Startups Bookkeeping Fractional CFO

Getting Your Books Ready for Due Diligence: A Founder's Pre-Raise Checklist

Published October 7, 2026 by Invisible LLC Team · 11 min read

You're a founder. You've been running books the way every sensible early founder runs books: fast, good enough, and with the assumption that you'd deal with it properly later. Then a term sheet gets real, someone sends you a diligence request list, and "later" becomes this week.

Here's the reassuring part and the annoying part in one sentence: what financial diligence usually turns up in founder-kept books isn't dishonesty — it's bookkeeping that was never asked to answer this kind of question. The problems are boring and predictable. Which means they're also fixable in advance, and the order you fix them in matters more than the effort you put in.

Key takeaways: The financial half of a data room is mostly five things — statements, the ledger behind them, revenue detail, payroll and contractor records, and the cap table. What breaks in founder-kept books is almost always the same short list: when revenue was recognized, whether deferred revenue exists at all, how contractors were classified, whether the cap table ties to the equity on the balance sheet, and whether personal spend ever ran through the business. Fix them in dependency order — the chart of accounts and reconciliations first, because everything above them inherits their errors.

What the financial half of a data room actually asks for

Every fund's list looks slightly different, but the financial section converges. Expect to be asked for:

Monthly financial statements for the trailing period since inception or for the last two to three years, whichever is shorter. P&L and balance sheet at minimum, usually monthly rather than annual, because monthly is where the story is.
The general ledger export, or read-only access to your accounting system. Assume someone will look at transaction detail, not just summaries.
Bank and credit card statements for the same period, plus proof the accounts reconcile to the ledger.
Revenue detail — a customer-level or contract-level breakdown that ties to the revenue line on your P&L. For anything subscription-shaped, expect questions about the MRR/ARR schedule and how it reconciles to booked revenue.
The contracts behind material revenue. Signed customer agreements, order forms, any non-standard terms.
Payroll registers and the employee/contractor roster, including which states you employ in and what registrations exist there.
Tax filings — federal and state income tax returns, payroll tax filings, sales tax returns if you collect, and franchise tax where applicable.
The cap table, plus the underlying documents: board consents, stock purchase agreements, option grants, the option pool, and any convertible instruments outstanding.
AP and AR aging, plus a list of commitments and obligations — leases, vendor contracts with term, any debt.
Corporate records — formation documents, bylaws, board minutes, IP assignment agreements.

Two things to notice about that list. First, roughly half of it isn't accounting at all — it's records hygiene, and it's the half founders most often haven't centralized. Second, the accounting half is checkable against itself. A diligence analyst's first move is usually to see whether three numbers that should agree actually agree.

The five things that break in founder-kept books

1. Revenue cut-off — when revenue got recognized

This is the most common finding, by a wide margin. Founder-kept books frequently record revenue when cash arrives, because that's when the notification hits. If you sold an annual contract in November and collected the whole thing in November, cash-basis books show a large November and a quiet January. Accrual books spread it across the service period, and that's the shape an investor is trying to read.

The formal framework here is FASB's revenue standard, Topic 606, which asks you to identify the contract, identify the performance obligations in it, determine the transaction price, allocate that price to the obligations, and recognize revenue as each obligation is satisfied — either at a point in time or over time, depending on how the promise is structured (FASB ASU 2014-09, Revenue from Contracts with Customers). You do not need to become fluent in that. You need your books to reflect it, and you need to be able to say out loud which month a given dollar belongs in and why.

The diagnostic: pick your three largest customers. Can you trace each one's contract to the specific months its revenue landed in? If the answer involves opening your bank statement, this is your first fix.

2. Deferred revenue that isn't on the balance sheet

Directly downstream of the first one. If you've collected for services you haven't delivered yet, that money is a liability, not revenue. Annual plans paid upfront, implementation fees, prepaid credits, multi-month retainers — all of it.

Founder-kept books often have no deferred revenue account at all, which is the tell. When it's missing, the P&L overstates revenue in the collection month, the balance sheet understates liabilities, and the trailing revenue curve an investor is extrapolating from is simply the wrong shape. This is also the item most likely to produce a re-cut of your own metrics mid-process, which is a bad moment to be having.

3. Contractor versus employee classification

Early teams are built from whoever will help. Some of those people were probably employees in substance while being paid as contractors on a 1099.

The IRS applies common-law rules here, weighing evidence in three buckets: behavioral control (does the company direct what work is done and how), financial control (who bears unreimbursed expenses, who has investment in the tools, how the worker is paid, whether the worker can realize a profit or loss), and the type of relationship (written contracts, benefits, expectation of permanence) (IRS, Independent contractor (self-employed) or employee?; detailed guidance in Publication 15-A). No single factor decides it; the whole relationship is examined.

Why diligence cares: misclassification creates a contingent liability, and contingent liabilities get quantified and then negotiated. It also intersects with state registrations — a full-time engineer in a state where you've never registered for payroll is two findings, not one.

The diagnostic: list everyone paid on a 1099 in the last two years who worked primarily for you, on your schedule, using your tools, for more than a few months. That list is the conversation.

