Short answer: they are maintained if every URL with rankings is migrated one-to-one. If that goes wrong, your organic traffic can drop significantly — and that is not always repairable.
A ranking is not tied to your page but to the address of that page. Everything a page has built up over the years is attached to that single URL. If the address changes without someone providing the new destination, that page starts from scratch again. No matter how beautiful the new site is, it does not matter.
The same applies to external links. A trade journal that linked to your services page four years ago will not adjust that link just because you have a new site.
The rule we follow
- URLs with rankings remain unchanged. Do not make them prettier, do not make them shorter, do not just clean them up.
- If a URL must change, we redirect one-to-one to the new page that is closest in terms of content.
- If pages are merged, we choose the primary one. The rest redirect there, so that the accumulated value is consolidated instead of fragmented.
- Pages that can be removed can truly be removed. Not everything deserves a redirect.
Where things go wrong in practice
- Redirecting everything to the homepage. Google reads that as “this page no longer exists”.
- Chains: old redirects to new, new redirects to newer.
- Temporary instead of permanent redirects.
- Forgetting variants: with and without trailing slashes, with capital letters, with search parameters, and other language versions.
- Only thinking about pages, while images and PDF files also have addresses that originate from somewhere.
And the honest part: this is not always reversible. If the old structure is still known and the error is found quickly, it usually recovers for the most part. But if the old site is already gone and no one has recorded which URLs existed, then a portion is permanently lost. That is why the initial inventory is not a formality: that document is exactly what allows an error to be corrected afterward.