Email bounces at 25 MB, upload forms cap at 10, and the PDF you need to send is 40. The usual next step is searching for a PDF compressor, and the usual result is disappointment: the file comes back 2% smaller.
That is not always the tool being bad. It is often the file having nothing left to give.
Find out what is making it large
PDFs get big for exactly two reasons, and the fix is completely different for each.
Embedded images. A scanned document is not text — it is a photograph of text, one per page, and those photos are frequently stored at 600 DPI when 150 would be indistinguishable on screen. Ten scanned pages can easily be 30 MB. This is the cause maybe 90% of the time.
Embedded fonts and vector content. Text-based PDFs from a word processor are usually small already — a hundred pages of text might be 2 MB. If your text PDF is enormous, the culprit is usually a font subset problem or a very complex vector graphic, and neither compresses much.
The quick diagnostic: divide the file size by the page count. Above roughly 500 KB per page, you have an image problem. Well under that, compression will not save you.
What actually works
If it is scanned images, the real fix is reducing the images, not “compressing the PDF”. A compressor that leaves a 600 DPI scan at 600 DPI is only re-packing it. What helps is resolution: dropping a scan from 600 to 150 DPI cuts the pixel count by sixteen.
Practically, that often means rendering the pages to images at a sensible DPI, compressing those images, and rebuilding. You lose selectable text, which a scan never had anyway.
If you only need part of it, extract the pages you need. Sending three pages instead of forty is a 92% reduction and takes ten seconds. This is by far the most underused answer, and often the correct one — the recipient rarely wanted the whole document.
If it is a text PDF, re-export it from the source rather than compressing the output. Most word processors have a “reduced size” or “optimised for web” export that subsets fonts properly.
Why we do not offer a “compress PDF” button
We could add one. It would strip metadata, deduplicate objects and re-save with better object streams — and on a typical scanned document it would shave off a few percent while the page title promised something much larger.
Genuine PDF compression means re-encoding the embedded images, which requires decoding the PDF, rasterising, re-compressing and rebuilding. That is a real piece of work rather than a checkbox, and until we have built it properly we would rather point you at what actually helps than ship a button that mostly does not.
The five-second version
Divide size by pages. Over 500 KB per page, it is scans — reduce resolution. Under that, it is text — re-export from the source. Either way, first ask whether you need to send all of it.