How to Compress a PDF Without Losing Quality
Why PDFs balloon in size, what's actually safe to strip out, and a practical workflow for shrinking a file without visibly hurting it.
Published January 15, 2026
Most "why is this PDF 40MB" problems come down to one of two things: images that were never compressed before they got embedded, or structural bloat the PDF format itself accumulates over edits, exports, and merges. Those two problems need different fixes, and using the wrong one is why compression sometimes does nothing.
Find out what's actually taking up the space
Open the file and look at it honestly. A 12-page text contract that's 8MB almost certainly has embedded fonts or leftover metadata bloating it, not images. A 12-page scanned document that's 40MB is almost certainly images, usually scanned at a much higher DPI than the page will ever need. Photos and screenshots pasted straight from a phone or a screenshot tool are frequently 3-10x larger than they need to be for anyone to read them clearly on screen.
For image-heavy PDFs: compress before you assemble
If a PDF is going to be full of photos or scanned pages, the highest-leverage move is compressing those images before they go into the PDF at all, not after. A PDF compressor working on the finished file has to guess which streams are images and how much it can touch them without visibly degrading the page; compressing the source images first means you're making that quality decision deliberately, with a preview, instead of leaving it to an algorithm.
A reasonable target for on-screen viewing is around 150 DPI for text-heavy scans and 100-150KB per photo for a full-page image at typical page dimensions. Anything scanned at 300 DPI or higher for a document nobody's printing is wasted size.
- Compress each image first, aiming for the smallest file where text and detail are still crisp at 100% zoom
- Assemble the compressed images into a PDF rather than compressing the whole PDF after the fact
- Keep an uncompressed original somewhere if the PDF might ever need to be printed at high quality
For text-heavy PDFs: structural compression does the work
A PDF exported from Word or generated by some other tool often carries redundant font subsets, uncompressed object streams, and editing history metadata that serves no purpose once the document is final. None of that is visible content, so removing it doesn't touch quality at all. This is what a "compress PDF" tool that works on file structure rather than images is doing: stripping unnecessary metadata and re-saving with object streams, which is exactly the kind of compression that's safe to apply blindly.
The tradeoff is that this approach won't do much for a PDF that's already lean, or one whose size is genuinely coming from embedded images rather than structure. If running it barely changes the file size, that's a sign the bloat is in the images, not the structure, and it's worth going back to the image-first approach instead.
A workflow that actually works
For scanned documents or photo-heavy PDFs: compress the source images to a sensible quality first, then combine them into a PDF. For text documents, reports, and contracts: run structural compression on the finished PDF and expect it to help more the more that file has been edited and re-exported over its life. For anything that mixes both, do the image pass first and the structural pass second, since order matters here; compressing structure on a file still carrying oversized images barely moves the needle.