The bounce message is always the same: attachment too large. Phone scans save every page as a full-resolution photo, slide decks embed the same logo forty times, and exported reports carry print-grade images nobody will ever print. The Compress PDF card re-encodes that weight away in your browser — downsampling images to screen resolution, deduplicating repeated resources, and cleaning dead metadata — while the text layer stays perfectly intact and searchable.
Know your target size before you start
Compress with a destination in mind: under 25 MB for Gmail, under 10 MB for most corporate inboxes and application portals, under 2 MB if humans will open it on phones. A good first pass typically cuts scans by 60–80% with zero visible change at reading zoom. Only if you're still over budget should you step quality down further — and always compare the result at 100% zoom, not thumbnail size, because thumbnails hide sins that full pages reveal.
What compression can't fix (and what can)
Compression shrinks bytes, not pages: a 200-page document stays 200 pages, so carve out what you don't need with Split PDF first — dropping half the pages beats any encoder. Already-small files that still bounce are usually one giant embedded image away from compliance; check the worst pages individually. And if the goal is archiving rather than sending, skip compression guilt entirely and keep the original — storage is cheap, re-scanning a faded receipt in five years is not. Everything here runs locally in your browser, so even sensitive contracts never leave your machine.