Email attachment limits (commonly 10-25MB) and upload portal caps (sometimes as low as 2-5MB) are hard limits, not suggestions, a file one kilobyte over gets rejected the same as one ten times over. Hitting a precise size target reliably takes a slightly different approach than general compression, since "compress it more" doesn't always land exactly under an arbitrary number.
Start with compression, and check the actual result
The first move is always image compression, since images account for the overwhelming majority of size in most oversized PDFs. Running the file through Compress PDF at a stronger quality setting than the default often gets a moderately oversized file under a target limit on its own. It's worth checking the actual resulting file size against the limit directly rather than assuming "compressed" automatically means "small enough", a 40MB scan compressed by 70% is still 12MB, which clears a 25MB email limit but not a 10MB upload cap.
When compression alone isn't enough
Some documents, especially long scanned files or ones already near-maximally compressed, won't get under a strict size limit through compression alone without becoming visibly degraded. In that case, splitting the document into multiple smaller files with PDF Split by Size is usually a better outcome than over-compressing a single file into illegibility, since it lets each piece land comfortably under the limit while keeping image quality intact. This is standard practice for large scanned reports or archives being uploaded to portals with strict per-file caps.
If the size problem comes specifically from unnecessary pages rather than genuinely needed content, checking whether every page is actually required is worth doing before compressing at all, removing blank or irrelevant pages with Remove PDF Pages or the PDF Blank Page Remover can meaningfully cut size with zero quality tradeoff.
A practical order of operations
For a file that needs to hit a hard limit: remove unneeded pages first, since that's free size reduction with no quality cost; compress remaining images next, checking the resulting size against the target; and only split into multiple files as a last resort if the limit still isn't met without unacceptable quality loss. Doing it in this order avoids over-compressing content that could have been removed or resized instead.