The keystroke that stalled a $145M deal for 19 days
A founder we worked with had a data infrastructure company under LOI at $145M. Clean revenue, sticky logos, the kind of unit economics a private equity buyer dreams about. Then the buyer's counsel started reconciling the cap table line by line and found a $3.4 million hole hiding inside a 2019 convertible note conversion. Years earlier, someone had manually overridden a formula in the master Excel sheet to honor a pro-rata right and never rippled the resulting dilution through the Series B and Series C rounds that came after. One keystroke. It stalled the deal 19 days, burned roughly $215,000 in emergency legal fees, and very nearly killed the transaction outright.
Here is the part founders underestimate: the buyer did not treat that error as a clerical slip. They treated it as evidence. If the ledger is wrong in one place, what else is wrong? The cap table is the only document at the closing table that defines who gets paid and how much, and when it does not foot, every other representation in the purchase agreement gets re-read with suspicion. In the $50M to $200M tech band specifically, too big for a quick founder-to-founder handshake and too small for the seller to dictate terms, that suspicion converts straight into price.
The numbers around this are not abstract. Carta's State of Private Markets data shows roughly 73% of Series B and C cap tables carry material mathematical or legal errors heading into diligence. And per PitchBook's M&A deal terms analysis, equity-ledger anomalies add 14 to 21 days to tech diligence timelines on their own. Time is leverage, and on a stalling deal the leverage is the buyer's. Every day lawyers argue over a mispriced warrant is a day terms can be renegotiated, or a day the buyer remembers they have other targets.
Founders treat the cap table as a living document they will tidy up later. Buyers treat it as an immutable legal ledger that already governs their money. Knowing your founder ownership benchmarks for Series A, B, and C tells you what you think you own. This is about proving it, to a hostile reader who profits from every gap.
Where the bodies are buried: ghost equity, SAFE math, and the waterfall
Three failure modes account for most of the bleed, and they are specific enough that you can go look for them this week.
Ghost equity. These are the grants that exist in spirit but not on paper: departed co-founders, early advisors, that contractor you promised "0.25%" in a Slack message. In companies that ripped from $10M to $50M ARR, the pattern is almost always the same. Leadership issued options to land a critical engineering hire, but the board consent approving the strike price lagged by months, or never got signed at all. That gap is a Section 409A landmine. A buyer will not absorb the tax exposure of mispriced options, so they do the next thing: carve out a special indemnification escrow and fund it from your proceeds. The SRS Acquiom M&A Deal Terms Study identifies cap table discrepancies as a leading driver of these special escrows, which lock up around $850,000 of founder wealth for 18 to 24 months after close. That is not a haircut. That is your money, sitting in a third party's account, hostage to a former employee's potential claim.
SAFE and bridge conversion math. The most common arithmetic error we find is not in the equity itself; it is in how convertible instruments resolve at exit. Founders routinely confuse pre-money and post-money SAFE mechanics, and when two or three notes with different caps and discounts convert simultaneously, the interactions are not intuitive. Say a founder ran a $2M bridge on post-money SAFEs thinking the dilution was capped at one number; at conversion it lands several points higher, because the post-money cap dilutes the existing holders, not the new money. We have watched that single misunderstanding silently transfer 3% to 5% of the company to seed investors at the precise moment of exit, value the founder had already mentally spent.
The liquidation waterfall. Participating preferred stock with multiple liquidation preferences is genuinely hard to model, especially in a flat or downside scenario where the preference stack eats most of the proceeds before common sees a dollar. Buyers know this, and they will run their own waterfall regardless of what you hand them. The test is brutal and simple: if your team cannot generate an accurate payout waterfall across five different exit prices, say $90M, $120M, $145M, $170M, and a downside $70M, in an afternoon, your cap table is not ready. Whoever runs the cleaner waterfall controls the conversation, and right now that is not you.
The 120-day fix: do this before you sign anything
Start the cleanup 120 days before you sign an LOI, not after. Once diligence opens, every correction looks like a confession. Here is the sequence that actually holds up.
Get off the spreadsheet, then audit paper-to-platform. If you are north of $20M in revenue and still running equity in Excel, fix that first, but understand that migrating to an equity management platform only digitizes whatever errors you already have. The real work is the trace: every single line on the cap table must connect back to a fully executed PDF signed by the grantee, an officer, and the board. Walk the list. Where the paper trail breaks, secure retroactive board ratifications now. They are uncomfortable to chase down and far cheaper than funding a settlement on the eve of close. The buyer will pull a capitalization representation into the purchase agreement; breach it because of a sloppy ledger and the penalty comes straight out of your pocket, not the company's.
Read every option agreement for acceleration triggers. Single- and double-trigger acceleration language is often boilerplate copied from a template years ago, and it quietly reshapes deal economics. In a $100M transaction, an unexpected wave of accelerated options can vaporize $8M earmarked for the rollover pool, money the buyer assumed they were buying. Know exactly which grants accelerate and on what condition; the mechanics are laid out in how acceleration provisions affect exit readiness.
Freeze the cap table. Once you are inside the 90-day pre-LOI window, stop issuing. No secondaries, no new option grants, no warrant tweaks. A moving cap table is an un-diligenceable cap table, and buyers pay for predictability above almost everything. Run your final pass against the cap table cleanup checklist for a technology exit to confirm lockdown.
The signal a clean ledger sends is bigger than the ledger itself. A cap table that foots to the penny tells a buyer the rest of the house is in order. A messy one tells them the opposite, and invites the 15% valuation haircut they were already hoping to justify.