4. The cap table doesn't tie to the general ledger

Two systems, one truth, frequently out of sync. Your cap table lives in Carta or Pulley or a spreadsheet; your equity accounts live on the balance sheet. They should reconcile — shares issued, consideration received, the option pool authorized versus granted versus exercised, and every convertible instrument outstanding.

They routinely don't, usually for mundane reasons: a founder purchase was never recorded, an advisor grant was agreed in email and never papered, an option exercise never made it to the ledger, or 83(b) elections can't be located. None of that is dramatic. All of it is friction at exactly the wrong moment, and it's typically counsel's workstream more than your accountant's — which is precisely why it falls between the two and goes unowned.

One deliberate omission here: the accounting treatment of convertible instruments — SAFEs in particular — is genuinely contested territory, and how your specific instruments should be presented is a question for your accountant and your counsel on your actual documents, not something to take from a blog post. What matters for this checklist is narrower and not contested: every instrument that exists should be findable, and the cap table and the ledger should agree with each other.

5. Personal and business expenses commingled

The one founders already know about and hope nobody notices. Somebody does. It's not usually treated as a character issue; it's treated as a reason to distrust every other number, because an analyst who finds a personal flight in software expense now has to check everything else.

The checklist, in the order that actually works

Sequence matters because these items sit on top of each other. Fixing revenue recognition on an unreconciled ledger means doing it twice.

Phase 1 — the foundation. Nothing above this is reliable until this is done.

Reconcile every bank, credit card, and payment-processor account for the full period, to the statement, month by month. No gaps.
Get the chart of accounts into a shape a stranger can read. Categories that mean something, no catch-all "other" absorbing 15% of spend, no duplicated accounts.
Separate personal from business, historically. Reclassify it, book it properly, and stop the practice today.
Confirm you're on the right basis. If your books are cash-basis and your business has contracts spanning months, you're going to be asked to present accrual.

Phase 2 — revenue, which is what gets read most closely.

Build the revenue schedule: by customer, by month, tied to contracts.
Set up and populate deferred revenue. Post the catch-up entries.
Reconcile your reported metrics to your books. If your deck says ARR and your P&L says revenue, know exactly how one becomes the other, in writing.

Phase 3 — people and obligations.

Review every contractor against the three-bucket test. Document the conclusion for anyone borderline.
Confirm payroll registrations in every state where you have a worker.
Collect payroll tax filings, income tax returns, and sales tax filings into one folder.
List every commitment with a term: leases, vendor contracts, debt, anything auto-renewing.

Phase 4 — equity and records.

Tie the cap table to the ledger. Resolve differences before anyone else finds them.
Locate the paper: stock purchase agreements, board consents, option grants, 83(b) filings, IP assignments for every person who wrote code or made a design.
Assemble the data room in the structure the request list asks for, not the structure your Drive happens to use.

Phase 5 — the read-through.

Have someone who has actually been through diligence read your own package cold and write down every question it raises. Those are the questions you'll get. Answer them before you're asked.

If there's one item to move earlier than feels natural, it's number 15. The cost of a surprise in diligence is rarely the finding itself — it's the credibility hit of not having known about it.

What "clean" means, and what it doesn't

Clean does not mean audited. An audit is an engagement performed by an independent CPA firm under professional standards, resulting in an opinion. It's a separate purchase from a separate firm, on a separate timeline. When founders say "Series A audit prep," they usually mean diligence readiness, which is what this post is about. If an investor genuinely asks for audited statements, that's a conversation with an audit firm — and for the record, it isn't something we do.

Clean does not mean pretty. You are not trying to make the numbers flattering. You're trying to make them explainable. A messy quarter you can account for reads as competence. A tidy quarter you can't reconstruct reads as luck.

Clean does mean reconciled, consistent, and traceable. Same accounting policies month over month, statements that tie to the ledger, ledger that ties to the bank, revenue that ties to contracts, cap table that ties to equity. Every number traceable to a source document by someone who doesn't work for you.

And clean means known. If something is unresolved — a classification question, a state registration you're catching up on, a contract with unusual terms — surface it yourself, with a plan attached. Disclosed and in hand beats discovered every single time.

Where this leads

Most of what's on this list isn't a project. It's a monthly habit that wasn't running. Reconciliations, accrual discipline, a close that happens on a date — do those continuously and there is no pre-raise scramble, because the data room is just a folder you export. That continuous version of this work is the controller layer, and we wrote the scoping guide for it in outsourced controller services: what they cover and when you need one.

A quick aside for software founders, because it comes up in the same conversation and belongs in a different post: if you've been capitalizing domestic research costs or wondering whether your engineering spend supports an R&D credit claim, the documentation habits overlap heavily with diligence readiness. We covered that separately in the R&D credit for software startups.

We keep founder books in a shape that survives outside scrutiny — chart of accounts set up properly from the start, Brex/Mercury/Ramp transactions categorized to something meaningful, accrual and deferred revenue handled monthly rather than retroactively, multi-state payroll set up the first time you hire out of state. Not because diligence is coming, though it is. Because you can't make a good hiring or pricing decision from books you don't trust either.

Show us your last month of books and we'll tell you specifically what a diligence process would find — before an investor tells you. You can see how we work with startups and founders, what the ongoing fractional controller layer covers, or read the companion guide on scoping a controller engagement.