notice does not tell the recipient the full terms. A signed stock purchase agreement without proof that the company received the purchase price leaves an incomplete issuance file. A clean row should point to the corporate action and signed instrument that created it.

Version drift does similar damage. The CEO models a round in one file, the lawyer updates a closing model, the finance lead tracks exercises elsewhere, and the board deck uses a screenshot. Several files may have been correct on different dates. None can claim to be current without a clear owner, an effective date, and a reconciliation process.

Control errors tend to hide outside the spreadsheet. A founder watches her common-stock percentage but does not model an investor board seat, a preferred veto, or a voting agreement amendment. She has tracked dilution and missed governance. Put a governance summary beside each financing scenario, even if the summary only points counsel and the board to the controlling clauses.

Bad data also reaches employees. Telling a candidate that 50,000 options equal "about 0.5%" without the fully diluted share count, measurement date, and future dilution caveat makes the number sound more stable than it is. Give the share count and the denominator used. Do not promise future value or a fixed percentage.

The cap table must reconcile to legal records

A reliable cap table can be rebuilt from the company's governing documents and transaction file. For a Delaware corporation, compare it with the charter's authorized shares, the stock ledger's issued and transferred shares, board and stockholder approvals, equity-plan records, signed purchase and grant documents, and the complete convertible schedule.

Run a cap-table close on a regular cadence and before every financing, major grant cycle, tender offer, or acquisition discussion:

1. Freeze an effective date and export the current outstanding and fully diluted views. Do not update assumptions during the reconciliation.
2. Tie every issued share, grant, cancellation, exercise, repurchase, transfer, SAFE, note, and warrant to its signed document and approval.
3. Recalculate totals independently, confirm that each class stays within charter authorization, and explain every difference from the prior close.
4. Have the responsible executive and company counsel resolve exceptions, then lock the dated version and update the live system.

The output should include an exception log. Typical entries are mundane: missing signature, grant date mismatch, stale address, exercise payment not matched, canceled option still counted, or SAFE amendment absent from the model. Give each exception an owner and a resolution date. Silent corrections make future diligence harder because nobody can explain why the table changed.

Delaware General Corporation Law Section 224 allows corporate records, including a stock ledger, to be kept through electronic systems if the company can convert them into clearly legible paper and the ledger meets statutory functions. The practical lesson is narrower than "use software." The company must be able to produce intelligible records and trace issuances and transfers. Software cannot repair an issuance that lacked required authorization or a document nobody signed.

Early companies can maintain a sound cap table in a carefully controlled spreadsheet. That stops working when transaction volume, instrument variety, or access by several people defeats version control. Move to a dedicated equity system before the spreadsheet becomes evidence you no longer trust, and keep counsel involved in the migration. Importing a wrong table creates a faster wrong table.

Tax records belong in the same operating rhythm even though the cap table is not a tax return. Option grants often depend on a current fair market value determination. IRS guidance under Section 409A generally treats a stock option with an exercise price at least equal to fair market value at grant, and without another deferral feature, differently from a discounted option. Ask qualified tax and legal advisors to set the process; do not backdate a valuation to make the spreadsheet look consistent.

Model the signed deal before it becomes binding

The best time to discover dilution or a control shift is while the document still says "draft." A financing model should reproduce the term sheet's defined capitalization, not a generic calculator's defaults. Build the pro forma from share counts and instrument terms, then compare it with counsel's model line by line.

For each proposal, keep a compact decision sheet showing the pre-money valuation, new cash, price per share, converting instruments, option-pool top-up, post-closing fully diluted shares, ownership by holder group, board seats, separate class approvals, and any pro rata or secondary component. Add cases for a smaller close, a larger close, and the next expected hiring plan. The goal is not to predict the company perfectly. It is to expose which assumption changes your outcome.

Read percentages back into share counts. If a term sheet promises an investor 20% and an employee pool 10% after closing, solve for the shares required instead of typing 20 and 10 into a presentation. Check whether "10%" means total plan size or unallocated capacity. Apply every SAFE definition exactly. Confirm that the resulting issuances fit within authorized shares and that the required corporate approvals appear in the closing plan.

Then test control separately. Mark decisions the board can approve, decisions requiring a stockholder vote, and decisions requiring a preferred-class consent. Ask counsel how voting agreements, proxies, drag-along provisions, and founder-vesting terms affect your actual position. This is specific legal work, and a generic cap-table article cannot answer it for your company.

Inside Sisters, founders can ask peers who have already worked through a US financing to pressure-test the questions they take to counsel, and use the community's workshops and practical knowledge base to prepare for the conversation. Peer experience is useful preparation, but it does not replace company counsel reviewing the signed documents.

A cap table earns trust when every row has a source, every percentage names its denominator, and every proposed deal has a dated pro forma. If your latest table cannot show the next round, the option-pool increase, and all convertibles in one view, do not sign from it.