Sonomir

Guides · 5 min read

How to convert a bank statement PDF to Excel — and prove nothing was missed

Getting transactions out of a PDF is the easy half. The half that matters is knowing none were dropped, doubled or misread on the way — because a spreadsheet that is missing one line looks exactly like one that is not. Here are the four ways to convert, in the order worth trying, and the check to run whichever you use.

1. Ask the bank for a download first

Before converting a PDF at all, check whether online banking will give you the transactions as a file. Look in the account's transaction history for "Download", "Export" or a download icon; most banks offer CSV, and many offer OFX or a format named after an accounting app.

This is the best option whenever it exists: it is free, and the numbers come straight from the bank's system, so there is nothing to misread. Its limits are practical. Downloads often reach back only a limited number of months, closed accounts usually lose access altogether, and a statement a client, tenant or lender sent you arrives as a PDF with no download attached. That is when the other three methods come in.

2. Excel's Get Data from PDF

Microsoft 365 Excel on Windows can read tables out of a PDF directly:

  1. Open a blank workbook, go to Data → Get Data → From File → From PDF, and pick the statement.
  2. The Navigator lists every table and page it found. Select the transaction tables — usually one per page — and choose Transform Data.
  3. In Power Query, append the page tables into one, remove the repeated header rows and any "balance brought forward" lines, set the amount columns to a number type, and load the result.

It is free if you have the right Excel, and it reads text-based PDFs accurately. It does nothing with scans or photos, it returns each page as a separate table for you to stitch together, and statements whose columns do not line up cleanly come through with amounts in the wrong column. Budget ten to twenty minutes per new statement layout.

3. Copy and paste

Select the transactions in a PDF viewer, copy, paste into Excel, then use Data → Text to Columns. For a single short page this is fine. For anything longer, it usually goes wrong in one of three ways: every column lands in one cell, a description that wraps onto two lines becomes two rows, or an empty money-in cell collapses so a withdrawal slides into the deposit column. The last one is the dangerous one, because the sheet still looks tidy.

4. A converter that checks itself

A converter does the stitching for you: every page into one table, descriptions joined, amounts in the right columns, scans read from the image. The one worth using is one that tells you whether the result adds up, because otherwise you are trusting it blind.

The bank statement converter here does both. It reads the PDF, scan or photos, then checks the opening balance plus money in minus money out against the closing balance, and every running balance against the row before, in whole cents. The first two pages a day are free, and the result tells you plainly whether the statement reconciles. If you need the transactions somewhere other than Excel, it also writes CSV, QIF and OFX.

The check: two formulas that prove the conversion

Whichever method you used, run this before relying on the sheet. Assume money out is in column C, money in in column D and the running balance in column E, with transactions in rows 2 to 200.

Total check. Put the statement's opening balance in H1 and its closing balance in H2, then:

=H1 + SUM(D2:D200) - SUM(C2:C200) - H2

The answer must be exactly zero. Anything else is the amount you are missing or have extra. If it equals one transaction's amount, that line was dropped or doubled; if it is a round-looking difference like 90 or 0.09, a digit was misread.

Row check. In F3, and filled down:

=ROUND(E2 + D3 - C3 - E3, 2)

Every row should show zero. The first non-zero row is where the problem is — the line itself, or one missing just above it. On statements that print one balance per day rather than per line, only the rows with a balance can be checked this way; the total check still covers everything.

A conversion that passes both checks has every amount right. It can still contain a misspelt description, so skim those, but the money is complete.

Problems you will meet, and the fixes

  • Amounts that will not sum. They are text. Select the column, use Data → Text to Columns → Finish, or =VALUE(). Check for currency symbols and non-breaking spaces first.
  • Negative numbers in brackets, or a trailing minus. "(45.00)" and "45.00-" both mean minus 45. Find and replace, or convert with a formula before summing.
  • CR and DR suffixes. Common on UK and Indian statements: "1,250.00 CR" is a positive balance, "DR" an overdrawn one. Strip the suffix and set the sign.
  • Dates without a year. Statements often print "03 Mar". Add the year from the statement period, and watch December–January statements, which straddle two.
  • Day and month swapped. 03/04 is 3 April in the UK and March 4 in the US. Excel guesses from your computer's region; set the column's format deliberately rather than trusting the guess.
  • Repeated headers and page totals. Every page's header row and any "carried forward" line must go, or they count as transactions.
  • Two-line descriptions. Join the continuation onto the transaction above; a description row with no amount is the giveaway.

Questions

Can I convert a bank statement PDF to Excel for free?
Yes. Your bank's own download is free and exact when it exists. Excel's Get Data from PDF is free if you have Microsoft 365 on Windows. Sonomir's converter does the first two pages a day free, including the balance check.
How do I convert a scanned bank statement to Excel?
Excel's PDF import cannot read scans, because a scan contains an image rather than text. Use a converter that reads images — and then run the total and row checks, since scans are where misread digits happen.
Why does my converted statement not balance?
Usually one of four things: a dropped or duplicated line at a page break, an amount in the wrong column, a misread digit on a scan, or a 'balance carried forward' line counted as a transaction. The row check finds the first place the balance breaks.
Is it safe to upload a bank statement to a converter?
Check what the service keeps. Sonomir deletes the uploaded file when the job ends, keeps only the last four digits of the account number, and never publishes results; the privacy page lists every service that handles the file.

The machines this uses