How to make a PDF small enough to email

Most mail providers cap attachments at around 20 to 25 MB. Gmail and Outlook both sit in that range, and corporate mail servers are often stricter. A scanned document can pass that easily, and a presentation exported to PDF frequently does.

Before compressing anything, it is worth knowing what is actually large, because the right fix is different in each case and the wrong one either does nothing or ruins the document.

What makes a PDF big

• Images, almost always. A scan is one large photograph per page. At 600 dpi in colour, a single A4 page can exceed 10 MB on its own.
• Embedded fonts, when a document embeds several complete families rather than subsets. Usually a few megabytes at most.
• Repeated content, such as a logo stored separately on every page instead of once.
• Revision history. Some editors append changes rather than rewriting, so the file carries versions you cannot see.

A text-only document is rarely large. If a mostly-text PDF is 30 MB, something structural is wrong with it rather than its content being heavy.

Two different kinds of compression

Structural cleanup is lossless. It removes duplicated objects, discards unreferenced data, subsets fonts down to the characters actually used, and compresses the internal streams. Nothing visible changes at all. On a bloated file this alone can be dramatic; on a well-made one it does very little.

Image downsampling is lossy. It reduces the resolution of the pictures inside the document. This is where the real savings are on a scan, and it is also where quality goes if you push too far.

The compress tool tries both and returns whichever result is actually smaller. That sounds obvious but it matters: on some files aggressive downsampling produces a *larger* file than lossless cleanup, because re-encoding an already-compressed image can add data rather than remove it. Returning the bigger one would be worse than doing nothing.

Choosing a level

• Light: lossless cleanup only. Nothing visible changes. Try this first.
• Medium: moderate downsampling. Good for scans you need to email but still read comfortably.
• Strong: aggressive downsampling. Fine for a receipt, poor for anything where fine detail or small print matters.

Compression is not reversible. Quality removed from an image cannot be restored, so keep the original if the document has any archival value.

Check what you got back

Compression is one of the few operations where the result deserves a look before you send it. Open the compressed file and check the pages that matter: small print, signatures, fine lines in tables and anything with a stamp or seal. Those degrade first and they are exactly the parts a recipient is most likely to need.

Text that was real text stays sharp at any compression level, because only images are downsampled. If the whole page looks soft, the page was an image to begin with, which means the document is a scan.

When compression is the wrong tool

If your file is large because it has 400 pages, compressing it will not help much, and the answer is to split it and send the part that is actually needed. If it is large because of one enormous embedded image on a single page, extracting or removing that page solves it directly.

And sometimes the honest answer is not to attach it. A shared link avoids the size limit entirely and lets you replace the document later without sending a correction. Most people reach for compression first because it feels like the technical solution, when the practical one is simply not putting a 40 MB file in an email.

And if the document was scanned at a much higher resolution than anyone needs, the underlying problem is the scanner setting. Scanning at 200 to 300 dpi rather than 600 produces a file a quarter of the size, and for ordinary printed documents nobody can tell the difference.