You've got a PDF to send and the email won't take it. Gmail allows attachments up to 25 MB, and Outlook.com stops at 20 MB. Go over that in Gmail and it quietly turns the file into a Google Drive link instead, which is fine for friends but not for a job portal or a client who won't click links.
In most cases the images are the problem, and shrinking them is the fix. A 30 MB file usually comes down to a few megabytes once its photos are resampled to a sensible resolution, and the pages look the same on a screen.
Why your PDF is so big in the first place
Text in a PDF is cheap. A page of plain text is a list of characters, positions and a reference to a font, which adds up to a few kilobytes.
Images are where the size lives. A scanned page isn't text at all; it's a photo of a sheet of paper. A phone camera shot is often 12 megapixels or more, and if someone dropped a few of those into a Word document and exported it, every one of them is sitting in the PDF at full size, even if it's drawn as a thumbnail in the corner of a page.
So when a PDF is bigger than you expect, the question is almost always "what images are in here, and how much detail do they carry that nobody will ever see?"
What ppi actually means here
Pixels per inch (ppi) only makes sense relative to how big the image is drawn on the page. The same 3000-pixel-wide photo is 1500 ppi if it's printed 2 inches wide, and 300 ppi if it fills a 10-inch-wide area.
So Compress PDF looks at each image's size on the page rather than its raw pixel count. If you pick 144 ppi and a photo is drawn 3 inches wide, the tool keeps about 432 pixels across. The photo shown at full page width keeps more. An image that's already below the target isn't scaled up or touched.
In practice:
- 72 ppi looks soft on a modern phone or laptop screen, though larger text in scans usually stays readable.
- 100 ppi is fine for reading on screen and noticeably smaller than 144.
- 144 ppi is sharp on most screens, including high-density ones at normal zoom.
- 300 ppi is what you'd want for printing.
Settings to pick in Compress PDF
Open Compress PDF and add your file. The compression runs on your own device, so the file never goes to a server.
- Leave Image quality on "Good" (the default, JPEG 80%).
- Open "More options" and check that "Image resolution" is at 144 ppi.
- If the PDF is a scan of black text on white paper, set "Image conversion" to "Grayscale".
- Click Compress PDF and compare the new size with the original.
If the result is still too large, I'd drop the resolution to 100 ppi before touching the quality setting. Going from 144 to 100 roughly halves the pixel count, and most people won't notice on screen. If the PDF is only going to be read on a screen and it's still too big, try 72 ppi.
"Medium" (JPEG 65%) is the next step. It's fine for photos of documents and for pictures of products, but on scans with small print you may start to see blotchy edges around letters. "Best" (JPEG 100%) exists for when you care about image quality more than size; it rarely helps with email.
The other options are more niche. "Multimedia files" set to "Discard" removes embedded audio, video or 3D content, which is unusual but huge when it's there. "Fonts (experimental)" with "Optimize" can trim fully embedded TrueType fonts and stores duplicate fonts once, which helps with PDFs exported from some design tools. "Remove document metadata" strips the title, author and similar details; it makes almost no difference to size but is worth ticking if you don't want your name in the file properties.
Why some files don't shrink
Sometimes the saving is a few percent, or the tool keeps the original because it couldn't get smaller. The usual reasons:
The images are already compressed. If the PDF came from a scanner app that saved JPEGs at a modest resolution, there's not much left to squeeze. Compress PDF only keeps a re-encoded image if it's actually smaller than the original, so it won't make things worse to try.
The file is mostly text and vector graphics. Text and vector drawings (charts, logos, line art) are never turned into images, because that would make them blurry and unsearchable. When an all-text file is big, it's usually fonts or something odd embedded in it.
The images are a type the tool skips. Compress PDF leaves CMYK images, 1-bit black-and-white scans, 16-bit images and image masks as they are. The 1-bit case is the common one: office scanners in "black and white" mode already produce a very compact format, and converting that to JPEG would make it bigger and uglier.
Every run also removes page thumbnails and unused objects, and compresses any data streams that were stored uncompressed. On badly made PDFs that alone can save a fair amount, even with no images touched.
When the upload limit is strict, split the file
Email is forgiving compared to some upload forms. Visa portals, court filing systems and university application sites often have hard limits like 2 MB or 5 MB per file, and no amount of reasonable compression gets a 60-page scanned contract under that.
That's what Split by size is for. You give it a target size and it splits the PDF into parts that each stay under it. Compress first, then split, and you'll end up with fewer parts. If the portal accepts several files, upload them in order and name them clearly (contract-part-1.pdf and so on).
If you're building the PDF from photos in the first place, it's easier to keep it small from the start. See turning photos into a PDF and scanning documents with your phone.
Read a line before you hit send
The most common mistake is compressing hard, sending the file, and only then noticing the signature or the account number at the bottom of page 3 has turned to mush. Open the compressed file, zoom to 100% on the page with the smallest text, and read a line. If you can read it comfortably, so can the person you're sending it to. The output is saved as compressed-yourfile.pdf, so your original is still there if you need to try again with a gentler setting.
Also useful: Split by size · Grayscale
