A UK SaaS companyA rebuild that kept every ranking it had

The problem they came with

The hard version of a rebuild. 412 URLs, a blog with years of accumulated rankings, and a marketing team whose pipeline could not survive a traffic dip while everyone waited for Google to recover.

The old site was also slow — 2.8s Largest Contentful Paint — which was costing conversions before search ever entered the conversation.

What we did

1. 01

   Crawl and map every URL before the new build shipped

   412 URLs matched to destinations and tested on staging. The redirect map existed before the design was finished.

2. 02

   Preserve the pages that were already earning

   The instinct in a rebuild is to rewrite everything. Pages with existing rankings were migrated with their structure intact and improved afterwards, not replaced on day one.

3. 03

   Rebuild the front end for Core Web Vitals

   Image handling, font loading and render-blocking resources dealt with during the build rather than in a later performance sprint.

4. 04

   Launch, then watch Search Console daily for a fortnight

   Migrations fail quietly. Coverage and impressions were checked every day for two weeks so anything broken was found in hours, not in the next quarterly review.

What happened

Load time fell from 2.8s to 0.9s and no rankings were lost across 412 migrated URLs. The second number is the one that mattered to them, and it is the one most rebuilds cannot claim.

2.8s LCP, 412 URLs0.9s LCP, 0 lost rankings

Search Console — client property, measured across the engagement