Two PDFs with the same page count can differ in size by a factor of a hundred, and the reason is rarely obvious just from looking at the document. A 20-page contract might be 400KB. A 3-page form might be 8MB. If you are trying to figure out why one specific file is enormous before you even get to compressing it, here is where that size is actually coming from.
Scanned or photographed pages are the biggest factor by far
Any page that started as a photo or a scan, rather than being typed directly into a document, is stored as an image, and images are expensive. A single page scanned at a high resolution can run one to three megabytes on its own before any compression. Multiply that across a multi-page document and you get files that dwarf anything typed from scratch. This is the single most common reason a PDF is large, and it is also the one compression addresses best, since resampling and re-encoding images is exactly what a compressor is built to do.
Exported PDFs can be surprisingly bloated too
A PDF does not have to contain a photo to be large. Documents exported from design software, or PDFs created by "printing" a webpage or a presentation to PDF, often embed full, uncompressed versions of every image and font they touch, including ones the final document barely uses. A slide deck with a handful of high-resolution background images, exported straight to PDF without any optimization step, can end up several times larger than it needs to be for something that is only ever going to be viewed on a screen.
Fonts, especially embedded ones, add up
Every font used in a PDF typically gets embedded directly into the file, so it displays correctly on any device without relying on that font being installed. A document using several custom fonts, especially ones with full character sets for multiple languages, can carry a meaningful amount of extra size just from the fonts alone. This is usually small compared to images, a few hundred kilobytes at most, but it adds up in documents that are otherwise fairly light.
Leftover structure from editing history
A PDF that has been edited, resaved, or had pages added and removed multiple times in the same software can accumulate objects that are no longer actually used, orphaned data left behind by the editing process rather than anything visible on the page. This does not usually account for a huge share of the size on its own, but it is one of the things a compressor cleans up as a matter of course, alongside the more significant image resampling.
What compression will not fix
Occasionally the images in a PDF are the actual content, not incidental, an architectural drawing, a detailed map, a high-resolution product photo someone specifically needs to zoom into. In those cases there is a real ceiling on how much you can shrink the file before it stops doing its job, since the detail being compressed away is the reason the file exists. This is the exception rather than the rule, but it is worth recognizing before assuming a setting is broken because the file did not shrink as much as a typical scanned document would.
How to tell which one applies to your file
The fastest check is simple: if the PDF came from a scanner, a phone camera, or a "photo to PDF" conversion, images are almost certainly the cause, and compression will make a real difference. If it is a document you typed and exported directly, and it is still unexpectedly large, the cause is more likely embedded fonts or leftover structure from editing, and while compression still helps, do not expect the same dramatic drop you would see on a scanned file. Running it through the compress tool either way costs nothing to try, and the before-and-after size tells you immediately which category your file fell into.
If you want the mechanics of what compression actually changes once you have identified the cause, this piece covers that in more depth, including what does and does not affect visible quality.