When a business closes, changes hands, or faces an audit, the accounting records become evidence — and the standards shift from "we have them somewhere" to "we can prove what they said and that they haven't been altered." This guide covers what authorities and counterparties actually need, how long records have to survive, and what a defensible archive looks like.
Most accounting records spend their lives in motion — invoices entered, payments applied, bank feeds reconciled, statements run, books closed for the month. The data is alive, changing daily, and the version that matters is the current one.
Three kinds of events break that pattern. Each one freezes the records at a specific point in time and gives them a job they didn't have before: defending themselves to someone other than the people who created them.
Each of these events shares a structural property: the records need to prove themselves later, often years after the people who created them have moved on. This guide walks through what that requirement actually means, how long records need to survive in different scenarios, and what discipline turns ordinary accounting data into evidence that holds up.
Related situations covered separately: If your situation is specifically a QuickBooks Desktop sunset migration, see What Happens to Your Accounting Records When QuickBooks Desktop Sunsets. If you're inheriting a QuickBooks file you didn't create, see Inheriting a QuickBooks File: A Guide for Executors, Buyers, and New Bookkeepers.
The end of operations is not the end of the records' obligations. When a business closes, the records continue to face demands from authorities and counterparties for years — often decades — after the last invoice was written. The closure itself doesn't extinguish those obligations. It just removes the people and infrastructure that were maintaining them.
The most common post-closure events that demand the records:
"How long do I have to keep records?" doesn't have a single answer. It depends on the record type, the jurisdiction, the underlying event, and what was reported on which return. The IRS publishes a general retention table:
| Period | Applies to |
|---|---|
| 3 years | The default. Records that support an ordinary return, where no fraud or substantial omission is involved. |
| 4 years | Employment tax records, after the tax becomes due or is paid. |
| 6 years | Returns where more than 25% of gross income was not reported. The IRS gets the longer window when substantial income is omitted. |
| 7 years | Returns claiming a loss from worthless securities or a bad-debt deduction. |
| Indefinite | If no return was filed, or if a fraudulent return was filed. No statute of limitations applies. |
State retention requirements add their own layer. The conservative practice — and the one most CPAs and attorneys actually recommend for closures — is seven years from the closure date, treating that as the minimum window that covers the broadest set of possible inquiries.
For businesses with PPP/EIDL exposure, regulated industries, real estate transactions, or pending litigation, longer windows are common. Some attorneys advise retaining closure records indefinitely.
The work of post-closure records isn't to be readable — it's to be defensible. There's a difference. A box of receipts in an attic is readable. It is not defensible against a sophisticated audit that asks: "How do we know these are the original records, and not something you assembled three years later when you got the IRS notice?"
Defensibility has three components, and "we have them" only addresses the first:
A defensible closure record-keeping process, ordered by what actually matters:
Step five — the seal — is the step most closures skip and the one that creates the most leverage years later. It's also the cheapest and fastest of the six. The cost asymmetry is real: the seal costs almost nothing at closure, and it can be the difference between a clean response to a 2031 audit notice and a multi-thousand-dollar reconstruction project.
A business sale is structurally a record-handoff event. The buyer is taking possession of historical accounting data they didn't create and didn't watch being created. The seller is releasing records they have to be able to stand behind, sometimes for years after the deal closes. Both parties have an interest in a clean, fixed-point-in-time version of the books that neither side can later dispute or modify.
That mutual interest is often handled informally — a copy of QuickBooks data emailed to the buyer's accountant, a folder of PDFs in a data room, a backup file on a thumb drive. These approaches work for the small transaction where both sides trust each other. They fall apart when the deal is contested, when post-closing adjustments are disputed, or when the buyer discovers something months later that they want to renegotiate.
The structural weakness of most M&A record-handoffs is that one party controls the canonical version. The seller hands the buyer "the books." The buyer takes possession. If a dispute arises later about what the books said at the moment of sale, the parties are arguing about which version is authoritative — typically the version each side happened to hold or modify after the closing.
