Failure mode 1: no audit trail.
When the trial balance does not match last month's, in Excel you have no way to tell what changed and who changed it. A formula reference moved, a row got inserted, a number got typed over a formula. Three hours later you have rebuilt the trial balance but you do not actually know whether the new version is right. At $1M+ revenue, this happens regularly enough that the time cost alone justifies switching.
A real accounting system records every change as a transaction with user, timestamp, and old-vs-new values. When something looks wrong, you can replay the history. When the tax authority or an investor asks "why did this number change in March?", you can answer with evidence. Excel cannot do this without scaffolding that nobody actually maintains.
Failure mode 2: multi-user collisions.
Two people editing the same Excel file at once: one wins, the other's changes are lost. Cloud Excel and Google Sheets help but only partially — formulas calculated against stale references produce silent errors. At 5+ users, the daily collision rate makes spreadsheet-as-ledger actively dangerous. People stop trusting the file, which means they stop using it correctly, which makes the file worse.
A real accounting system handles concurrency at the database level. Two AP clerks entering bills at the same time both succeed. The owner pulling a P&L while a sales clerk records a sale gets the latest committed view. This sounds basic because it is — every business application built since 1995 handles this — but Excel has never been one of them.
Failure mode 3: no branch P&L.
Multi-branch in Excel means either one tab per branch (consolidation is manual misery) or one file with branch tags in a column (pivot tables that break when a column moves). Either way, by branch 3 the consolidation process introduces errors and by branch 5 nobody trusts the consolidated view. Decisions get made on gut because the numbers are unreliable.
A real accounting system tags every transaction with a branch at entry. Branch P&L is a query, not a build. Consolidation is automatic. The owner can see consolidated P&L, branch P&L, and branch-vs-branch comparisons in seconds. This single capability is worth the migration cost for any business with 3+ branches.
Failure mode 4: no inventory ledger.
Excel inventory tracking depends on someone updating quantities by hand after every sale and purchase. Across 5 branches with 2,000 SKUs, the gap between recorded inventory and actual inventory grows daily. Stock-outs happen on items that "should" be in stock. Dead stock piles up because nobody can see it. Annual inventory counts produce 8-15 percent variance every year and the cause is "Excel got out of sync" which is unfixable.
A real inventory system updates quantities on every sale and purchase, automatically. Branch-scoped inventory ledgers (which Nonari uses) keep per-branch quantities and weighted-average cost. Annual count variance drops to 1-2 percent and the variance is real (theft, breakage, miscount), not a system artifact.
Failure mode 5: AR and AP visibility.
Excel-based AR tracking quickly becomes a separate sheet that gets out of sync with the GL. Aged receivables reports are built once and then become stale. Nobody knows the current AR balance without rebuilding. AP is the same: the bill arrives, gets paid, but the AP listing still shows it open. By $1M+ revenue, the gap between sheet and reality is large enough to mismanage cash.
A real accounting system has AR aging and AP aging as live queries against the ledger. They are always current. Dunning automation runs against the live data. Cash forecasts use the live data. The SMB stops mis-managing collections by 2-3 weeks because the data is finally trustworthy.
Failure mode 6: tax and compliance.
Excel does not calculate withholding correctly across vendor classifications (1099 vs not, CIS vs not), does not maintain vendor tax-ID masters, does not generate sales tax / VAT / GST returns from the underlying transactions, and does not produce audit-ready financial statements. SMBs handle these manually as a separate process, which is where errors creep in. At $1M+ revenue, tax errors cost real money in penalties and reassessments.
A real accounting system handles withholding automation, vendor masters, indirect-tax return generation, and audit-ready statements as standard features. Compliance becomes a system output, not a manual exercise. The risk of penalties drops sharply.
When to switch and what to switch to.
The trigger points: revenue above $1M, more than 2 active users on the books, more than 1 branch, or any plan to raise capital, sell, or get audited. If two of these are true, start the switch. If three, you are already late.
Options for growing SMBs: QuickBooks (multi-branch and global-ops pain points, see the dedicated post), Xero (clean UI, similar limits at scale), Oracle NetSuite or SAP Business One (capable but enterprise-priced), ERPNext (free but heavy, see the dedicated post), or modern AI-native systems like Nonari (multi-branch and AI-bookkeeping built in, designed for SMB scale). The right choice depends on team size, complexity, and budget. Pick something built for your stage, not the enterprise tier or the spreadsheet tier.
- Revenue above $1M
- 2+ active users on the books
- 2+ branches or locations
- Plans to raise, sell, or get audited
- Inventory above 200 SKUs