Redirects are the safety net of a website migration. When someone follows an old link, from Google, a newsletter or another site, a redirect sends them to the right new page. Get them right and the move is invisible. Get them wrong and traffic leaks away quietly for months.
The redirects themselves take an afternoon. The work that matters is the inventory that comes before them.
Step 1: Find every URL that exists
No single source has the full list, so we combine several:
- A full crawl of the live site, following every internal link
- The sitemap, which often lists pages the navigation has forgotten
- The CMS export, for blog posts, collections and draft pages that went live once
- Google Search Console, for every URL Google has indexed or shown in results
- Analytics, for landing pages that have had visits in the last year
- Backlinks, from Search Console’s links report or an SEO tool, for pages other sites link to
The overlap is large, but the gaps are where problems hide: an old campaign page still linked from a partner site, or a PDF that ranks for a valuable search.
Step 2: Don’t forget the non-page URLs
These catch people out most often:
- PDFs, brochures and downloadable files
- Images that rank in image search
- Old blog category and tag pages
- Pages with tracking parameters still linked from emails
- Different versions of the same URL: with and without
www, with and without a trailing slash
Step 3: Rank by value
Not every URL deserves the same attention. We sort the list by traffic, backlinks and conversions. The top twenty or so get a hand-checked redirect to the best possible match. The long tail gets sensible rules.
Every old link is a promise someone else made on your behalf. Redirects keep it.
Step 4: Map old to new
For each URL we decide one of three things:
| Situation | What we do |
|---|---|
| The page exists on the new site | Keep the same URL if possible. If not, 301 redirect to it. |
| The content merged into another page | 301 redirect to the page that now covers it. |
| The page is gone with no equivalent | Redirect to the closest parent page, or let it return a proper 404 if nothing fits. |
A few rules we follow:
- Use 301s, not 302s. A 301 says the move is permanent, which is what you want search engines to remember.
- No chains. Old page to new page, in one step. Redirects that hop through several URLs slow things down and dilute their value.
- Use patterns carefully. Rules that redirect a whole folder at once are useful, but we test them against the full list so they don’t catch pages they shouldn’t.
Step 5: Test before launch
We load the redirect map onto staging and run every old URL through it with a script. Each one should land on a working page in a single hop. Anything that returns an error or lands on the homepage gets a second look.
On Cloudflare, redirects can live in a simple file alongside the site or as rules at the network level, so they respond almost instantly.
Step 6: Watch after launch
Even with a careful inventory, something occasionally slips through. For the first month we check Search Console for new 404 errors each week and add redirects as needed. After that, the list is stable and the old site can be retired.
The short version
- Combine every source to find every URL
- Include files, images and odd URL variations
- Rank by value and hand-check the important ones
- Map each one with a single permanent redirect
- Test the whole list before launch
- Monitor for a month afterwards
If you’re planning a move and want this done properly, it’s part of every migration we run.