A sealed archive solves this differently. The same data, cryptographically hashed at the closing date, is held by both parties. Neither side can modify their copy without the hash mismatching the other side's copy. Disputes shift from "whose version is correct" (unanswerable, since both sides have their own version) to "what does the data actually show" (tractable, since both sides have the same data).
This pattern is borrowed from how serious litigation discovery has worked for decades. Document preservation in federal litigation requires that produced documents be demonstrably unmodified from the moment of production — typically with chain-of-custody documentation and hash verification. The same discipline applied to business sale transactions costs almost nothing at the deal table and prevents an entire category of post-closing disputes.
A buyer's diligence team or post-acquisition integration manager is looking for:
A QuickBooks backup file hands the buyer the data but doesn't satisfy any of these except the first. They have to install the right version of QuickBooks to read it, take the seller's word that the file is complete, and trust that nothing was modified between the closing date and the handoff. The buyer's accountant typically asks for additional formats — Excel exports, PDF reports, sometimes raw database access — to get the verifiability the QBB file alone doesn't provide.
The seller's interest is symmetric and often underappreciated:
The seller often needs the records more than the buyer does. The buyer gets the going-concern business and its forward-looking value. The seller gets the cash and walks away — but also retains all the personal liability for what was represented at the closing table. A sealed archive is, in effect, the seller's defense file.
This adds a few hours of work to the closing process. It avoids the standard post-closing scenario where one side accuses the other of misrepresenting the books and the parties spend tens of thousands of dollars on forensic accountants to reconstruct what should have been a simple reference point.
An audit — IRS, state, lender, financial-statement, regulatory — is the canonical case where accounting records become evidence. The auditor's job is to assess what the records say and whether the records can be trusted. Most of the value at stake in an audit isn't in the underlying numbers; it's in whether the auditor accepts the records as authoritative or treats them as a starting point for additional inquiry.
The structural problem most small businesses face when preparing for an audit is the same problem that makes closures hard: the records are alive. Books continue to change after the audit period closes. By the time the auditor arrives, the live system shows the current state, not the state at the audit-period close. The auditor wants the latter. The business has to reconstruct it.
A typical small-business audit — whether routine IRS examination, financial-statement review, or lender due diligence — works from these inputs:
The last item is the one that creates the most friction. If the auditor suspects the records were reconstructed or modified after the period-end date — either to fix errors or to support a particular tax position — they shift from accepting the records to investigating them. That shift is what turns a routine audit into an expensive one.
The technical name for what auditors want is "contemporaneous records." Records produced in the ordinary course of business, dated when the underlying transactions occurred, unmodified since. Contemporaneous records are trusted at a different level than reconstructed records, even when the reconstructed records show identical numbers.
The challenge for a small business is that the accounting system itself doesn't preserve contemporaneity by default. A transaction entered in March can be modified in November without leaving a clear audit trail. Account balances at year-end can shift months later if anyone corrects an old entry. By the time the auditor asks for the year-end balance sheet, it may not be the same balance sheet that existed on January 5th when the books were originally closed.
The most useful sealing pattern for audit preparation is annual: at fiscal year-end, after the books are closed and final adjusting entries are posted, the records are sealed into an archive that can be referenced for any audit involving that year.
This pattern has several practical advantages over reactive sealing (sealing only when an audit is announced):
If the audit notice has already arrived, sealing the records as they currently exist still adds value, but it serves a different purpose. The seal doesn't make the records contemporaneous (that ship has sailed) — it documents that no further modifications were made after the notice was received. This matters because modifications after notice can be characterized as obstruction or evidence destruction depending on the audit type.
For audits in progress, the typical pattern:
This is more work than handing over a clean year-end seal that was made at the time. It's still substantially better than handing over only the current live version of the books, which gives the auditor nothing to anchor against and invites questions about what changed between period-end and audit.
