PDF compression works by reducing or rewriting the data inside a PDF. The best approach depends on whether the document contains scanned images, photographs, vector graphics or mostly text. Start by checking the original file size and the quality you need after compression.
For a text-heavy PDF, structural optimization may reduce some overhead without visibly changing the document. Image-heavy PDFs are different: their size is often dominated by embedded images, so meaningful reduction may require downsampling or recompressing those images.
Before sharing a compressed PDF, open the output and check every page. Look for missing images, altered fonts, blurry scans, broken links or unexpected page changes. Keeping the original file is also useful if you later need a higher-quality copy.
ToolsPoint Online provides a browser-based PDF Compressor for supported workflows. Because browser capabilities vary, the tool page explains what the current implementation can do instead of pretending every PDF can be reduced by the same percentage.