# Scan Invoice to Excel: Paper Invoices into Rows

- Author: FastBankConvert (https://fastbankconvert.com)
- Published: 2026-10-17
- Category: Other financial documents
- Canonical: https://fastbankconvert.com/blog/scan-invoice-to-excel

> Scan an invoice to Excel from paper: scan settings, OCR on supplier layouts, checking totals against line items, and a log sheet for many invoices.
To scan an invoice to Excel, scan the paper at 300 dpi or more as a PDF, run OCR so the text becomes readable characters, then copy the header fields and each line item into rows. Before you trust the sheet, add up the line amounts and compare them with the printed subtotal, tax and total.

## Scanner settings for a paper invoice

A paper invoice scans best at 300 dpi in grayscale, saved as a PDF, with the page flat and square to the glass. That combination gives text recognition enough detail to tell a 5 from a 6 without making the file huge. Lower resolutions save space but blur small print, and small print is where the unit prices live.

### Resolution and color

Most invoices mix large header text with tiny line-item text, often 7 or 8 point. At 200 dpi that small text loses the gaps between characters, so "1.50" can read as "1 50" or "150". At 300 dpi it stays separate. Go to 400 dpi only when the invoice uses a very small font or a dot-matrix printer.

Color mode matters less than people expect. Grayscale keeps faint pencil marks and light shading readable. Pure black and white makes smaller files, but it can erase a pale blue company stamp or turn a gray table band into a solid black bar that hides the text on it. Use color only when the supplier prints amounts in red or highlights totals, because the color carries meaning.

### Paper handling

Remove staples and flatten folds before scanning. A fold through the middle of a line-item table creates a shadow, and OCR often reads a shadow as a row of dashes or drops the characters it crosses. Thermal-paper receipts and invoices fade over time, so scan them soon after they arrive.

If you use a phone instead of a flatbed, use a scanning app that corrects the angle and crops the page. A photo taken at a slant makes the right side of each row smaller than the left, and columns that should line up drift apart.

### One invoice, one file

Save each invoice as its own PDF, named with the supplier and the invoice number, for example "acme-supply-INV-20417.pdf". One file per invoice makes the log sheet in a later section easy to build, and it lets you go back to the image when a number looks odd.

## Why does OCR misread some supplier invoice layouts?

Optical character recognition, or OCR, turns the picture of each letter into a text character, and supplier invoice layouts trip it up when text sits on shading, columns lack ruled lines, or one description wraps over two lines. The characters may be read correctly while their position in the table is read wrongly, which is harder to spot.

### Layout features that cause trouble

Every supplier designs its own invoice, so no two are alike. These features cause most errors:

- **Shaded or colored table bands.** Alternating gray rows look tidy on paper, but the shading lowers contrast and can swallow decimal points.
- **Columns without lines.** When quantity, unit price and amount are separated only by white space, a long description can push into the next column, and the amount lands under the wrong heading.
- **Wrapped descriptions.** A product name that runs onto a second line may be read as a new row with no quantity or price.
- **Logos and stamps over text.** A "PAID" stamp across the totals block is common on supplier copies, and it can hide the very numbers you need.
- **Handwritten changes.** A crossed-out quantity with a new number written beside it gives OCR two candidates and no way to know which one counts.

### Characters that swap

Some characters look alike in print. A zero and the letter O, the digit 1 and a lowercase l, a 5 and an S, and an 8 and a B are the usual pairs. In a description, a swap is a nuisance. In an amount, it changes the money. Excel will also treat "1O5.00" as text, not a number, which at least makes that error visible when a sum ignores the cell.

### Header fields are often easier than lines

The invoice number, date, supplier name and total usually sit in labeled boxes near the top, so they come through more cleanly than the line table. If you only need those header fields for a payables list, you can skip the line items entirely. When you do need every line, the separate guide to [pulling PDF invoice line items and header fields into Excel](/blog/pdf-invoice-line-items-to-excel) covers digital invoices, where no OCR is needed.

## Checking an invoice total against its line items

Every invoice carries its own proof: the line amounts should add up to the subtotal, the subtotal plus tax and shipping should equal the total, and each line's quantity times unit price should equal its amount. If those three checks pass in Excel, the scanned numbers are almost certainly right. If one fails, the gap tells you where to look.

### A worked example

Here is one scanned invoice checked line by line. The supplier, items and figures are an example, made up for illustration.

| Line | Description | Qty | Unit price | Printed amount | Qty x price |
| --- | --- | --- | --- | --- | --- |
| 1 | Copy paper, A4, box | 4 | 38.50 | 154.00 | 154.00 |
| 2 | Toner cartridge, black | 2 | 89.90 | 179.80 | 179.80 |
| 3 | Desk organizer | 3 | 12.75 | 33.25 | 38.25 |
| | **Line sum** | | | **367.05** | **372.05** |

The printed subtotal on the paper reads 372.05. The printed amounts add to only 367.05, while quantity times price gives 372.05, so line 3 is the suspect. Looking at the scan, the printed amount is 38.25: OCR read the 8 as a 3. One corrected cell, and both sums agree with the subtotal.

