Why a redesign usually loses the traffic it inherited
The pattern is familiar. A new site launches, everyone is pleased with it, and eight weeks later somebody notices that enquiries are down and a page that used to arrive high in search results has disappeared. Nothing dramatic went wrong. The URLs changed, the redirects covered the top twenty pages rather than the whole estate, several pages were consolidated without anyone deciding where their content went, and the search engine has been quietly reassessing the site ever since.
Avoiding it is procedural rather than clever. Before anything is designed we crawl the existing site and pull every page that receives traffic or holds inbound links, then produce a mapping document with a destination for each one. Pages being retired get a specific redirect to the closest equivalent, not a blanket rule sending everything to the homepage, which search engines treat as a soft error. Query strings, trailing slashes, uppercase paths and the old sitemap all get considered.
After launch the work continues for a few weeks. Server logs and search console coverage reports show what is actually being requested and what is returning errors, which is usually a longer list than the redirect map anticipated, because old links exist in places nobody thought to check. Keeping page titles, heading structure and internal linking recognisably similar for the pages that were already performing is also worth doing, even when the temptation is to rewrite everything at once.
- A crawl of the existing site before design, listing every URL with traffic or inbound links
- A redirect map with a specific destination per page, not a catch all rule to the homepage
- Query strings, trailing slashes and case variants included in the mapping
- Error monitoring and coverage reports reviewed for several weeks after launch
- Titles, headings and internal links kept recognisable for pages that were performing