Mail providers cap attachments at around 25 MB, and many company servers and government portals are far stricter — 10 MB, sometimes 5. A folder of scans sails past that without trying.
The instinct is to run the file through the first compressor available. That works for scans and actively harms text documents, so it is worth knowing which one you have.
Open the PDF and try to select a line of text. If the cursor selects words, the document is text: it is probably already efficient, and there is far less to gain than you hope. If nothing selects, each page is an image, which is where the megabytes are and where compression pays off enormously.
Mixed documents — a report with photographs, a contract with a scanned signature page — sit in between: the images carry the weight, the text does not.
Many tools shrink every PDF the same way: they turn each page into an image and compress that image. On a scan this is exactly right. On a text document it is destructive — crisp vector text becomes a picture of text, which looks blurred when zoomed, stops being searchable and selectable, and is frequently larger than the original.
A correct compressor checks the document first. Abzar's does: text-heavy files get a lossless structural optimisation that keeps the text exactly as sharp as it was, and image-heavy files get their images recompressed, which is where the actual weight is.
In order of how much they help, without wrecking the document:
Two things routinely get lost and matter later. A digital signature becomes invalid as soon as the file is modified, so compress before signing, never after. And heavy image compression can render small print — the fine text in a contract, a serial number on a scan — genuinely unreadable. Always open the compressed file and check the smallest text on the page before sending it.