Name Your Files So Automation Rules Can Find Them

A bakery owner I worked with last spring had set up what should have been a simple automation. Every supplier invoice saved to a shared folder would be copied to her bookkeeper's folder and logged in a spreadsheet. It ran for three weeks. Then the bookkeeper called asking why only about half of March's invoices had arrived.

The automation was fine. The file names were not. The rule looked for files with "invoice" in the name, and the folder held things like "scan0047.pdf", "Flour March.pdf", "INV_final(2).pdf" and "Doc - 14 Mar 2026.pdf". A rule cannot match what nobody wrote down. Most small business automations that fail quietly fail this way, long before anyone gets to the clever part.

Why automations depend on names more than you think

Almost every common automation tool decides what to do with a file by looking at its name, its folder, or both. Zapier and Make filter on text in the filename. Power Automate conditions check whether a name contains or starts with something. Hazel on a Mac, and Dropbox and Google Drive rules, sort files by matching patterns in the name.

Names are cheap for software to read. The content of a PDF is not. To look inside, a tool needs to run text recognition, which costs time, often costs money per page on paid plans, and gets things wrong on scanned receipts with coffee rings on them.

So a naming habit is the foundation that every later automation sits on. When names follow a predictable pattern, a rule can be written in a minute. When they do not, people either give up on automating or build fragile workarounds that break the first time someone saves a file slightly differently.

A naming pattern that works for most small teams

After trying a few variations with clients, I now suggest the same structure almost every time:

date_client-or-supplier_type_reference.pdf

For example: 2026-10-14_harlow-mills_invoice_INV-1182.pdf. Each part has a reason.

• The date comes first, written year, month, day. This format, from the ISO 8601 standard, sorts correctly in every folder view. "14-10-2026" sorts by day, which mixes months together. "Oct 14" sorts alphabetically, which puts April before January.
• Underscores separate the parts, hyphens join words within a part. That way a rule can split the name into fields reliably, and "harlow-mills" stays one unit.
• Lowercase throughout, except the reference number if it came that way. Some systems treat "Invoice" and "invoice" as different. Lowercase removes the question.
• No spaces. Spaces turn into "%20" in web links and cause odd failures in some tools.
• The type comes from a short fixed list. More on that below, because it is where most of the value lives.

Keep it under about 60 characters if you can. Long names get cut off in folder views and phone apps, and people stop reading them.

The fixed vocabulary

The type field only works if everyone uses the same words. For a typical small business I suggest no more than eight, written on a sticky note by every computer: invoice, receipt, quote, contract, statement, payslip, letter, other. That is it.

The bakery owner's list ended up with seven. Once "invoice" meant invoice, and nobody wrote "bill" or "inv" or "supplier-doc", her rule caught every file. She went from about 51 percent of invoices reaching the bookkeeper to all of them, the next month, without touching the automation itself.

Smart tools will not fix messy names

A common reply I hear is that AI tools can now read documents and work out what they are, so names do not matter anymore. Some tools can. I have tested several, and they are getting better.

But they are slower, they often cost money per document, and they make occasional confident mistakes, like filing a supplier credit note as an invoice because the word appears in the header. For a bookkeeping folder, an occasional mistake is a real problem. The human in the loop catches it, but that is a person checking a machine's guess about something a two second filename could have stated plainly. I wrote about why that human check still matters, and names are part of how you keep the checking light.

A good file name is an instruction you write once and every tool can follow for free.

Use document reading tools for the stuff you cannot control, like paperwork other people send you. For anything you save yourself, name it properly at the moment you save it.

Getting the habit to stick

Naming rules die when they live in a document nobody reads. They survive when saving a file the right way is easier than saving it the wrong way. A few things that help:

• Set your scanner app's default name to start with today's date in the right format. Most apps allow this, including Microsoft Lens, Adobe Scan and the scanner in the iPhone Files app.
• Pin the naming pattern and the vocabulary list to the top of the shared folder as a text file called 0_how-to-name-files.txt. The zero keeps it at the top.
• Rename anything that arrives from email before you file it. It takes five seconds and saves a search later.
• Do not rename everything from the past. Start from today, and rename older files only when you open them anyway.

Pick one folder to start, the one an automation already touches or the one you want to automate next. Spreading the habit across the whole business at once is how it gets abandoned, which is the same reason I recommend teams automate one task at a time. Habits and automations both stick better in small doses.

What to do before Friday

Choose the one folder that matters most for your money, usually supplier invoices or receipts. Write your eight word type list and the naming pattern on a note. Rename the files from the last two weeks, which for most small businesses takes under twenty minutes, then save every new file that way until the end of the month. By then you will have a folder that any rule, in any tool, can sort without guessing. For more small habits like this one, browse the tips and habits section.