When scanned and digital statements go into one batch, both come out as transaction rows, but they get there differently. A digital PDF already holds the bank's text, so it is read directly. A scan is an image, so its characters are recognized first. That extra step is why scanned statements deserve a closer review after the run.
How is a scanned statement read differently from a digital one?
A scanned statement is read through optical character recognition before any field can be found, while a digital statement skips that step. A digital PDF downloaded from online banking stores each character the bank printed as text. A scan stores a picture of a page, and the converter has to decide which shapes are digits, letters and column lines.
After that first step, the two paths join. Both kinds of file go through the same layout reading, the same field mapping and the same balance check. So the difference is not in what comes out, which is the same columns for both. It is in how much confidence you can place in each character before the checks run.
In a mixed batch you do not have to split files by type. AutoBankConvert reads scanned and digital PDFs in the same run and checks every statement's running balance against its totals. The type of file changes what you review afterwards, not how you upload.
Why do scanned statements in a mixed batch need more review?
Scanned statements need more review because recognition errors can produce values that look valid. A blurred 8 read as a 3 is still a number. A faint decimal point that disappears turns 41.50 into 4150. Digital PDFs do not produce these errors, because nothing is being recognized.
Errors on scans fall into three groups, and the balance check catches them unevenly:
- Amount errors usually break the running balance, so a statement with one is flagged rather than passed.
- Date errors can move a line by a day or a month without touching the balance, so they slip through more easily.
- Description errors, such as a zero in place of the letter O, leave the money right but can stop a categorization rule from matching next month.
That is the case for spending more review time on scans even when they pass. A passed balance check tells you the amounts add up. It says less about dates and descriptions.
Telling scanned from digital PDFs before the batch runs
Telling scanned from digital PDFs takes a few seconds per file and saves time in review. Open the file and try to select a word in the transaction list. If the text highlights and you can copy it, the file is digital. If the selection draws a box around the whole page, or copies nothing, it is an image.
Other signs help when you are looking at a folder rather than opening each file:
- File size: a scan of a three-page statement is often several times larger than the bank's own PDF of the same pages.
- Appearance: a slight tilt, a grey background, a staple shadow or a punched hole means paper was involved.
- Source: files that arrived as email attachments from the bank or as downloads from a client's portal are almost always digital. Files from an office scanner or a scanning app are not.
Some files are both. A client may print a digital statement, sign it and scan it back. Treat any file that went through paper as a scan.
Sorting mixed batch results by scanned and digital source
Sorting mixed batch results by source lets you give each file the review it needs and no more. Digital statements that pass the balance check need a light look. Clean scans need a few targeted checks. Poor scans need a line-by-line read or a better copy. The table is an example of how to split the work.
| File type | Balance check passed | What to review | Example time per statement |
|---|---|---|---|
| Digital PDF | Yes | Period and account match the client | 1 minute |
| Clean scan | Yes | Dates inside the period, five descriptions read against the page | 3 minutes |
| Poor scan | Yes | Every date and amount against the page | 10 minutes |
| Any type | No | Find the gap before anything else | Varies |
The times are illustrative, not measured. Adjust them to your own clients. The order is what matters: failed statements first, then scans, then digital files.
If the batch covers a checking account and a card for the same client, the card statement may be digital while the checking account came in on paper. The guide to checking and credit card statements in one batch covers how the two kinds of statement differ beyond file type. Once every statement is reviewed, the steps to export converted statements per client apply the same way to both.
Phone photos that should stay out of a scanned and digital batch
Phone photos are the one input that usually does not belong in a scanned and digital batch, at least not unreviewed. A photo taken at an angle bends the columns, a shadow can hide a row, and a page cut off at the edge loses lines without any sign on the image itself.
Keep a photo out of the batch, and ask for another copy, when any of these is true:
- The page is photographed at a visible angle, so column edges are not straight.
- Part of the page is outside the frame, especially the bottom rows or the balance column.
- Fingers, folds or glare cover any part of the transaction list.
- Several pages were photographed in one shot.
The better copy is almost always available. The bank can provide the PDF through online banking, and the client can download it in less time than it takes to photograph it. Mixed banks bring their own variety on top of this, which the post on bank statements from multiple banks works through.
For a new practice, a decent document scanner and a habit of asking for downloads first are worth more than any later fix, and the list of tools to start a bookkeeping business puts scanning in context with the rest of the setup. To see a mixed batch go through checks and into export files, start from the AutoBankConvert homepage.