Skip to content

Work & documents

What your files say about you before anyone reads them

Photos carry the coordinates of where they were taken. PDFs carry the name of whoever made them. Neither is visible, and both are trivial to remove.

QuickFileToolkit Team6 min read

A photograph is not only a picture. A PDF is not only its pages. Both formats carry structured fields that travel with the file, describe how it was made, and are shown to nobody by default.

Most of the time this is harmless and occasionally useful. But it is worth knowing what is in there, because the moments when it matters tend to be the moments you find out the hard way.

The short version

  • Photos can carry GPS coordinates, the exact date and time, the device, and the settings.
  • PDFs carry an author name, the software that made them, and creation and modification timestamps.
  • Office documents are the worst of the three for this, and PDF export does not always clear it.
  • Nothing on screen tells you any of it is there.
  • Removing it takes seconds and does not degrade the file.

What a photo is carrying

The container is called EXIF, and a photo from a modern phone typically holds:

Location. Latitude and longitude, often with altitude, accurate to a few metres. If location services were on for the camera, the photo knows where it was taken. This is the field that matters most and the one people are least aware of.

Timestamps. The exact second the shutter fired, usually in local time.

Device. Make and model, sometimes the operating system version. Enough to identify which of your phones took it.

Camera settings. Exposure, aperture, ISO, focal length, whether the flash fired. Genuinely useful to photographers and harmless to everyone.

An embedded thumbnail. A small copy of the image. Historically this occasionally survived edits to the main image, so a cropped photo could ship a thumbnail of the uncropped original.

Sometimes more. Editing software adds its own fields, and some cameras write a serial number.

The awkward case is not usually one photo. It is a set. Sell a bicycle online with five photos and the buyer has your street. Post a picture from your desk and the coordinates identify your office. None of that requires any skill to read — the file properties panel on both Windows and macOS will show it.

What a PDF is carrying

Less dramatic than coordinates, more likely to be professionally awkward.

Author. Very often the account name of whoever created the file. Sometimes a full name, sometimes a corporate username.

Producer and creator. The software and version. This quietly announces your toolchain, and occasionally that a document was made by exporting a template.

Title. Frequently left as the original filename, which can be more revealing than the document. offer_v4_final_revised.pdf says something the offer letter does not.

Creation and modification dates. Including the case where a document dated last month was modified this morning.

Keywords and subject, if anyone filled them in.

The practical risk is context leakage. A proposal whose author field carries a colleague’s name, or whose title still reads pricing_template, tells the recipient something you did not intend to send. Nobody is hacking anything — they clicked File, Properties.

The one that surprises people: office documents

Word, Excel and their equivalents keep considerably more than a PDF does — author, last modified by, total editing time, and often tracked changes and comments that are not currently displayed.

Exporting to PDF removes the change history, which is the main reason it is good practice. But it does not reliably clear the author fields, and depending on the exporter it can carry the original document’s metadata straight through. A PDF is a safer thing to send than a .docx, not a guaranteed clean one.

When it actually matters

Being proportionate about this — most files, most days, it does not matter at all. It matters here:

  • Selling or renting anything online. Photos taken at home carry your home.
  • Anything involving safety. If there is a reason someone should not know where you are, photographs are the leak, not the message.
  • Journalism, research, whistleblowing. Source protection fails on metadata more often than on content.
  • Job applications and proposals. Author fields and stale titles say things about how a document was produced.
  • Legal and regulated work. Timestamps and author names can become evidence about a process, separately from the document’s contents.
  • Anything a client’s data passes through. Some duties of confidentiality cover the file, not just the words in it.

What to do about it

For a single file, strip it before you send it. Seconds, no quality cost.

For photos specifically, note that resizing or compressing an image usually drops the metadata as a side effect, because the file is being rebuilt. That is convenient, but it is a side effect rather than a guarantee — if the point is privacy, remove it deliberately and confirm it is gone.

For a batch, do it in one pass rather than file by file. A set of photos is exactly the case where one missed file undoes the effort on the other nine.

Turn location off at the source if you never want it. Both major phone platforms let you deny location access to the camera specifically, which solves the problem permanently for photos you have not taken yet — and does nothing for the thousands already in your library.

The thing that is not metadata

One important distinction, because conflating them is genuinely dangerous.

Metadata is data about the file. Removing it is easy and complete.

Redaction is removing content from inside the document, and it is a different problem. Drawing a black box over a paragraph in most PDF editors adds a black rectangle on top of the text. The text is still there. It can be selected, copied, and extracted with no special tools — this has produced a long history of embarrassing disclosures from organisations who should have known better.

If a document needs redacting, the text has to be deleted, not covered. And whatever you use, verify afterwards the same way anyone else would: try to select the hidden text and see what you get.

The short checklist

Before a file leaves:

  1. Photo? Check for location. Strip it unless you need it.
  2. PDF? Check the author and title fields.
  3. Came from a Word document? Check again — export does not always clear it.
  4. Sending several? Do the whole set, not the ones you remembered.
  5. Redacted anything? Try to select the covered text before you trust it.

Frequently asked questions

Do social networks already strip this for me?

The large platforms generally strip location data from photos when they re-encode uploads, so a picture posted publicly is usually safe on that count. What they do not cover is everything else — email attachments, messaging apps that send files as documents rather than photos, cloud folder links, and anything you upload to a portal. Those deliver the original file intact.

Does removing metadata damage the file?

No. For a JPEG the metadata sits in separate blocks alongside the image data, so it can be removed without touching or re-encoding a single pixel. For a PDF the document information fields are similarly separate from the page content. The file gets slightly smaller and looks identical.

Can metadata be recovered after it is removed?

Not from the file you send — the fields are gone. But your original still has them, and any copy you sent earlier still has them. Removing metadata protects the next recipient, not the previous one.

Is there metadata I should keep?

Sometimes. Colour profile information affects how the image is displayed and is worth keeping. Orientation data tells the viewer which way up the photo goes, and stripping it carelessly can leave a photo sideways. Copyright and author fields may be exactly what you want present on work you are publishing.

What about redacting text in a PDF?

That is a different and more dangerous problem. Drawing a black rectangle over text in most editors only covers it visually — the text is still in the file and can be selected, copied or extracted. Real redaction has to delete the underlying content. If a document is sensitive enough to redact, verify by trying to select the hidden text afterwards.