Reconciliation goes either well or badly depending almost entirely on what happened in the ten minutes before it started. A bookkeeper working from a properly prepared file moves through reconciliation in a quarter of the time it takes someone working from a messy one. Same work. Different preparation.
The seven checks below are the preparation. They take maybe twenty minutes once you've practiced them. Skipping them can easily add an hour to the reconciliation itself, plus the awkward client conversation when something turns out to be wrong.
These apply whether you reconcile in QuickBooks, Xero, Wave, or a desktop tool. The cleanup is upstream of the software.
1. Statement is complete and the balance chains
Open the PDF. Confirm three things. The statement period is what you expected. The starting balance matches the previous month's ending balance, exactly. The page count matches whatever the footer says (most banks include "page X of Y").
The chain check is the one most people skip and the one that causes the worst pain. If last month ended at $12,847.16 and this month starts at $12,801.04, somebody is going to spend an hour finding the gap. Find it now while both statements are open, not two weeks from now mid-reconciliation.
If pages are missing, request them from the client or pull the statement again from the bank. Don't try to reconcile around a hole. The transactions you can't see will haunt you.
2. The PDF is converted to structured data
Reconciliation tools work best when fed clean structured data. They work worst when fed PDFs that have to be eyeballed line by line. The first real cleanup step is conversion.
Upload the PDF to Ledgertome (or whatever converter you use). Pick the output format your accounting software wants, CSV for general purpose, QBO for QuickBooks Online's bank feed, Excel if you want to do further cleaning, QIF for older tools. Save the file in the client folder with a consistent name like ClientName_AccountName_YYYY-MM.csv.
Doing the conversion at the start, rather than partway through, means every subsequent check runs against structured data instead of a PDF. Sorting, filtering, summing, all of it is trivial in a spreadsheet and painful in a PDF.
3. Totals match the statement summary
Most bank statements include a summary with total deposits, total withdrawals, and ending balance. Use them. Sum the deposit column in your converted file and confirm it matches the statement's total. Same for withdrawals. Confirm starting balance plus deposits minus withdrawals equals the ending balance.
If everything matches, the conversion captured every transaction. If not, find the problem before going any further. Common causes: a transaction that wrapped across pages and got split or dropped, or a transaction with an unusual sign that landed in the wrong column. Two minutes here saves an hour of mid-reconciliation confusion.
4. Date format is standardized
Bank statements come in every imaginable date format. "1/3/24." "Jan 03, 2024." "03 Jan." Sometimes different formats in different columns of the same statement.
Pick your firm's standard, most teams use ISO YYYY-MM-DD because it sorts correctly everywhere, and convert every date in the file. A few minutes of Find and Replace or a quick formula handles it.
While you're standardizing, scan for obviously wrong dates. A transaction dated next year, or before the statement period, is a red flag. Usually it's a typo in the original statement (it happens), but sometimes it indicates a conversion error worth investigating.
5. Descriptions are clean
Bank descriptions are notoriously messy. Extra whitespace, inconsistent capitalization, vendor names that change format month to month, transaction codes that mean nothing to a human.
Trim leading and trailing whitespace from every description (Excel's =TRIM does this in one column). Strip non-printable characters with =CLEAN. Use =SUBSTITUTE or a vendor normalization lookup to collapse common variants, turning "AMAZON.COM*MKTPLACE" into "Amazon Marketplace" so downstream categorization rules can match it.
This is also a good moment to apply your firm's standard memo column if you use one. The cleaner the description data, the less manual categorization you'll do downstream.
6. No duplicates and no anomalies
Scan for two problems.
First, duplicates. Transactions that appear twice in the file, usually because the conversion picked up a header row or the original statement showed the same transaction in two views. Excel's Remove Duplicates handles this, but be careful, a legitimately identical transaction (same vendor, same amount, same day) isn't a duplicate. Use date, description, and amount together as the match criteria, and review any flagged rows before deleting.
Second, anomalies. Transactions that look unusual for the client. A wire transfer ten times larger than anything else in the month. A foreign currency charge for a client that does no international business. An unfamiliar vendor with a round-number amount. These aren't necessarily errors, but flag them now and ask the client one focused question instead of discovering the issue mid-reconciliation.
A simple practice: add a "review" column and mark each anomaly with a short note. That becomes your client question list for the month.
7. Cross-check against other records
Before starting the reconciliation proper, do one final cross-check against any other records you have for the period. Receipts. Vendor invoices. Sales reports. Take ten minutes to confirm that significant transactions appear in both places. A $4,200 deposit on the bank statement should trace to a corresponding invoice paid or sale recorded. A $1,800 withdrawal should trace to a vendor payment or payroll run.
You don't need to match every transaction this way, that's what reconciliation itself is for. You're just confirming the big-ticket items line up with other documentation, so you go into the reconciliation knowing the foundation is solid.
A note on order
The checks aren't arbitrary. Each one depends on the trustworthiness of the previous. Completeness before conversion prevents wasted work on a partial file. Conversion before standardizing prevents manual cleanup on a PDF. Totals before description cleaning prevents polishing data that's missing transactions. Cleaning before duplicate-scanning prevents the dedupe from being thrown off by formatting noise.
Skip a step and you usually pay for it later. The cost compounds.
After doing the seven checks for a few months, they become automatic. You'll run them in less time than it takes to read this article. Reconciliation itself will feel like the easy part of the work, which is what it should have been all along.
Visit Ledger Tome