Why email limits are stricter than they look
Attachments are not sent as raw bytes. They are encoded for transport, and that encoding inflates them by roughly a third — a 500KB file arrives as about 680KB of message. Many mail systems apply their stated limit to that encoded size rather than to the file on your disk.
The limit usually applies to the entire message too, not to each attachment: the body, any images in your signature, and every attachment together. This is why a file comfortably under a stated 1MB cap can still bounce, and why aiming meaningfully below whatever number you were given is the sensible habit rather than paranoia.
At this size the question is how much to keep
500KB is roomy for a document, which shifts the problem. Rather than fighting to fit, you are deciding how much resolution to retain, and for most text documents the honest answer is that you will not see the difference — a scan at 150 DPI and the same scan at 300 DPI look identical on a screen and very similar in print.
Photographic content is where the budget actually goes. A report that is mostly full-page images will use the entire allowance on a handful of pages, and no setting changes that arithmetic. If image quality genuinely matters, sending fewer images at good quality beats sending all of them badly.
A useful habit at this size is to compress once and then actually look at the result at the scale it will be viewed. Judging a compressed document by zooming to 400% will always disappoint, because you are inspecting artifacts nobody else will ever see. Judged at reading size, a well-compressed 500KB document is usually indistinguishable from the original, and that is the standard that matters.
When not to compress at all
If your PDF is already under the limit, running it through a rasteriser does nothing but degrade it — there is no prize for arriving further under a cap you had already cleared. This tool returns the original untouched whenever compression would produce a larger or pointless result, but it is worth knowing so you do not go looking for a setting that would have made things worse.
The same applies to digitally created PDFs. Exported from a word processor, presentation tool or design application, they are already efficient and their text is real text. Rasterising one to shave a few kilobytes trades away searchable, selectable, accessible content for a saving nobody asked for.
Better options than compressing, when you have them
If the recipient can accept a link, sending one avoids the problem entirely and delivers the document at full quality. Most cloud storage services generate a shareable link in a couple of clicks, and unlike an attachment it does not sit in the recipient's mailbox consuming their quota. The obvious caveat is that a link is only as private as its permissions, so check who it is actually shared with before sending anything sensitive.
It is also worth asking what the recipient needs. A fifty-page report emailed in full because it was easier than choosing is a common source of size problems that need not exist. Extracting the pages that are actually relevant produces a smaller file, a faster reply and a better impression than a heavily compressed complete document nobody was going to read in full.