When the cap table folder is opened in an investment review, it is entirely possible to find a share register that reconciles to the statutory ledger, an option pool cleanly separated into allocated and unallocated tranches, and vesting schedules tracked month by month; and yet, when the same file is asked for the aggregate nominal value of convertible instruments executed over the preceding two years, the answer tends to arrive from the founder's recollection rather than from any table. This is less a symptom of disorder than a structural consequence. Because SAFE-style instruments issue no shares at signature and therefore open no line in the share register, they fall outside the company's recording architecture by default — appearing on the accounting side as a liability or an equity item, resting on the legal side inside a contracts folder, and belonging nowhere at all on the ownership side. The instrument itself does not create the gap; the gap arises because searching the share records for an agreement that produces no shares is nobody's assigned task.
The practical appeal of SAFE-style instruments — simple agreements for future equity, which decline to price today's participation and defer conversion to the next priced round — rests exactly on this capacity for deferral. The company takes cash without conducting a valuation negotiation; the investor, having absorbed early-stage risk, is protected through a valuation cap or a discount; and both sides push the most expensive and most adversarial conversation to a later date. That choice is rational to the extent that it admits capital quickly and at low transaction cost. The difficulty is that what has been deferred is not merely the negotiation but the arithmetic itself: if the combined effect of cap, discount, MFN clause and conversion threshold is only to be computed at the priced round, then throughout the intervening period the company does not know its own ownership structure.
These instruments accumulate in a way that is diagnostic as a set rather than individually. Fifteen pages of text signed at different moments, with different investors, at different cap levels, each look defensible when read alone; read together, they form an interlocking system. An MFN clause in one instrument reaches backward to capture a lower cap granted in a later one; a conversion threshold in another makes the size of the planned round a function of the instrument rather than of the company's financing plan; and when a post-money instrument and a pre-money instrument convert in the same round, the question of whose ownership absorbs the dilution resolves in two entirely different ways. The reviewing party is looking for exactly these relationships, and when they are absent, what is missing is not a document but a mind that has read the documents against one another.
On the documentation dimension, the most frequently observed finding is that the principal agreements are stored properly while the surrounding layer of undertakings remains scattered. Side letters, supplementary correspondence granting information rights, single-paragraph emails conferring pro rata participation, cap levels revised by verbal understanding after execution — none of these sits in the main directory of the contracts folder, and all of them are legally operative at conversion. The point that matters for institutional memory is narrower than it first appears: these ancillary undertakings typically live inside one individual's mailbox and depend on that individual's recall. When such a document surfaces mid-diligence, it does more than alter the terms of the instrument it modifies; it lowers the confidence coefficient applied to the file as a whole.
What the practice dimension measures is not the execution of the instrument but what happens after execution. When a convertible instrument is signed, a line should open in the fully diluted cap table, conversion assumptions should be modeled against then-current conditions, and that model should be refreshed at every subsequent round. What is observed instead is that the signature is celebrated at closing and the model is not opened again until a priced round approaches. Because this behavior generates no cost in the short term, it appears sustainable; the cost materializes in aggregate, during the most strained week of the round negotiation, at the moment existing holders discover they have diluted several points beyond expectation. At that point what is being negotiated is no longer the new investor's price but the relative priority of the older instruments among themselves.
Measurement is the dimension most companies never build here, largely because the object to be measured is not thought of as a metric. The quantities that warrant tracking are, however, few and explicit: the aggregate nominal value of live instruments, the weighted average valuation cap, the dilution band produced across different round sizes, and the residual position of existing holders under each of those scenarios. The distinction between a company that refreshes these four figures quarterly and one that never refreshes them lies not in the legal quality of the instruments but in management's field of vision over its own capital structure. When a reviewing party asks for that table, the thing being measured is not the dilution percentage; it is how many scenarios the company is able to compute about its own future.
The ownership question produces a sharper result here than expected. Responsibility for convertible instruments is customarily declared to sit with the founder or the CFO, but when the chain is followed, two distinct responsibilities are found consolidated in one person: whoever negotiates the instrument is also whoever models its effect. That consolidation makes it structurally difficult for a concession granted during negotiation to be priced dispassionately afterward, since no one reports the dilutive consequence of a cap level bearing their own signature with a cold eye. The separation is simple and carries no cost: instrument negotiation belongs on one side, maintenance of the fully diluted table on the other, and the second should not report into the first.
The continuity dimension examines whether the knowledge can be detached from the founder. A founder's ability to recite cap levels from memory in a meeting is often presented as a mark of command; at the diligence table it is read in precisely the opposite direction, since single-source knowledge is by definition non-transferable knowledge. The channel into valuation is direct here. Where the information is concentrated in one individual, the acquiring party prices that individual's post-closing tenure as a discrete risk item, and that risk typically returns not as a headline price reduction but as a structural change — broader representations and warranties, a dedicated indemnity heading for cap table accuracy, a pre-closing condition requiring confirmation of the fully diluted table, or an escrow percentage moved up a notch.
Structural remediation requires treating this dispersion as a problem of record and cadence rather than one of document housekeeping. BEIREK's work in this area begins by consolidating every convertible instrument into a single-source instrument register, in which each row carries nominal amount, execution date, cap structure, pre-money or post-money designation, discount rate, presence of MFN language, conversion threshold and any side letter reference as separate fields — and no field is left blank, since anything unknown is marked as unverified and drops into the open items list. Layered over that register is a dilution model running at least three round-size scenarios in parallel, the output of which is not a single figure but an ownership band for existing holders.
The second layer is cadence. The register and the model are refreshed when the calendar arrives, not when a transaction occurs; at each quarter close the fully diluted table is presented to the board as a single page, and that page is produced even in quarters when no new instrument has been signed. While a new instrument is under negotiation, its effect on the existing register is computed as a mandatory pre-signature step, and the decision is taken alongside that computation — dilution discovered after signature is dilution that has ceased to be negotiable. Custody of the record is separated from the person conducting the negotiation, so that whoever grants the concession is not also whoever reports it.
The most visible effect of this discipline surfaces in the review itself. When the first item produced from the cap table folder is the instrument register and the dilution scenarios, the counterparty's question list compresses from several weeks into several days; more consequentially, the character of the questions changes. The reviewing party moves from asking whether other instruments exist to asking how a particular cap level was determined — and the second question builds a thesis rather than eroding a price. The existence of the documents is a hygiene condition here; what distinguishes a file is that the documents have been read against one another and their combined result assembled into a single table.
Convertible instruments disclose two things about a company's capital structure at once: how quickly it can raise money, and how long it waits before computing what that money costs. The first is a statement about capacity, the second about maturity, and it is almost invariably the second that carries weight at the diligence table. A company able to present its own ownership structure under three distinct scenarios, with the founder absent from the room, has produced the cheapest and most persuasive evidence of capital discipline available to it — evidence that stands independently of the legal quality of any individual instrument.