Most founders prepare for diligence by organizing documents. That is not wrong, but it is not what determines the outcome. A buyer's team is not grading your filing. They are testing a small number of specific claims, and every one of them can move the price.
I have sat on both sides of this — preparing companies for financing and sale, and doing the diligence on the other side of multi-million dollar transactions. The pattern is consistent enough to plan around.
The one question underneath every other question
Everything a diligence team does is a version of the same question: can I trust these numbers, and can I trust the people who produced them?
Those are separate. A company with slightly messy books and a controller who can explain every irregularity on the spot tends to sail through. A company with beautiful reports and nobody who can say where a figure came from does not. Buyers know that the second situation is where surprises live.
This is why the first thing a good diligence team does is not read your P&L. It is pick two or three numbers and ask you to walk them back to source.
What they check first
1. Quality of earnings — is the EBITDA real?
The headline number gets adjusted, and the adjustments are where the negotiation happens. Expect them to look for one-time revenue booked as recurring, owner expenses running through the business, deferred revenue recognized early, capitalized costs that should have been expensed, and related-party transactions on non-market terms.
Prepare by doing it to yourself first. Build your own quality-of-earnings bridge from reported EBITDA to what you believe adjusted EBITDA is, with every adjustment labeled and supported. Handing the buyer your own bridge changes the dynamic entirely — you are now negotiating from a document you authored rather than defending one they built.
2. Revenue recognition — does it survive contact with the contracts?
Buyers pull a sample of customer contracts and check them against how you recognized the revenue. Multi-year deals, usage-based pricing, implementation fees, and anything with a milestone are where discrepancies surface.
If revenue recognition currently lives in a spreadsheet, assume it will be questioned. Not because spreadsheets are wrong, but because a spreadsheet cannot show its own history.
3. Working capital — the adjustment nobody models until it costs them
Purchase agreements almost always include a working capital target, set from a trailing average. If your working capital at close is below the peg, the price drops dollar for dollar.
Founders routinely leave money here because they did not model it. Track your net working capital monthly for at least twelve months before a process, understand its seasonality, and k now what a normalized level actually is. This single item moves more money at close than most of the diligence findings that get more attention.
4. Customer concentration and retention
They will build a cohort analysis whether or not you give them one. If a third of revenue sits with two customers, that is not fatal, but it is a discount unless the contracts are long, assignable, and recently renewed.
Check assignability now. A change-of-control clause requiring customer consent gives your largest customers leverage over your transaction, and you will discover it at the worst possible moment.
5. The cap table and the 409A
Option grants that were never formally approved. Advisor equity promised in an email. A SAFE whose conversion nobody has modeled. A 409A that is stale relative to the price being discussed. These are common, they are all fixable in advance, and every one of them is expensive to fix during a live process.
6. Employee classification and payroll
Contractors doing full-time employee work. Unpaid overtime exposure. Equity issued without a plan document. State registrations missed for remote employees. None of this is glamorous and all of it can produce an indemnity holdback.
The thing that quietly decides how the process feels
Can your accounts be reconciled, on demand, by someone other than you?
This is the item I care most about, and not only because I built reconciliation software. Itis the closest thing to a leading indicator of how a diligence process will go.
Here is what happens in practice. The buyer asks for the bank reconciliation supporting a month you closed nine months ago. If the answer arrives in an hour with the supporting detail attached, you have just told them something important about the whole company. If the answer takes four days and comes with a caveat, you have told them something else — and they will respond by widening the sample.
The scope of a diligence process is not fixed at the start. It expands when the answers are slow or inconsistent. Every discrepancy that takes a week to explain buys you two more requests. I have watched a three-week process become a three-month process for no reason other than that nobody could quickly tie a merchant settlement file back to the general ledger.
Where this most often breaks:
- Payment processors. A processor deposits net of fees, refunds and chargebacks, so a single bank line corresponds to hundreds of underlying transactions, and no field ties them together.
- Multiple collection channels. Cards, ACH, wires, checks and a marketplace, each with a different export format, timing and identifier.
- Timing. A transaction dated the 31st that settles on the 2nd sits in different months on the two sides. Both are correct. Neither matches.
- Intercompany. Two entities book the same transaction under different naming conventions, and the elimination never quite eliminates.
None of this means anything is wrong with your business. It means the evidence that nothing is wrong is expensive to produce — and diligence is entirely an exercise in producing evidence cheaply.
What to do, and when
Twelve months out. Get monthly close under ten business days. Start tracking net working capital. Move revenue recognition out of a spreadsheet and into something with an audit trail. Fix contractor classifications.
Six months out. Run your own quality-of-earnings analysis. Read every material customer contract for assignability and change-of-control. Clean the cap table — every grant approved, every SAFE modeled, 409A current.
Three months out. Build the data room before you need it. Reconcile every balance sheet account and keep the support. Write the narrative for anything unusual in the numbers, because you will explain it either way and it is better done in writing on your schedule.
During. Answer fast, answer completely, and never let a question go unanswered while you investigate — say what you know, say when you will have the rest, and hit that date.
The uncomfortable part
Most of the value in diligence preparation is captured in the twelve months before anyone starts. By the time a buyer is in the data room, you are managing a process rather than improving a position.
That is the real argument for having senior finance in place before you think you need it. Not because a fractional CFO makes the diligence go faster — though it does — but because the practices that make diligence uneventful are the same practices that make the business easier to run in the meantime. Closing in eight days, reconciling on demand, and knowing your working capital are not diligence tasks. They are just what a well-run finance function looks like, and diligence is simply the first time somebody else grades it.
Doug Schiller is a fractional CFO for technology and fintech companies and the founder of MatchData.ai, patented reconciliation software (U.S. Patent11,475,026). He has led due diligence on multi-million dollar transactions and prepared companies for financing and sale.
Schedule a call to talk through where your company stands.