A digital bank statement PDF stores its words and numbers as text; a scanned one stores a picture of a page. To tell them apart, try to select a line of text, search for an amount you can see, and check the font list. A digital file passes all three. A scanned file fails, so it needs OCR first.
What makes a bank statement PDF digital
A bank statement PDF is digital when its characters are stored as text, usually because the bank's system wrote the file directly from its records. That stored text is called the text layer. It holds each letter and digit exactly, along with the font and the position on the page, so software can copy it without guessing.
The most common source is online banking. When a client downloads a statement from the bank's website, the PDF was generated by the bank's statement system, not photographed. Every amount on the page exists in the file as the characters 1, 2, 3 and so on, in a named font.
A scanned PDF is different. A scanner or phone captures an image of the paper, and the PDF is a wrapper around that image. To the computer, a scanned "1,234.56" is a pattern of gray pixels. Nothing in the file says it is a number until optical character recognition (OCR) is run on it.
Some PDFs contain both. A scanner or app can run OCR at scan time and add an invisible text layer on top of the image. Canon's manual for MF Toolbox scan settings, for example, describes creating a searchable PDF at 300 dpi so that the characters can be searched and edited as text. That file behaves like a digital one in some tests, but its text came from recognition, not from the bank. The hybrids section below covers how to treat it.
The difference is not about how the file looks. A digital statement and a good scan of the same statement can look identical on screen. The tests in the next section look under the surface.
How do you tell a scanned bank statement PDF from a digital one?
You tell a scanned bank statement PDF from a digital one with three tests that take under a minute: select text, search for an amount printed on the page, and open the font list in the document's properties. A digital file passes all three, an image-only scan fails all three, and a hybrid gives mixed results.
Test 1: select text. Choose the text selection tool and drag across a transaction row. In a digital file, the words highlight one by one, following the letters. In an image-only scan, nothing highlights, or the whole page is selected as one picture.
Test 2: find an amount. Pick an amount you can see on the page, such as a deposit of 2,005.10, and search for it with the reader's find command. A digital file finds it. An image-only scan finds nothing, because there is no text to search. A hybrid may find it, or may miss it if OCR misread a digit.
Test 3: check the fonts. Open the document properties and look at the fonts. In Adobe Acrobat, the Texas Commission on Environmental Quality's guide to embedding fonts in Acrobat describes the route: open the Document Properties window from the File menu or by right-clicking, then the Fonts tab, which lists the fonts under "Fonts Used in this Document." A digital statement lists the bank's fonts. An image-only scan lists none.
Three tests, three kinds of file
The results below are examples for three versions of one fictional March statement for Northgate Supply, checking account XXXX-4821.
| Test | Digital (downloaded from online banking) | Scanned, image only | Hybrid (scan with an OCR text layer) |
|---|---|---|---|
| Select a transaction row | Words highlight along the letters | Nothing highlights, or the whole page selects as one block | Words highlight, sometimes slightly off the printed letters |
| Search for "2,005.10" | Found | Not found | Found if OCR read it right; not found if a digit was misread |
| Fonts tab in document properties | Lists the bank's statement fonts | Empty | Not decisive; check with the copy test below |
For a hybrid, add one more check. Copy a few amounts and paste them into a text editor or a blank worksheet, then compare them with the page. If a pasted 1,284.56 sits next to a printed 1,234.56, the text layer came from OCR and carries its errors.
Why digital bank statement PDFs convert more reliably
Digital bank statement PDFs convert more reliably because the converter copies exact characters instead of recognizing them. Every digit in a digital file is the digit the bank's system wrote. A scan adds a recognition step that can misread look-alike digits or drop a decimal point, and that step is where most conversion errors start.
Think of two routes into Excel.
The digital route. The converter reads the text layer, finds the transaction table from the positions of the words, and writes rows. Characters are never in doubt. What can still go wrong is structure: a description that wraps to a second line, a subtotal read as a transaction, or a row near a page footer left out. A balance check catches those.
The scanned route. The converter, or a separate OCR step, first has to decide what each mark on the image is. Microsoft's guidance on inserting scanned text into Word says recognition accuracy depends on the quality of the scan and the clarity of the text, and that the result should always be proofread. Google's help page on converting PDF and photo files to text adds that tables and columns are not likely to be detected. Then the same structural problems apply on top.
So a scanned statement carries two layers of risk, and a digital one carries one. The specific digit errors OCR makes on statements, and the size of the balance gap each one leaves, are set out in why OCR misreads amounts on bank statements.
The practical rule follows: when a client sends a scan of a statement that is available online, ask for the download instead. It costs the client a minute and removes the recognition step entirely. Keep the scanned route for statements that exist only on paper, such as older periods, closed accounts, or a bank that mails statements.
Is a printed-then-scanned bank statement PDF still digital?
A printed-then-scanned bank statement PDF is not digital. Once a statement is printed on paper and scanned back in, the text layer is gone and the file is an image. Printing to a PDF file on a computer is a different action, and that file can keep its text. The three tests settle it.
The two actions sound alike and produce different files.
Printing on paper, then scanning. A client downloads a digital statement, prints it, signs or annotates it, and scans it to send. The PDF that arrives is a picture of the printout. It fails the selection and search tests, even though it started as a digital file. The original download, if the client still has it, is the better source.
Printing to a PDF file. A "print to PDF" option, where the computer offers one, writes a new PDF instead of sending the page to a printer. That new file may keep the characters as text, because no paper and no camera were involved. It may also be flattened into an image, depending on the software. The tests settle it in seconds.
Scans with an OCR text layer added. A client or an office scanner may run OCR at scan time. The file passes the selection test and often the search test, which makes it look digital. It is still a scan. Its text came from recognition, and any misread digits are baked into the text layer. Treat it as scanned: run the balance check with the same care, and do not trust the copied amounts without it.
Faxed statements belong in the scanned group too: a fax is an image, so run the same tests and expect OCR.
The deciding question is always the same: did the characters in this file come from the bank's system or from recognizing a picture? The three tests answer it.
Password-protected and image-only bank statement PDFs
Password-protected and image-only bank statement PDFs are the two files that stop a conversion before it starts. A password-protected statement cannot be opened or read without the password. An image-only statement opens normally but has no text to read. Each has a different fix, and neither should be forced.
Password-protected statements. When a statement arrives locked, the firm cannot read the file until it is unlocked, and a converter cannot either. Ask the client to send the password through a separate message from the file, or to download the statement again and save an unlocked copy if the bank's site allows it. Note in the client's file how the statement was unlocked, so next month's request can say the same thing.
Image-only statements. A file that fails all three tests has no text layer. If the client has online access, ask for the digital download of the same period. If the paper is the only source, run OCR and plan on the balance check. The scanning settings that give OCR the best page to read are covered in scanning a bank statement to Excel.
Flattened files. A statement that a client signed, marked up or shrank to fit an email has been saved again by some app, and that app may have flattened the page into an image. If a statement that should be digital fails the tests, ask how it was saved and request the original download.
Name and file each statement with its type noted, such as northgate-4821-2026-03-digital.pdf or -scan.pdf, so whoever converts it knows which route to take. How firms name, sort and split client files generally is covered in the sorting and organizing client files category.
Once the route is chosen and the rows are out, the next step is often a CSV. The post on converting a bank statement to CSV covers the save steps and encoding, and the bank statement CSV format explains which columns the file should hold.
Schema CPA takes digital downloads and scans in the same client file, classifies each statement page by document type, bank and period, ties the balances from month to month, and exports the rows to Excel or CSV. The client tracker demo shows how statements sit alongside a client's other documents.