### Formulas to add

Add two helper columns beside the extracted lines:

1. In the "Qty x price" column, multiply quantity by unit price, rounded to two decimals.
2. In a check column, subtract the printed amount from that product. Anything other than zero needs a look.

Then, under the table, sum the printed amounts and compare the result with the subtotal you typed from the header. Add tax and shipping and compare again with the total. A difference of a few cents is usually a rounding choice by the supplier. A difference that matches one digit swap, such as 5.00 or 50.00, is usually OCR.

### Discounts and credits

Some invoices print a discount as a negative line, and others take it off the subtotal in the totals block. Decide which you have before you sum, or the check will fail for the wrong reason. Keep discounts as negative numbers in the amount column so the sum still works.

## A log sheet for scanning many invoices into Excel

When you scan invoices into Excel by the dozen, keep one log sheet with one row per invoice, holding the header fields and a link to the scanned file, and keep line items on a second sheet keyed by invoice number. The log answers "what do we owe and to whom", and the line sheet answers "what did we buy".

### Log sheet columns

A layout that works for most small offices, with example values:

| Column | Example value | Notes |
| --- | --- | --- |
| Supplier | Acme Supply Co | Spell it the same way every time |
| Invoice number | INV-20417 | Format as text so leading zeros survive |
| Invoice date | 2026-09-14 | A real date, not text |
| Due date | 2026-10-14 | From the payment terms |
| Subtotal | 372.05 | |
| Tax | 29.76 | |
| Total | 401.81 | |
| Lines check | OK | From the line sheet sum |
| File | acme-supply-INV-20417.pdf | Or a hyperlink to it |
| Paid on | 2026-10-10 | Filled in from the bank statement |

### Keeping it clean

Format the invoice number column as text before you paste. Supplier numbers such as "000814" lose their zeros the moment Excel decides they are numbers. Use a data validation list for the supplier column so the same company does not appear as "Acme", "Acme Supply" and "ACME SUPPLY CO".

Add a duplicate check on supplier plus invoice number. Paper invoices arrive twice more often than you would think, once by post and once as a reminder, and paying both is easy when they sit in different piles.

Paper checks raise the same scanning questions. If a client pays suppliers by check, the guide to [scanning checks into Excel by payee, amount and check number](/blog/scan-checks-into-excel) uses the same scan settings and adds the written-amount check.

## When a scanned invoice should be typed by hand instead

Some invoices are faster to type than to scan and fix. A short invoice with two or three lines, a carbon copy, a heavily handwritten bill from a tradesperson, or a page with a stamp across the totals will take longer to correct after OCR than to enter directly. Typing is the right tool for those pages.

A rough rule: if the OCR output needs more than a handful of corrected cells, the next invoice from the same supplier is probably better typed. Keep a note in the log of which suppliers fall into that group, so the decision is made once.

Also ask suppliers for the digital version. Many invoicing systems can email a PDF that already contains real text, and a digital PDF skips OCR completely. The difference matters most for long invoices with many lines, where every misread character means another check to run.

A mixed pile needs sorting before you start. Separate the digital PDFs, the clean printed paper and the handwritten or damaged pages. Run the first two groups through extraction and the third through a quick manual entry. The same split between summary pages and detailed tables shows up in other documents too; [a brokerage statement in Excel is easier once holdings and activity sit on separate sheets](/blog/brokerage-statement-to-excel), for the same reason the invoice log and line sheet are kept apart.

## Where do scanned invoices meet the bank statement?

Scanned invoices meet the bank statement at the payment: each invoice in the log should match one withdrawal, check or card charge for the same total, made on or after the invoice date. Matching them is how you confirm a supplier was paid once, in full, and that nothing on the statement lacks an invoice behind it.

Fill the "Paid on" column from the statement. Sort the log by total and the statement withdrawals by amount, and the matches mostly line up. A payment that covers several invoices at once will not match a single row, so add a column for a payment reference and give every invoice in that batch the same one. Then the sum of those invoices should equal the one withdrawal.

Getting the statement into Excel is the easier half of this job, because the bank's PDF is usually digital text rather than a scan. FastBankConvert is built for that half: it turns a bank statement PDF into a spreadsheet with dates, descriptions, amounts and balances in their own columns. It is getting ready to open, and [the get started page](/get-started) will let you know when trial conversions begin. For owners weighing which tool fits a business with several accounts, [choosing a bank statement converter for a small business](/blog/bank-statement-converter-for-small-business) compares the options.