What does flattening a PDF actually do?
A PDF can hold live objects on top of the page: form fields with values, sticky notes, highlights, drawn signatures and stamps. They are separate from the page content, which is why a recipient can click a field and retype it, or drag an annotation away.
Flattening draws those objects into the page itself and removes the live versions. The document looks identical and behaves like a printout: nothing to click, nothing to edit, nothing to delete.
When is flattening the right answer?
Returning a filled form, so the answers cannot be altered after you sign. Sending a signed document, so the signature is part of the page. Sharing a marked-up draft where the comments should be visible but not editable. Printing reliably from a reader that ignores annotations.
It also solves the common complaint that "my form answers disappeared when I emailed it": some readers never save field values at all, and a flattened file cannot lose what is now printed on the page.
When do you need redaction instead?
Whenever the point is that somebody must not read something. Flattening a document with an account number on it produces a document with an account number on it — now permanently.
The two combine well in one direction: redact first, then flatten, so the black boxes become part of the page and the removed text is gone. Doing it the other way round is harmless but pointless.
A black rectangle drawn as an annotation and then flattened hides the text visually, but the text object underneath is still selectable. Redaction is what removes it.
Does either one make the file smaller?
Not usually. Flattening removes the annotation objects but adds their drawn appearance to the page, so the size lands close to where it started. Redaction removes a little text and adds a few rectangles.
If size is the goal, compression is the tool for that job — it works on the images, which are where the megabytes live.