Why Merging Doesn't Require Re-Rendering Any Content
Combining multiple PDFs into one document is a structural operation — concatenating each source file's internal page objects into a single new page tree, in the order you specify — not a visual re-rendering of any page's actual content (see Complete Guide to PDF Processing for the deeper explanation of why this makes merging fast even on large files).
Because merging doesn't touch the underlying content, source documents with different page sizes or orientations stay exactly as they were after merging — a mix of portrait and landscape pages, or different paper sizes, doesn't get automatically normalized into a consistent format unless that's done as a separate, explicit step. This is worth checking before merging documents from different sources, since an unexpected mix of page formats in the final merged file is a common surprise for anyone assuming merging implies normalization.
Reordering files before merging (not just after) matters for the same reason getting image order right matters for image-to-PDF compilation — fixing a wrong page sequence after merging generally means starting the merge over, rather than reordering pages within an already-merged document.