Blog › Published 2026-06-19

A PDF workflow for accountants: from shoebox receipts to clean spreadsheets

Month-end for an accounting practice looks the same everywhere: forty emails with attachments named “scan0001”, photos of receipts taken on car dashboards, bank statements exported from five different portals, and a client who swears the missing invoice was “definitely sent”. You cannot control how documents arrive. You can control what happens in the first five minutes after they do — and that is where a PDF routine earns its keep.

One folder, one naming rule

Before any tool, a convention: every incoming document is renamed on arrival to a fixed pattern — year-month, client, type, amount, as in 2026-06_rossi_receipt_84.50.pdf. Alphabetical order becomes chronological order, a missing month is visible as a gap rather than a surprise, and the rule survives holidays, new hires and software migrations. It is the cheapest piece of infrastructure a practice will ever deploy.

Make every receipt searchable on arrival

A photographed receipt is a picture; the amounts in it are invisible to search. A pass of OCR turns the picture into text you can query, so that finding a parking charge from last February is a two-second search instead of an archaeology dig through folders. Do it on arrival, not at year-end: OCR-ing one receipt is a reflex, OCR-ing three hundred is a project.

Bank statements into Excel without retyping

Statement PDFs contain perfect tables that spreadsheets cannot see. Instead of retyping — or worse, copy-pasting columns that shift and merge — use the PDF to Excel converter to extract the transaction table into real rows and columns, with numbers as numbers. Then apply the thirty-second control every bookkeeper should trust: sum the extracted amount column and compare it with the statement’s closing balance movement. If they match, the extraction is sound; if not, you know before the error propagates into the ledger.

The month as a single, numbered file

Once the month is complete, merge the receipts and statements into one bundle in date order and add page numbers. The payoff comes later, in every conversation about the file: “the charge on page 12” is an instruction, “the receipt from that Tuesday, I think it was a PDF” is a treasure hunt. One file per client per month also makes handovers between colleagues trivial.

Send lighter files

Scanned bundles are heavy, and mail servers still bounce attachments over 20–25 MB. A pass of compression typically cuts a scanned bundle by half or more with no visible difference, which is the difference between “sent” and “the client never received it and the deadline is today”. Compress the copy you send; keep the original in the archive.

Ten years is the horizon

Invoices and accounting records commonly must be kept for about a decade, and a decade is long enough for fonts to go missing and software to change hands. Converting closed years to PDF/A embeds everything the file needs to render identically in 2036 — which is precisely what you want to hand an auditor, a judge, or your own successor. Archive the PDF/A version, and treat it as read-only from that day on.

“The client sent sixty photos of receipts”

It happens every quarter: a shoebox emptied onto a kitchen table, photographed in bursts, delivered as a cascade of JPGs. Do not file photos — they sort by camera timestamp, preview badly and get lost among holiday pictures. Convert the batch to a single PDF in one pass, run OCR so the amounts become searchable, and rename it by month. Sixty loose images become one queryable document, and the kitchen-table delivery format stops mattering — which is the point, because the client will do it again next quarter.

“The invoice PDF won’t import into the accounting software”

Two distinct failures hide behind this symptom. If the software reports the file as unreadable or damaged, a pass of repair usually rebuilds the internal structure — export-and-email chains are rough on files. If the file opens fine but the import finds no data, you are looking at a scanned invoice with no text layer: run OCR first, then import. Diagnosing which of the two you have takes one try each — still faster than asking the supplier to reissue.

When the auditor asks for Q3 only

Requests are rarely for everything; they are for a quarter, a supplier, a single dispute. Split the relevant page ranges out of the annual bundle instead of sending twelve months of a client’s financial life to someone who needs three. It is faster for you, safer for the client, and it signals a practice that knows where everything is — because, with this routine, it does.

Counted honestly, the whole routine costs perhaps an hour per client per month — five minutes at a time, mostly on arrival of each document. What it buys is the thing accounting practices actually sell: the confidence that any number can be traced to its paper in under a minute, in front of the client, the auditor or the tax office. Firms bill for that confidence. The routine is how you manufacture it.