Why your PDF won't get smaller
5 min read
Clicking compress and getting 3% back is common. That's usually the document's nature rather than the tool failing.
PDF is already a compressed format
Text and vector data inside a PDF are normally stored compressed already. Re-compressing them is like zipping a ZIP file — there's little left to squeeze.
So what a lossless pass really does is tidying rather than compressing: dropping duplicated objects, removing unused data, and rewriting the structure. A well-built document has almost nothing to gain from that.
Images are where the bytes are
In a 10MB PDF, 9MB is often images. A scanned document is the extreme case: every page is one large photograph.
To shrink meaningfully you have to re-encode those images — lower resolution, lower JPEG quality. That discards information and cannot be undone.
Which documents shrink and which don't
You can usually predict the outcome before you start.
- Shrinks a lot: scanned documents, photo-heavy material, decks with full-resolution images dropped straight in
- Barely shrinks: text-first documents exported from a word processor, files that were already compressed
- May grow: text-only documents rebuilt as images
What losing the text really costs
Redrawing pages as images shrinks a lot, but the result can no longer be selected or searched. For anything going into a document management system, that can be a dealbreaker.
It's fine for a viewing copy. Think twice before it replaces your original.
A reasonable order to try
Start with the lossless pass and look at the result. When you're only a few megabytes over an attachment limit, that's often enough.
If it isn't, split the document, or — for scans — use the image rebuild and keep the original somewhere safe.