How to Analyse a Bank Statement

Read the balances, then confirm the statement reconciles

Every statement opens with a balance carried forward and ends with a closing balance. Subtract one from the other and you've got the net change for the period — whether the account grew or shrank — before you've read a single line. But those headline figures only mean something if the statement is complete, and the running balance is what proves it.

Banks recalculate the running balance after every transaction, so it's the account's true position at that moment and doubles as a free checksum. Take the opening balance, apply each credit and debit in order, and you should finish on the closing balance to the penny. That's reconciliation: opening, plus all money in, minus all money out, equals closing. Land there and the statement is whole, so build on the figures. Come up short — even by a few pounds — and a transaction is missing or misread, and that gap will sit silently inside whatever you analyse next. Find it first.

It matters most when the numbers came out of a PDF. Extraction slips in predictable places: a row vanishes in the seam between two pages, a 1,480 payment reads as 1,460 when a digit gets fumbled, a debit lands in the credit column and flips its sign. None of those flash red — the export looks clean. The running balance catches them, because the moment one number is wrong, the balance stops tying out.

Export Bank Statement runs this check before you ever see the spreadsheet. It extracts every line, walks the running balance from opening to closing, and confirms the totals reconcile — or it flags the statement instead of handing you a clean-looking file that's quietly short a payment. Most converters skip this; they extract and hope, and the checking falls to you. For analysis that gap is the whole game, because one dropped direct debit can tip a category, hide a shortfall, or bury a red flag you needed to see.

Separate income from outgoings

With the figures verified, split the statement two ways: money in and money out. Most statements already give you those columns, so in a spreadsheet it's a couple of sums. Total each side and the difference between them should match the gap between opening and closing balance you worked out earlier. If those two numbers disagree, something's still off.

On the income side, look past the total to the pattern. Is the money arriving regularly or in lumps, from one source or several? A salary landing on the same date each month reads very differently from a self-employed account where invoices settle whenever clients feel like paying. Confirming income is real, regular and explainable is its own job — the business cash flow analysis guide covers it.

Group the outgoings into categories

A single "money out" total tells you almost nothing; the value is in the breakdown. Sort every debit into a category, keeping the list short enough that you actually read it. For a personal account the spine usually runs housing, utilities, groceries, transport, subscriptions, eating out, and a small "other". For a business: payroll, rent, suppliers, software, marketing, fees, tax, and "other". Keep two habits — every transaction lands in exactly one category, so the parts add back to the verified total, and when "other" creeps past roughly a tenth of total spending, it's hiding a category that deserves its own line.

The built-in analyser does the first pass after conversion, reading each transaction description and sorting it automatically. What you bring is the judgement it can't have, like knowing a payment tagged "supplier" was really a one-off equipment buy. The expense analysis guide has the fuller walkthrough.

Pick out the recurring payments

Recurring items are the easiest spending to lose track of and the easiest to act on — standing orders, direct debits, subscriptions, finance repayments, anything landing at a steady amount on a steady rhythm. A small repeat compounds: a forgotten 9.99 subscription is 120 a year doing nothing, and most accounts carry two or three.

Sort on the description and scan for the same merchant month after month, or let the analyser's recurring detection flag and total them for you. One caution: recurring isn't the same as essential. A payment you can cancel tomorrow is a candidate for cutting; a mortgage payment isn't. The tool finds the repeats — you decide which earn their place. The recurring transactions guide goes deeper.

Scan for red flags

The last read is the suspicious one. With the data verified and grouped, look for what doesn't fit: a balance that hugs zero or dips into the overdraft repeatedly, round-number transfers that match no normal pattern, gambling payments, large cash withdrawals, or returned direct debits and unpaid-item fees that say the account was stretched. For anyone reviewing someone else's statement — a lender, a landlord, an accountant — a doctored closing balance is the classic one, and reconciliation exposes it instantly, because a tampered figure breaks the running-balance maths. The bank statement red flags guide lists them in order.

How to analyse a bank statement, step by step

You can do every bit of this by hand: foot each page, tick the running balance, then key each line into a spreadsheet and tag it. For one quiet month, fine. For a busy account across several months it's a long afternoon, and the re-keying is where fresh errors creep in. The quicker route, on data you can trust:

1. Gather the statements. Download the PDFs from online banking — several consecutive months if you want patterns to show. A scan or phone photo of a paper statement works too; it'll be read by OCR.
2. Convert them to Excel. Upload the PDFs at /convert. Every line comes across — date, description, money in, money out, balance — across every page, so you can sort, sum and filter instead of reading down a PDF.
3. Check the reconciliation result first. Confirm the totals reconcile before you trust a single figure. A flagged statement means you've caught a missing line a manual skim would probably have waved through.
4. Read the balances and net movement. Opening to closing, and the difference between them; make sure money in minus money out agrees with that gap.
5. Split income from outgoings, then categorise. Open the analyser: spend grouped by category, recurring payments and merchants surfaced, each group totalled. Fix any miscategorised lines — only you can do that.
6. Scan for recurring drains and red flags. Cancel what you don't use; flag what looks wrong.
7. Export. Take the clean .xlsx for your own working, or the CSV in the native bank-import format if the data's headed for Xero, QuickBooks or Zoho Books.

One verified month usually takes well under a minute, and that's the point — when it's that quick, you look monthly instead of once a year.

A worked example

A small services account, three months, roughly 200 transactions a month across a six-page PDF. Footing the first month by hand, page four didn't tie — a 126.40 payment had dropped into the join between pages three and four, understating the outgoings by exactly that. Nothing on screen looked wrong; the totals just sat quietly short, which is the worst kind of error because you trust it and act on it.

Run the same PDFs through the converter and the reconciliation check flagged that month in seconds. Once fixed, the real shape showed. The closing balance ended slightly up — looked fine. But the running balance had dipped into the overdraft twice mid-month, both times the week before the largest client settled, and five subscriptions had crept to 73 a month between them, two unused. None of that shows in the headline numbers; all of it is obvious once verified figures are sorted and grouped.

What this is, and what it isn't

Worth being precise, because it's easy to assume more than the tool does. Export Bank Statement converts your PDF statements, verifies they reconcile against the running balance, and exports a CSV in the native bank-import format Xero, QuickBooks and Zoho Books expect. The honest caveat: this is a file import, not a live bank feed. It doesn't push transactions into your ledger through a certified bank-feed API — you convert, then you import the CSV, and it lands as reconcilable statement lines. It's a tool you run yourself, not a bookkeeping service: the reading of the statement, and the calls you make off it, stay yours. It just hands you complete, verified figures. Files are processed then deleted, and never used to train AI.