Is your bank statement PDF scanned? How to tell, and what to do

A PDF can hold a bank statement in two quite different ways. A text PDF is generated by the bank’s systems: every character is stored as text with its position on the page. A scanned PDF is a photograph of paper: the page is one picture, and the computer does not know there are numbers on it until something reads them. On screen the two can look the same, but only the text PDF can be converted to Excel or CSV exactly.

Why it matters

Tools that convert a statement to a spreadsheet read the text layer. With a text PDF, the dates and amounts are exactly what the bank printed, to the cent, with nothing to guess. With a scan, the characters must first be recognised from pixels, which is called optical character recognition (OCR). OCR is good, but it is not exact, and on a bank statement one wrong digit is one wrong balance.

Steps

1. Open the PDF in a normal viewer. Use your browser, Adobe Acrobat Reader, Preview on a Mac or any PDF reader. Do not use an image viewer.
2. Try to select a line. Drag across a transaction line with the mouse. If individual words highlight, there is text. If a whole rectangle highlights, or nothing does, the page is an image.
3. Search for an amount. Press Ctrl+F or Cmd+F and type an amount you can see on the page, such as 45.20. A hit on that exact spot confirms a usable text layer.
4. Copy and paste a line. Copy one transaction line and paste it into a text editor. Readable text means the layer is good. Gibberish or nothing means it is broken or missing.

How you end up with a scan

• Paper statements that were scanned at home or in an office, or photographed with a phone.
• Statements sent by fax or as image attachments, still common for some loan and mortgage documents.
• Old statements requested from the bank’s archive, which some banks return as image files.
• “Print to PDF” from a screenshot of online banking.
• Documents passed through a scanner app that saved images without running OCR.

The reverse also happens. Scanning apps often add a hidden OCR layer, so a scan can pass the selection test. That layer comes from OCR, so it can still contain misread digits.

What to do if your statement is a scan

1. Get a digital copy from the bank. This is the best fix. Most banks let you download statements as PDFs from online banking, often going back several years. Those are text PDFs. If online banking only goes back a short time, ask the bank for older statements as electronic PDFs rather than paper copies.

2. Download transactions instead of the statement. Many banks also offer a CSV, OFX or QBO download of recent transactions. If that covers the period you need, you can skip the PDF entirely. The guide on banks that export CSV lists which banks say they offer which formats.

3. Run OCR yourself. Several PDF editors, including Adobe Acrobat, can recognise text and add a text layer to a scanned PDF. After OCR, repeat the selection test above. Then check the result line by line against the paper, paying special attention to amounts.

4. Type it in. For a one-page statement with a dozen lines, typing is often quicker than fixing OCR output, and it is exact if you check the total afterwards.

Checking an OCR result

Whatever route you take from a scan, finish with the balance test. Add all money in to the opening balance, subtract all money out, and compare the result with the closing balance printed on the statement. If they match to the cent, no amount is wrong in a way that changes the total. If they do not, the guide on a converted statement that does not balance shows how to find the row, usually in a few minutes.

Watch for the classic OCR swaps: 3 and 8, 1 and 7, 5 and 6, and a decimal point read as a comma or dropped altogether. A dropped decimal point turns 12.50 into 1250, which jumps out in the total.

What Rowkeel does with scanned statements

The Rowkeel converter reads text PDFs only. It does not run OCR. When a PDF has almost no text on any page, the converter reports it as a scan and suggests downloading the statement from online banking, instead of producing an empty or half-filled table. Every row is checked against the statement’s balances, and that check means little on guessed characters.

If your scan already has an OCR layer, the converter reads that layer like any other text, and the balance check shows whether the result adds up. Treat a pass on an OCR’d file with more caution than one on a bank-generated PDF, because two misread amounts can occasionally cancel out.

For text PDFs, the PDF to Excel and PDF to CSV converters work in your browser, without uploading the file.