A large PDF and a large photo file both need to shrink, and it's tempting to assume compressing either one is basically the same process. Under the hood, they're actually fairly different, mostly because of what a PDF typically contains versus what a photo file is.
A scanned document PDF is usually a series of embedded images — one per page — wrapped inside a PDF structure that also holds text layers, metadata, and formatting information. Compressing a PDF effectively means compressing each embedded image individually while leaving the document structure intact, which is a slightly different job than compressing one standalone photo.
A PDF that's actual selectable text (a Word document exported to PDF, for instance) is already fairly small and compresses with plenty of headroom left, since text data is inherently lightweight. A PDF that's a scanned image of a document — common for signed contracts, ID copies, receipts — behaves much more like a big collection of photos, and that's where file size balloons and compression makes a real difference.
Because a scanned document's compression is really image compression applied to each page, pushing it too aggressively can blur the text within the scan to the point of being genuinely hard to read — a real risk that doesn't apply the same way to a photo of, say, a landscape, where a bit of softness rarely matters.
Compressing a 20-page scanned PDF means processing every embedded page image, not just one — the tool needs to handle the whole document consistently rather than treating it as a single image, which is part of why dedicated PDF compression tools exist separately from basic image compressors.
For a single photo, straightforward image compression works fine. For a scanned multi-page document you need to keep legible, it's worth using a moderate compression level and checking the text is still clearly readable afterward — rather than pushing for the smallest possible file size at the cost of legibility.
Both file types genuinely benefit from compression, but a scanned PDF deserves a bit more care about the compression level than a standalone photo does, precisely because there's readable text riding along with the image data.