Is it acceptable to keep business books in a spreadsheet?

Yes, for a simple business: cash basis, one or two bank accounts, no payroll, modest transaction volume. The IRS does not mandate software — Publication 583 requires accurate, complete records, not a particular tool. The spreadsheet must still be a real system: every transaction listed, categorized consistently, and reconciled to bank statements monthly.

When should a business move from spreadsheets to bookkeeping software?

At the first of these triggers: hiring an employee (payroll liabilities need real accounts), switching to accrual accounting (receivables and payables need subledgers), a second person working in the books, invoicing customers on terms, inventory, or monthly volume high enough that manual reconciliation slips. Any one trigger is sufficient; waiting for several makes the migration harder.

Spreadsheets vs. ledger software: when a spreadsheet is genuinely enough

A spreadsheet is not a lesser form of bookkeeping — it is bookkeeping, if it is kept with discipline. For a cash-basis service business with one bank account, no employees, and a few dozen transactions a month, a well-structured spreadsheet meets every real requirement, including the IRS's: Publication 583 demands accurate and complete records, not a software subscription. The mistake is not using a spreadsheet; it is staying on one past the specific triggers that make it fail.

What a spreadsheet system must include

If you run on a spreadsheet, run it like a ledger:

List every transaction — date, payee, amount, category — from the bank statement, not from memory.
Use a fixed category list (a chart of accounts by another name) and never invent one-off categories.
Reconcile monthly: your spreadsheet's ending balance must equal the bank statement's, and you should be able to show it.
Keep the source documents organized alongside — the folder structure in document management for bookkeeping works for either system.
Separate business and personal completely. A spreadsheet cannot fix commingling; nothing can.

What a spreadsheet cannot do is enforce double-entry. It will happily let debits and credits disagree, because it has no concept of either — the discipline explained in double-entry explained lives entirely in your habits.

The graduation triggers

Each trigger, and why the spreadsheet fails at it:

Trigger	Why the spreadsheet breaks
First employee	Payroll needs liability accounts for withholdings and employer taxes that tie to Form 941 — see Publication 15's employer obligations
Accrual accounting	Receivables and payables need subledgers and aging reports, not a column
Second user	No access control, no audit trail, and simultaneous-edit errors
Invoicing on terms	Tracking who owes what, and matching payments to invoices, outgrows rows fast
Inventory	Cost tracking and quantity on hand need structure a sheet fights
Volume	Past roughly 100–150 transactions a month, manual reconciliation errors compound

Payroll is the sharpest trigger. The employer's deposit and reporting obligations in Publication 15 are unforgiving, and payroll liability accounts kept "in your head" produce exactly the discrepancies that cost penalties.

How to graduate cleanly

Pick the platform using the criteria — not the brand ads — in choosing bookkeeping software.
Reconcile the spreadsheet through the cutover date.
Enter opening balances from the reconciled figures.
Run one month in parallel if you can bear it; confirm the new system reconciles.
Archive the spreadsheet read-only. It is now a historical record — retain it like any other book of account.

The honest summary: the spreadsheet-versus-software question is about triggers, not virtue. Below the triggers, the spreadsheet's simplicity is an asset. At the first trigger, its flexibility becomes the liability.