How to do Bank Reconciliation Step by Step Guide

Bank reconciliation is the routine process of comparing your internal cash or ledger records against bank statements to confirm that balances and transactions agree. Doing bank reconciliation consistently reduces errors, uncovers missing transactions, and prevents surprises at month end.

This article outlines a practical, step-by-step approach you can use whether you perform reconciliations manually or with reconciliation software. It combines concrete actions, checks to run, and signals for when automation will save time and reduce risk.

We start with why the task matters, break down core components, then walk through a repeatable process you can apply to monthly bank statement reconciliation.

Why this topic matters

Timely and accurate reconciliations are foundational for cash visibility, month-end close, and financial control. For finance teams of SMBs, startups, and accounting firms, unresolved bank variances can lead to misreported cash, delayed closes, and harder audits.

A disciplined reconciliation process helps teams detect missing deposits, duplicate or unauthorized payments, bank fees, timing differences, and bookkeeping errors before they compound into larger problems.

When paired with reconciliation software, teams can reduce manual ticking and tying, maintain clear audit trails, and shift effort from searching to investigating true exceptions.

Core components of bank reconciliation

A reliable reconciliation workflow has three main parts: the data inputs, the matching engine (rules + AI), and the reconciliation outputs and controls.

Data inputs: Side A and Side B

• Side A: Your internal records — cash book, ERP bank ledger, payments register, or sales reports. These should include date, amount, and an identifier where possible.
• Side B: The bank statement file or bank-provided transaction report. This typically contains value date, amount, bank reference or narration, and running balance.

Supporting data such as fee schedules, refund reports, or remittance files can be used to enrich either side without being reconciled directly.

Normalization and matching rules

• Standardize dates, amounts, and text fields before comparison.
• Use identifier matching where available (UTR, transaction ID, remittance reference). This is the highest-confidence signal.
• Fall back to date + amount matching when identifiers are missing; allow reasonable timing windows for cleared vs posted dates.
• Support grouping and net-to-net matches for summarized bank entries versus detailed internal records.

Output categories: matched, partially matched, unmatched, skipped

• Fully matched: Identifiers and amounts reconcile according to the configured logic.
• Partially matched: Identifiers match but amounts differ — likely requires an adjustment or split/fee handling.
• Unmatched: Present only on one side and needs investigation.
• Skipped: Records excluded due to invalid or missing required data; these are visible so the user can fix input issues.

Practical implementation steps

Follow these steps to perform bank statement reconciliation consistently.

Step 0: Prepare and gather files

1. Export your internal cash ledger or bank ledger for the reconciliation period.
2. Download the bank statement or transaction report covering the same period (CSV/XLS/XLSX). Include sub-ledger exports if needed (e.g., payment gateway settlements).
3. Collect supporting files: fee schedules, refund logs, merchant settlement reports, or order metadata if relevant.

Step 1: Standardize and map columns

1. Confirm headers and data types: date column, amount column, and primary identifier column(s).
2. Normalize date formats and currency signs; trim text fields and clean narration strings.
3. Create or derive columns as needed (for example, a net amount after fees, or a cleaned reference extracted from a narration).

Tips:

• If using automation or a reconciliation tool, configure a reusable template per bank or partner to avoid remapping each period.

Step 2: Run deterministic matching

1. Apply strict identifier equality matches first (UTRs, transaction IDs, invoice numbers).
2. Allow one-to-many and many-to-one matches for grouped bank entries or aggregated settlements.
3. Use date + amount matching with a sensible timing window when identifiers are missing.

Deterministic rules resolve the high-confidence matches quickly and reduce the volume of exceptions.

Step 3: Review AI or manual matches and resolve exceptions

1. Review partially matched items (same identifier, different amounts) to identify fees, chargebacks, or posting errors.
2. Investigate unmatched items by checking supporting files and timing differences (e.g., deposits in transit or outstanding checks).
3. For ambiguous cases, document reasoning and, if required, perform a manual match with a clear audit note.

When available, AI-assisted matching can suggest likely matches for unstructured references or inconsistent formatting, but always verify low-confidence suggestions before approving.

Step 4: Approve, document, and post adjustments

1. Create a short reconciliation report summarizing matched totals, exceptions, and recommended adjustments.
2. Post correcting journal entries for true bookkeeping errors (and link them to reconciliation references).
3. Archive the reconciliation file and supporting evidence for auditability.
4. Convert learnings into improved mapping and derived column rules for the next period.

Numbered checklist for monthly close:

1. Upload files and run matching within 48 hours of receiving the bank statement.
2. Triage high-value exceptions first.
3. Document every manual match and adjustment with a reason and preparer initials.
4. Complete final sign-off and store the report with supporting files.

Common mistakes to avoid

• Relying only on narration text without standardizing references first.
• Ignoring small unmatched amounts; they often indicate systemic issues like rounding or fee handling.
• Forcing low-confidence matches to artificially clear exceptions.
• Not keeping a clear audit trail for manual matches and journal adjustments.
• Repeating manual remapping each period instead of creating reusable templates.

Key Takeaways

• Reconcile using standardized inputs, deterministic rules, and a controlled review of exceptions to maintain accurate cash records.
• Prioritize identifier-based matches, then date+amount and grouped matching for summarized entries.
• Use derived columns and supporting files to handle fees, refunds, and mapped references consistently.
• Document manual matches and adjustments with clear audit notes and retain supporting evidence.
• Consider reconciliation software to reduce manual work, speed high-confidence matches, and keep an auditable trail.

Conclusion

A repeatable bank reconciliation process — combining data standardization, rule-based matching, exception review, and clear documentation — improves cash accuracy and shortens close cycles. Implementing these bank reconciliation steps with the right tools helps finance teams move from manual ticking to exception investigation.

If you want to see how reconciliation software can speed the process and provide audit-ready results, Start your 14-day free trial with Cointab. No credit card required. 14-day free trial.