Across all three scenarios — closures, M&A transactions, audits — the same five properties separate a defensible archive from a folder of PDFs:
The archive must be readable by someone other than the person who made it, using tools they have available. A QuickBooks backup file requires QuickBooks to read. A proprietary accounting export requires the original software. A SQLite database with a documented schema can be queried by any analyst with SQL access — your CPA, the buyer's diligence team, the auditor, opposing counsel, your future self in 2031. The open format is what makes the archive useful when the original software is gone, the original vendor is out of business, or the records need to be reviewed by someone unfamiliar with your specific tools.
The hash is what proves the archive hasn't been modified since sealing. The hash being held independently of the archive — typically by your attorney, your CPA, and at least one third location — is what proves the hash itself hasn't been modified. A hash stored alongside the archive proves nothing, because anyone who could modify the archive could also update the hash to match.
The hash combined with a trusted timestamp (a third-party-issued time-stamp using a standard like RFC 3161) is what proves when the seal was made. Without the timestamp, the seal proves only that the data hasn't changed since the hash was generated — but the hash itself could have been generated last week. With the timestamp, the seal proves the data was in its current form at a specific moment that a third party can independently verify.
The archive is only useful if someone can read it. A documented schema means the table structures, column meanings, and relationships are explicit and published. Anyone with the schema can write queries against the archive. This matters because the people who will need to query the archive years from now — auditors, attorneys, the buyer's team — won't be the same people who created it. The documentation has to be self-contained.
The archive is yours. It lives in your possession, not on a vendor's servers. The vendor that produced the seal — whether that's Sealed Ledger, a forensic accounting firm, or an enterprise litigation-support platform — should be irrelevant to whether you can read your own archive in 2031. If accessing your records requires the vendor's continued existence, you have a custody arrangement, not a preservation arrangement.
The right next step depends on where you actually are:
Don't wait until the closure date to think about records. The cleanest seal is one made before operations wind down, while the data is still being entered cleanly and the people who know the books are still engaged. The week of closure is the right window: final reconciliations done, all entries posted, books closed, then the seal made before the entity dissolves.
Raise the closing-data seal during deal negotiation, before letter of intent. It's a single paragraph in the LOI specifying that closing data will be sealed at the transaction date, with both parties holding copies. Adding it later, after deal terms are agreed, often gets pushed off as "we'll handle it informally" — which is how the disputes that the seal would prevent get created.
Request the seal as part of standard due diligence. The seller almost always benefits from the same arrangement they may not have thought to ask for. Frame it as protection for both parties, not as distrust of the seller. The seller often welcomes it once they understand the symmetric protection.
Seal annually. Build it into the year-end close process the same way bank reconciliations and tax-return preparation are built in. The cost is small. The defensive value is large. The seal is most useful when you don't yet know which year will be audited.
Seal the current state of the books immediately upon receipt of notice, then work backward to reconstruct the period-end version the auditor is asking about. The immediate seal documents that no further modifications happened after notice. The reconstruction gives the auditor what they came for. Both pieces, together, defend against the questions a less-prepared response would invite.
The discipline scales down to the simplest case. If the records matter at all and you're not sure when they'll matter, sealing them at any clean point creates a defensible reference for any future event. The seal is cheap. The reconstruction it prevents is not.
This article relies on the following authoritative sources. Each is publicly accessible and independently verifiable.
Sealed Ledger is built for exactly the situations this article describes — businesses where a closure, transaction, or audit creates a moment when the records need to defend themselves later. We don't do live bookkeeping or tax filing. We produce the seal, hand over the archive, and step out. The archive is yours and remains yours.
If you're at one of these inflection points and aren't sure whether what we do is a fit, the right next step is a short conversation. Tell us where you're stuck and we'll tell you honestly whether a sealed archive helps your situation. If it doesn't, we'll say so.