You need to understand what are the reasons for a file weight. You can probably analyze the file and see what is the best action.
Let me name a few elements and explain a bit about the variables, and some method to evaluate them.
I. Is the text, really text?
On a long file, this is probably the most important element. Using PDF adobe reader, and if when using the marker or select tool, you are able to actually select the text, the test is passed. You have text.
But sometimes the text is just a “photo of a text” (see point II) and sometimes is what we called curves. They are the shape of letters, but each is an independent figure. They take more space than real text and can not be compressed.
If the text is made of curves, the only way to fix it is by having the original document and generate a new PDF without this text-2-curves conversion.
II. The images.
In some cases, this is the easier part to compress, and is the main target with “compression” software. If the images are high resolution they can be resampled, and if they are not compressed they could be, using a jpg algorithm.
But if the document is mainly images, for example, a bunch of scanned documents, the “quality” could be reduced, of course depending on what is the objective of the images, for example, if they are meant to be for high-quality print.
They could be also in CMYK mode, with compresses less efficient than RGB color mode.
If you can control the compression and resampling try different combinations, for example, assigning 100PPI instead of 300PPI or adding some more Jpg compresion value. Convert the images to RGB if needed.