Why bank statement PDFs come out messy in Excel

Why does a statement that looks tidy convert badly?

Because the tidiness is only visual. A PDF is a set of instructions for drawing letters at positions on a page. It does not say “this is row 14, column Debit”. Your eye sees a table because the words line up; a converter has to rebuild that table from coordinates, and statements are full of layout habits that break a simple guess.

Generic PDF-to-Excel tools treat every page the same way: find lines of text, split them into cells wherever there is a gap. That works for a clean grid.

A statement is not a clean grid. It wraps long descriptions, prints a running balance, repeats its header on every page and puts summary boxes above the transactions. Each of those becomes a broken row.

The seven causes below cover almost every messy conversion we have seen. The example rows are made up; they copy common layouts, not any real customer’s statement.

Descriptions that wrap onto a second line

Long merchant names and transfer references wrap. To the converter, the second line looks like a new row with no date and no amount.

Fix: join the orphan line onto the row above it. A row with no date and no amount is almost always the end of the previous description.

One amount column, or two?

Some banks print separate Debit and Credit columns. Others print one Amount column with a minus sign, a trailing sign (“30.00+”), a bracket (“(30.00)”) or the letters DR and CR. A converter that keeps only the digits loses the direction, so money in and money out look the same.

The running balance sits on another line

Many statements print the balance on the last line of a wrapped transaction, or only once per day. The converter puts that balance against the wrong row, and every balance after it looks shifted by one.

Fix: trust the order of the rows, not the cell the balance landed in. Each balance should equal the previous balance plus or minus the row’s amount.

Summary boxes read as transactions

The first page usually has a box with “Opening balance”, “Total money in”, “Total money out” and “Closing balance”. Those are totals, not transactions. Pulled into the sheet, they double-count everything.

Page headers repeated on every page

The bank’s name, your account number, “Page 2 of 6” and the column headings print again at the top of each page. A converter copies them into the middle of your data, sometimes with a date that is the statement date, not a transaction date.

A date inside the description

Card payments often print the purchase date inside the description, while the row’s own date is the day it cleared. Some converters read the wrong one as the transaction date, so rows move days or even months.

Numbers split, dates without a year

Narrow columns split a long figure over two lines (“200,000,000.0” and “0” underneath), so it reads as two numbers. On a scanned page every digit has to be read with OCR first, which adds its own errors, so check whether your PDF is scanned or digital. And most statements print dates as “03/04” with no year, which is either 3 April or 4 March depending on the bank’s country. Excel guesses, and the guess can be wrong for half the rows.

Fix: find one date with a day above 12 (such as 25/03): it proves the order for the whole statement. Take the year from the statement period printed at the top.

What does a statement-specific converter do differently?

It reads the page as a statement, not as a generic table. A dated line with an amount starts a new transaction, and the lines under it without a date are joined to its description. Summary boxes and page headers are recognised and kept out of the rows.

The direction of each amount comes from its column or its sign. Then the arithmetic is checked: each running balance against the one before it, and the opening balance plus money in minus money out against the closing balance.

FastOCR’s bank statement converter works this way and shows whether the balances add up as Reconciled, Not reconciled or Could not be checked. It does not check dates or descriptions, so skim those against the PDF. For scanned or photographed statements, read what to expect from OCR on bank statements.

How do I check any conversion, from any tool?

Add up the money in and money out in your sheet and test one formula: opening balance + money in − money out = closing balance. If it holds, no row is missing or flipped. If it does not, the size of the gap usually tells you what went wrong. The 2-minute balance check walks through it with a formula you can paste.

When is converting the wrong answer?

If your bank’s website offers a CSV, OFX or QuickBooks download for the dates you need, use that instead. It is the bank’s own data and needs no checking for layout. Converting is for closed accounts, old statements beyond the download window, and banks that only give you a PDF. If the rows are headed for QuickBooks, our guide to importing a statement PDF into QuickBooks Online covers the upload.