All posts

The accountant's guide: how to turn bulk PDF statements into clean Excel sheets in seconds

5 November 2026

Year end. A client sends a folder: twelve monthly statements, two of them scans, one PDF with three months merged into it, and a statement for a different account that slipped in by mistake. Here is a routine that turns that folder into clean, checked Excel, one workbook per account. With soft copies it takes seconds. Scanned pages take longer, and Parserje says how long before it starts.

1. Sort the folder by account first

One workbook covers one account, meaning one bank and one account number. That is what makes the checks mean something: a running balance only runs within a single account. So before converting anything, group the PDFs by account, and name them so they sort, for example Client Maybank 2026-01.pdf.

If one slips through, Parserje catches it. A file from another bank or another account is left out of the workbook, and the result says which file and why. A single PDF that holds several accounts has to be taken out and done on its own.

2. Drop the whole account in at once

Select every PDF for the account and drop them onto the page together. A merged file is split statement by statement, and each month is read on its own: soft copies from their text layer, scans by OCR, and a file that mixes the two is fine. If any pages need scanning, Parserje tells you how many and roughly how long they will take, and asks before it starts.

3. Let it remove the overlaps

Client folders overlap. A month gets downloaded twice, or a branch printout covers the same dates as an e-statement. Rows that appear in both are kept once. The removed rows go to their own sheet, Duplicates Removed, with the file each came from and the row it duplicated, and a note at the top of the main sheet points to it. Nothing goes missing without a trace.

4. Read the result before you download

The result shows the account, the closing balance and whether the figures tie out against the statements. Look at two things: that the closing balance matches the ending balance printed on the last statement, and any rows Parserje has flagged. A flag means it was unsure, not that it failed. A glance at the PDF settles it.

5. Download one workbook

You get one .xlsx file for the account. Inside it:

• All: every transaction, with Date, Description, Type, Amount, Balance and Check columns. Debits are negative, so the Amount column sums to the net movement. The account and the closing balance are stated above the table.
• Debit and Credit: the same rows split by direction.
• Monthly sheets, if you add them: a Debit and a Credit sheet for each month, for a month by month reconciliation.
• Summary, if you create one: it opens as the first sheet.
• Duplicates Removed, when any statements overlapped.

6. File it next to the PDFs

Save the workbook in the same folder as the statements it came from, with a matching name, so that in six months the answer to "where did this figure come from" is one click away. Leave that file as it is. When the rows are headed for accounting software, work on a copy, as described in using the file with AutoCount, Bukku and Excel.

Then the next account

Three accounts means three runs and three workbooks. It takes longer to say than to do, and it is the only way the checks can tell you each account is whole. For why those checks matter so much, read why manual data entry ruins reconciliations. If the folder is mostly scans, read how a scanned Maybank statement is read. And for the basics, start with how to convert a bank statement PDF to Excel.