Why PDFs get large in the first place
PDF size is usually driven by images, embedded fonts, and uncompressed object streams. A 50-page report with high-resolution photos can easily exceed 50 MB.
Scanned documents are especially heavy because every page is stored as a full-resolution image.
Common compression techniques
Image downsampling reduces the resolution of embedded images to what the page actually needs (typically 150-300 DPI).
Image re-encoding swaps PNG images for JPEG where appropriate, which can dramatically shrink photographs.
Object stream compression and font subsetting remove unused glyphs and pack the file's internal objects more efficiently.
Quality trade-offs
Light compression is essentially invisible: re-saving the PDF with optimised streams can shrink size by 20-40% with no quality loss.
Aggressive compression downsamples images and increases JPEG compression, which is fine for screen reading but can soften photographs and fine print.
Best practices
Start with light compression and only increase if the file is still too large. Always keep a copy of the original. For documents that will be printed, use medium compression at most so images remain crisp on paper.