April 2, 2026 · 6 min read

Scanned Bank Statement to Excel: OCR That Actually Works

How to convert a scanned or photographed bank statement to Excel — DPI tips, format gotchas, and when OCR genuinely fails.

Native PDF bank statements (the kind your bank emails you) embed real text — every dollar amount and date is selectable. Scanned statements, in contrast, are just images of paper. To get them into Excel you need OCR (optical character recognition) — and not all OCR is created equal.

Why scanned statements break most converters

A generic PDF-to-Excel tool looks for a text layer first. If there isn't one, it gives up or returns gibberish. A scanned statement has no text layer. So you need a converter that:

  • Detects that the page is image-only and switches to OCR automatically
  • Uses a financial-document model (digits and totals matter more than prose)
  • Reconstructs the table layout from pixel positions, not just raw text order

Our PDF to Excel converter does all three. Drop in a scanned PDF — even a phone photo — and you get the same XLSX output as a native PDF: one transaction per row, ISO dates, separate debit and credit columns.

DPI and image quality

OCR accuracy is mostly about pixels per character. Aim for:

  • 200 DPI or higher for flatbed scans
  • At least 1500 px wide for phone photos
  • Even lighting — no shadows across the page
  • No skew — rotate within ±2° of horizontal

Below 150 DPI, accuracy drops noticeably for small fonts and tightly-spaced columns. If you only have a low-resolution scan, the tool still runs — but expect to spot-check a few rows.

When OCR genuinely fails

  • Carbon-paper or thermal-printed receipts with faded ink — usually unrecoverable
  • Handwritten amendments — no OCR engine will read these reliably
  • Very dense statements photographed at a slight angle, where two columns merge into one

In those cases you'll get most rows correct and a handful flagged for review. Far less work than retyping the whole statement.

Bank-specific tips

The most common scanned formats we see are older Chase, Bank of America, and Wells Fargo statements pulled from filing cabinets — see the Chase, Bank of America, and Wells Fargo walkthroughs for layout-specific notes. International users on HSBC, Barclays, or HDFC Bank get the same OCR pipeline — there's no separate "scanned" plan.

Need a different output?

If your accountant wants CSV instead of XLSX, use the PDF to CSV converter — same OCR pipeline, different output. For QuickBooks, the PDF to QBO converter carries the OCR'd transactions through with stable FITIDs.

Try it on your own statement

Convert a PDF to Excel free — no sign-up required.

Convert a PDF →