Preparing a rebuild: the SEO decision checklist

Inventory what exists before drawing anything

List the pages and what they are actually used for: offers, content, tools, contact pages, resources. Add the available data on visits and on enquiries received. The goal is not to keep everything, but to know what to keep, improve, merge or remove — and to decide it consciously.

A page with few visits that generates enquiries is worth more than a heavily visited page that produces none. The inventory must cross both dimensions, otherwise it leads to deleting exactly what should have been kept.

Tie every page to a precise need

A service page must explain the intervention; an article must answer a question; a tool must do something useful. Plan the links between them instead of designing three separate silos that never cite each other.

This internal linking is not an SEO trick: it is the real path of a visitor who compares, hesitates and comes back. If they have to go through the menu at every step, they will mostly go through the search engine instead.

Prepare the URL mapping

Document the old addresses and their destination wherever pages move. That mapping is decided with the technical team during framing, not after go-live.

• One permanent redirect for every address that moves.
• No redirect chains: a single hop to the final destination.
• A relevant destination, never the home page by default.
• An explicit fate for deleted pages: a useful redirect or a deliberate 410.

Write the migration acceptance plan

The plan covers internal links, forms, reachable pages, metadata, indexing rules and the sitemap. Keep the staging settings carefully separate from the public site: one forgotten `noindex` in production costs several weeks.

Add one simple but decisive check: is the page served to a crawler the same as the one a visitor sees? Conditional rendering, caches and country redirects produce differences the human eye never catches.

Follow up after go-live

Keep a baseline before the switch: list of URLs, positions, volumes, enquiries received. Then watch the crawl errors, the pages dropping out of the index and the forms that no longer complete.

Finally, name someone to follow up and set a correction window. A rebuild does not end at deployment: it ends when the curves are back to their level, or when you understand why they will not come back.

The best redirect plan is the one that exists before go-live, not the one reconstructed from the errors observed afterwards.