Guide
Two PDFs with the same page count can need completely different compression strategies because PDF is a container, not one uniform kind of document.
Open the original and try to select one sentence. If individual letters highlight, the page probably contains real text. If the whole page behaves like one picture, it is probably scanned. Mixed PDFs are common: a cover letter may contain selectable text while attached certificates are page images.
A scanner can capture every page at high resolution and in full color even when the page is mostly black text on white paper. Ten letter-size pages at photo-like resolution contain far more pixels than a screen or upload portal needs. Downsampling and JPEG encoding can remove a large amount of that data.
A clean export from Word, Pages, or Google Docs may already store text and vector shapes compactly. Embedded fonts and photographs can add weight, but there may be no giant page image to shrink. A text-preserving rewrite can remove some overhead without promising a dramatic result.
Rasterization draws each page into a new image and places that image in a new PDF. The page looks similar, but selectable text, hyperlinks, form fields, bookmarks, layers, annotations, and some accessibility information are not preserved by that process.
Rasterization may be acceptable for a simple scan submitted to a portal that only needs a readable visual copy. It is a poor default for a resume, tagged accessible document, fillable form, digitally signed file, or any document whose searchable structure matters after upload.
When a PDF mixes text and scans, check both kinds of page. A setting that leaves typed pages looking sharp can still blur a faint stamp, and a setting that preserves a scan can leave the file above target. Page-by-page review is more reliable than judging the first page.