Bank Reconciliation

Your month-end close slipped three days last quarter because bank reconciliation kept turning up items nobody could match, and ownership of the exceptions was unclear for two full weeks. Bank reconciliation is the process of comparing your company's internal cash records (the cash account in the general ledger) with the bank's records (the bank statement) to identify and resolve any differences. It confirms that every deposit and payment recorded internally matches what actually cleared the bank account, and that every item on the bank statement has a matching entry in the books. When it works, it's a routine control that takes hours. When it doesn't, it turns into a multi-day investigation that blocks the close. The process itself is rarely what trips teams up. The trouble sits at the boundary between automated matching and manual exception handling. Modern systems match most items automatically. The 3 to 5% that don't match automatically are where the real work is, and that work needs clear ownership that many teams haven't formalized.

How it works

As a process, bank reconciliation follows a standard sequence: get the bank statement for the period, get the GL cash account balance, identify items in one record that don't appear in the other, classify those differences as timing items (transactions in transit) or discrepancies (errors, missing transactions, unauthorized items), and bring the two records into agreement. The reconciliation 'balances' when the adjusted GL balance equals the adjusted bank balance after timing differences. As a control, bank reconciliation does something different: it provides evidence that the cash balance on the balance sheet is accurate, that there are no unauthorized transactions, and that all transactions were properly recorded. These two functions can come apart. A reconciliation can 'balance' in the process sense (the numbers agree) without working as a real control. If timing items are carried forward month after month without investigation, if exceptions are cleared with offsetting adjustments that don't represent real transactions, or if the same person prepares and reviews the reconciliation, the control is compromised even when the numbers balance. Good bank rec processes separate preparation from review, set age limits on unresolved timing items, and require a documented explanation for every exception.

Automated match rates, manual exceptions

Automated bank reconciliation matches bank statement items to GL entries on amount, date, and reference number. When those three fields agree within a tolerance, the system auto-matches. When they don't, because of timing differences, rounding, or data quality problems, the item becomes an exception for manual review. High-volume businesses processing thousands of transactions a month typically see automated match rates of 85 to 95%. The 5 to 15% handled manually can carry significant dollar volume even when the count is small. The distinction teams miss is that a completed reconciliation (every item matched or explained) is not the same thing as an accurate cash balance. A reconciliation can be completed with incorrect matches: two transactions offset against each other that don't actually correspond, or a timing item carried forward instead of investigated. The report shows a balance, but the underlying cash position is wrong. This happens most when teams are under close pressure and clear exceptions to hit a deadline instead of investigating them. The reconciliation schedule should report the age and dollar value of unresolved exceptions at close alongside the match rate.

What vendors mean by 'automated bank reconciliation'

Every bank reconciliation module on the market, whether it sits inside accounting software or a standalone tool, advertises automation. What they mean varies considerably. Basic automation is transaction import from bank feeds (no manual CSV upload) and rule-based matching on amount and date. More sophisticated automation adds fuzzy matching (catching items where amounts are close but not exact due to fees), multi-transaction matching (one bank item matching multiple GL entries), and machine learning that improves match rules over time based on how exceptions are manually resolved. The only reliable way to evaluate what a vendor's automation actually delivers for your organization is to run a reconciliation on your own historical data during the proof-of-concept phase. Request a sandbox, import three months of your bank transactions and the corresponding GL entries, run the automated matching, and measure the exception rate. Vendors who won't support this during evaluation are telling you something about their confidence in their match rates. Also ask specifically: does the system enforce a review and approval step after matching, or can a user mark a reconciliation as complete without a second set of eyes? The control only works if the workflow enforces it.

Questions to ask

• What is your typical automated match rate for companies with our transaction profile? Can you show us data from reference customers with similar volume and transaction types?
• How does the system handle multi-transaction matching, where one bank deposit corresponds to multiple individual AR payments in the GL?
• Is there a mandatory review and approval workflow, and does it create a separate audit trail for the reviewer distinct from the preparer?
• How are unresolved exceptions tracked over time? Is there an aging report for open items, and does the system flag exceptions open longer than a defined threshold?
• What happens when a bank feed connection breaks? Does the reconciliation fail silently, or does the system alert the team before the exception count becomes unmanageable?

Common mistakes

The first is confusing a balanced reconciliation with an accurate one. Teams under close pressure sometimes clear exceptions by forcing matches or booking offsetting adjustments: the rec balances, but the underlying cash records are wrong. This usually comes out at the year-end audit, when auditors ask for support for specific transactions and the documentation doesn't hold up. Fixing it means going back through every forced match and unsupported adjustment, which can take days and delay the audit. The second failure is having no documented owner for exceptions. When the reconciliation turns up unmatched items and it isn't clear whether AP, AR, treasury, or the controller should resolve them, the exceptions sit. Two weeks later, ownership is still disputed and the items are still open. Bank rec processes that work assign exception ownership when the exception is identified, by type (vendor payment, customer receipt, interbank transfer), so there's no question about who investigates what.

Matched against accurate

A high automated match rate is not the same as a reconciled account. Matching confirms that two records agree on an amount and a date. It does not confirm the transaction was coded to the right account, or that a duplicate was not matched twice. When a vendor quotes a match rate, ask what happens to the unmatched remainder and who reviews the exceptions.

Comparing accounting software

The dividing line in this category is whether a tool owns the ledger or sits beside it running reconciliations and the close checklist. Once bank reconciliation is settled as a requirement, read the Accounting Software category guide for how the market splits, then weigh the BlackLine profile against the FloQast profile on that requirement rather than on the feature list. Our BlackLine pricing breakdown records what the vendor publishes and what is quote-based. The BlackLine against FloQast comparison sets two of them side by side on the same criteria. Before you book demos, put your own numbers through the accounting software ROI calculator so the business case is written down before a vendor writes it for you.

Bank Reconciliation questions

Related terms

Account Reconciliation, Accrual Accounting, and Audit Trail come up in the same workflow as bank reconciliation, and the rest of the accounting software cluster is indexed in the finance operations glossary.