You scan four pages and end up with a 38 MB PDF that your email client refuses to send. The document is four sheets of black text on white paper. Something has clearly gone wrong, and it is worth knowing what.
Why PDFs get big
A PDF is a container. It can hold text stored as actual characters, images, embedded fonts, vector graphics, and various metadata. Which of those it contains makes an enormous difference to the size.
Scanned pages are pictures, not text
This is the usual culprit. When you scan a document, the scanner does not read the words — it takes a photograph of each page. A PDF of ten scanned pages is ten large images in a wrapper. At 600 DPI in colour, each page can run to several megabytes on its own.
A genuinely text-based PDF of the same document, where the letters are stored as characters, might be 100 KB for the whole thing. The difference is routinely a factor of a hundred.
Embedded images at full camera resolution
Drop a phone photo into a document and export to PDF and the full 4000 × 3000 image usually goes in at native size, even though it is displayed at postcard scale on the page.
Embedded fonts
PDFs embed fonts so the document looks identical everywhere. Usually only the characters actually used are included, which is small — but a document using many fonts, or fonts embedded in full, can add several megabytes.
Know your limit before you start
| Service | Attachment limit |
|---|---|
| Gmail | 25 MB |
| Outlook / Microsoft 365 | 20–25 MB (often lower on company servers) |
| Yahoo Mail | 25 MB |
| Many corporate mail servers | 10 MB |
| Government / legal upload portals | Frequently 5 MB |
What to try, in order
1. Compress the PDF
Always start here. Compression re-encodes the images inside the document at a more sensible resolution and quality. On a scan-heavy PDF this alone often cuts 60–90% of the size, and for a document that exists to be read on screen the visible difference is usually negligible.
2. Send only the pages you need
Frequently overlooked, and often the biggest single win. If the recipient needs pages 3 to 7 of a 60-page report, splitting out those five pages reduces the file by more than any compression setting could. It is also more considerate.
3. Reconsider the source
If you created the PDF yourself, going back is often better than compressing after the fact. Resize images before placing them; export at "screen" or "smallest file size" rather than "press quality"; and if you scanned it, rescan at 200–300 DPI in greyscale rather than 600 DPI in colour. A text document gains nothing from colour scanning.
4. Split it into parts
When the content genuinely cannot be reduced further, split the document and send it in two or three messages. Ugly, but it works when a portal will not budge.
5. Share a link instead
Above roughly 20 MB, email stops being the right tool. Upload to cloud storage and send a link — it avoids limits entirely and lets you revoke access later.
Scanning settings that prevent the problem
- Resolution: 200–300 DPI is plenty for a document meant to be read. 600 DPI is for archiving photographs and quadruples the data.
- Colour mode: greyscale for anything printed in black ink; colour only when colour carries meaning. This alone can cut the size by two-thirds.
- Black and white / bilevel: for plain printed text with no photographs, this produces dramatically smaller files, though it looks harsh on anything with shading.
How far can you go before it looks bad?
Compression on a scanned PDF mostly reduces the resolution and quality of the page images. Moderate compression on a 300 DPI scan is essentially invisible on screen. Aggressive compression starts to show as soft or slightly speckled text, most noticeably on small print.
The test that matters is simple: open the compressed file and read the smallest text in it at 100% zoom. If it is comfortable, you are fine. If you are squinting, step back one level.
Scanned PDFs and searchable text
One consequence of scans being images: you cannot search them. Ctrl+F finds nothing, because there is no text in the file, only pictures of text. OCR solves this by recognising the characters and adding them as an invisible layer behind the scan — the page looks identical but becomes searchable and selectable. Worth doing before you archive anything you will need to find again.
Common questions
Why is my PDF huge when it's only a few pages?
Almost certainly a scan, or full-resolution photos placed inside it. Page count is a poor predictor of size — what those pages contain is everything.
Does compressing a PDF lose quality?
Images inside it are re-encoded, so yes, in the same way a JPG loses quality. Text stored as real text is unaffected. For a document meant to be read on screen, moderate compression is normally invisible.
Can I compress a PDF that's already been compressed?
You can, but returns fall away quickly and quality degrades faster the second time. If one pass has not got you under the limit, splitting the document or sending a link is a better move than compressing repeatedly.