In February 2003, the British government posted a briefing document on Iraq to the 10 Downing Street website as a Word file. Researchers soon found that the file’s hidden revision log listed the usernames of the people who had edited it and the file paths it had passed through. Four names appeared, along with a reference to a government communications unit. The episode became a standard example of how much a document can say about where it came from.
Two decades later, the formats have changed and the problem hasn’t. PDFs and Office files still carry information the sender never sees on screen. For a law firm, accounting practice or clinic, that hidden data can include client names from an earlier matter, internal comments, draft figures and the structure of your file server.
What metadata is#
Metadata is information about a file stored inside the file. Some of it is useful: a title and author help search tools and screen readers. The trouble comes from the fields that fill themselves in without anyone noticing, and from content that stays inside the file after you think you’ve removed it.
The standard PDF properties#
Every PDF can carry a set of document properties:
| Field | What it often contains |
|---|---|
| Title | The title of the source document, which is frequently the name of the template or an earlier client’s document |
| Author | The username or full name from the computer that created the source file |
| Subject and keywords | Whatever the author or template put there |
| Creator | The application that produced the original, such as a word processor and its version |
| Producer | The software that wrote the PDF |
| Created and modified dates | Timestamps that show when the work was done |
The Title field causes more embarrassment than any other. A firm reuses last year’s engagement letter for a new client, exports it to PDF, and the PDF’s title still names the old client.
PDFs can also carry an XMP metadata stream, an XML block that duplicates and extends these properties. Cleaning the properties panel without cleaning XMP leaves the same information in a second place.
What else hides inside a PDF#
Metadata fields are the start. These are the other places sensitive information turns up.
Earlier revisions#
PDF supports incremental saving: when you edit and save, the software can append the changes to the end of the file instead of rewriting it. The earlier version stays inside, and tools can recover it. A form filled in, corrected and saved can still contain the first answers.
Comments and annotations#
Sticky notes, highlights and markup are annotations. Some viewers hide them or show them only in a side panel, so a reviewer’s note like “client wants to hide this from the IRS?” can travel with the file unseen.
Embedded files and attachments#
PDFs can hold attachments: spreadsheets, emails, other PDFs. Portfolio PDFs and some scanned packets bundle files this way.
Hidden layers and form data#
Optional content layers can be switched off in the viewer and still exist in the file. Form fields keep their values, including fields that were later hidden.
Text under black boxes#
A black rectangle drawn over text doesn’t remove the text. Anyone can select it, copy it and paste it into another document. This failure is common enough that it has its own guide: how to properly redact a PDF.
Scripts#
PDFs can contain JavaScript for forms and actions. A document from outside the firm that runs scripts is a security question as well as a privacy one.
What hides in Word, Excel and PowerPoint#
Most PDFs start life as Office files, and Office files carry even more.
- Author and “last modified by” names, the company name and the template file name
- Tracked changes, including deleted text that is still stored in the file
- Comments, with the commenter’s name and the date
- Hidden text formatted to be invisible
- In Excel, hidden rows, columns and entire hidden sheets, cell comments, and links to other workbooks that include the full server path
- In PowerPoint, speaker notes and slides marked as hidden
- Embedded objects, such as a whole spreadsheet pasted into a Word table
File paths deserve their own warning. A link from a spreadsheet to another workbook, or an image inserted from a network drive, can store a path like \\fileserver\Clients\Okafor-Divorce\Settlement figures.xlsx. That one line tells a reader your server name, your folder structure and the name of another client.
Why this matters for a professional firm#
Lawyers, accountants and clinicians have a duty to protect client information, and hidden data is information. Sending a file that names a different client or exposes an internal comment can breach confidentiality as surely as sending the wrong attachment. In negotiation and litigation, metadata can reveal drafting history, which terms were changed and who changed them.
There is a security cost too. Usernames, server names and folder paths are useful to an attacker planning a phishing email or looking for a way into your network.
How to check a document#
Look before you send. A few ways to see what a file carries:
- In Adobe Acrobat or Reader, open File, then Properties, and read the Description tab. Acrobat Pro’s sanitize and “remove hidden information” tools also list hidden content before removing it.
- In Word, Excel or PowerPoint on Windows, go to File, Info, Check for Issues, then Inspect Document. The Document Inspector lists comments, tracked changes, properties, hidden text and more.
- On a Mac, Preview’s inspector shows basic PDF properties.
- For a fuller view, upload the file to a checker that reports every category: author and usernames, server paths, tracked changes, earlier PDF revisions, hidden sheets, text under black boxes, comments, and macros or scripts.
How to clean a document#
The order matters. Clean the source file first, then the PDF.
Step 1: Clean the Office file#
Accept or reject every tracked change, delete comments you don’t intend to share, unhide and review hidden sheets, rows and slides, and break links to other files. Then run the Document Inspector and remove document properties and personal information. Doing this by hand in Office is safer than an automatic tool, because you decide what goes and nothing you need disappears.
Step 2: Export to PDF#
Use your word processor’s export or save-as-PDF function. Check the export options for settings that include comments, markup or document properties, and turn off what you don’t want.
Step 3: Sanitize the PDF#
Use a tool that removes the document properties, the XMP metadata, earlier revisions, embedded files, scripts and other hidden content, then saves a fresh file. In Acrobat Pro, the sanitize function does this. Document Security Check’s cleaning step removes document properties, XMP data, earlier PDF revisions, embedded files, scripts and printer settings. It deliberately leaves comments, tracked changes and hidden sheets for you to fix in Office, so nothing you need is lost.
Step 4: Check the result#
Open the cleaned file and check the properties again. Select all the text, copy it and paste it into a plain text editor. If anything appears that you didn’t expect, go back a step.
Preservation comes first
If a document is subject to a litigation hold, a regulatory request or a production obligation, don’t strip its metadata. Clean a copy for sending, and keep the original as it is. When in doubt, ask the responsible attorney before you remove anything.
A sending routine for the whole firm#
Metadata problems are habits, so the fix is a habit too. Make these steps part of how every document leaves the firm:
- Start new documents from a clean firm template, never from another client’s file
- Before export, accept or reject tracked changes and remove comments
- Unhide and review hidden sheets, rows, columns and slides
- Break links to other files
- Run the Document Inspector in Office
- Export to PDF with markup and properties turned off
- Sanitize the PDF and save it as a new file
- Check properties and do a copy-and-paste test on the final file
- Keep the uncleaned original where your retention rules require it
The templates step prevents more trouble than any other. Most leaked client names come from documents that began life as someone else’s file.
Questions people ask
Does converting a Word file to PDF remove its metadata?
Only some of it. Comments and tracked changes may or may not carry over depending on your export settings, and the author name, title and application details usually copy straight into the PDF's own properties. Check the PDF itself after you export it.
Does printing to PDF clean a document?
Printing to PDF drops some of the original file's structure, but the new PDF gets fresh properties, often including your system username, and text under a drawn black box usually stays selectable. Use a proper sanitize or clean step instead.
Can metadata be used as evidence?
Yes. In litigation, metadata often shows who created or edited a file and when, and courts can treat it as part of the document. Removing metadata from files you are obliged to preserve or produce can cause serious problems, so follow your preservation and production obligations first.
Is it safe to upload a client document to an online metadata checker?
Check the service's handling terms first. Document Security Check processes each file in memory, never stores the file, and deletes findings automatically after 24 hours.
Sources
Keep reading
This guide is general information for small firms. It is not legal, insurance or compliance advice for your situation.