Website Migration: Building a Redirect Map to Keep Rankings — Web & E-Commerce article by Rui Codex

To keep rankings through a website migration, build a redirect map that pairs every indexable old URL with the closest matching new URL using a permanent 301 redirect. Then verify the map against crawl data, analytics and Search Console before launch. Pages without a sound destination should go to the most relevant replacement, never automatically to the homepage.

Key takeaways

  • A redirect map is the single source of truth that tells search engines and visitors where each old URL now lives.
  • Build the inventory from several sources, because no single export lists every URL that has search value.
  • Match pages by intent and content, and only consolidate pages when the merged page genuinely answers the same questions.
  • Test chains, loops and status codes before launch, and monitor crawl errors and traffic for weeks afterwards.

What is a redirect map and why does it matter?

A redirect map is a structured list that pairs each old URL with its new destination and the HTTP status code that connects them. Its purpose is to preserve the value that search engines have already assigned to old pages, including links, relevance signals and history. Without it, the new site starts from zero for every page that moved.

The permanent status code matters. The HTTP semantics standard defines 301 Moved Permanently as a signal that the target resource has a new permanent address (RFC 9110, HTTP Semantics). Google's guidance on site moves recommends permanent redirects for pages that have moved for good (Google Search Central, site moves with URL changes).

Treat the map as an engineering artefact, not a spreadsheet afterthought. It should be versioned, reviewed by someone who knows the new information architecture, and loaded into the server or CMS as code. Teams that do this well can explain every row, which makes the testing phase far shorter.

How do you build a complete inventory of old URLs?

You build it by combining at least four data sources, because each one misses something. Start with the old XML sitemap, then run a full crawl of the live site, then export URLs from server logs and analytics. Add the Search Console performance report and backlink data as the final layer.

  • Crawl export: finds linked pages, paginated lists and parameter variations.
  • Server logs: reveal URLs that bots and visitors request, including orphaned pages.
  • Analytics: shows which URLs carry real traffic and conversions.
  • Search Console: lists pages that earn impressions and clicks.
  • Backlink data: identifies external links that point at URLs no internal page references.

Merge these lists and deduplicate them carefully. Trailing slashes, uppercase letters, http and https variants and tracking parameters often create duplicates that look like separate pages. Normalise them first, then decide which canonical version each one maps from.

Tip: Keep the source of every URL in a column. A URL that appears only in backlink data is a signal to check whether the external link still matters.

How do you match each old URL to its new destination?

Match each old URL to the new page that answers the same question for the same searcher. Exact matches come first, then close matches in the same category, and only then consolidated targets. Pages with no realistic successor are the exception and need an explicit decision.

Exact, category and consolidated matches

An exact match is a page whose content and purpose survived the move, often with only a new path. A category match sends a discontinued product or article to the closest surviving category. A consolidated match merges several thin pages into one stronger page, which can work well but only when the merged page answers every question the old pages answered.

Removed pages

Some pages should not have a destination at all. If a page is intentionally retired and has no links or traffic, returning a 404 or 410 is honest and correct. Redirecting it to an unrelated page only creates confusion for the reader and for the search engine.

The trade-off is specificity against simplicity. Consolidation reduces the number of pages to maintain, but it can blur the topic a specific URL used to rank for. Make that call per URL, with the traffic and link data in front of you.

When do redirects fail to protect rankings?

Redirects fail when the chain is long, the target is irrelevant, the status code is wrong, or the new page cannot be crawled. Each failure is common, and each is visible in logs if you know where to look. The most damaging one is usually the blanket redirect of many pages to the homepage.

  • Redirect chains: old URL to intermediate URL to final URL adds latency and risks losing signals along the way.
  • Redirect loops: two rules that point at each other make the page unreachable.
  • Temporary codes: a 302 used by mistake signals a short-term move, which is the wrong message for a permanent migration.
  • Homepage catch-alls: sending unmatched pages to the homepage makes them look like soft errors, as Google's redirect documentation explains.
  • Blocked new pages: a robots rule or noindex tag on the new site can stop the destination from being indexed at all.

Protocol and host variants are the quiet failure. If http, https, www and non-www versions each resolve differently, search engines may index several copies of the same page. Decide one canonical host and redirect every other variant to it in a single hop.

Tip: Map the chain for every row. If the final destination is not reached in one hop, rewrite the rule so the old URL points directly to the final page.

How should you test the redirect map before launch?

Test the map on a staging environment that mirrors production routing, then crawl the complete list and compare results against the expected destinations. Any mismatch in status code, target or final URL must be fixed before the switch. Testing is cheap compared with recovering from a failed launch.

  • Every old URL returns a single 301 with the expected target.
  • No chains longer than one hop and no loops.
  • Every destination returns 200 and is indexable.
  • Canonical tags on destination pages point to themselves.
  • Internal links on the new site no longer point at old URLs.
  • XML sitemap for the new site lists only final URLs.

Run a crawl of the old URL list against the staging host, and keep the output as a baseline. Repeat the crawl after every change to the map, so you can see exactly which rows moved.

Hypothetical examples of redirect mapping

Here are two hypothetical scenarios:

Scenario one. A hypothetical online shop renames its category paths from a legacy structure to a clearer taxonomy. Product pages keep their content, so each product URL receives a direct 301 to its new path, and the old category pages redirect to the matching new category. The team checks that no category redirects to the homepage, because each category has a clear successor.

Scenario two. A hypothetical company blog has many short posts on the same narrow topic, written years apart. The team consolidates them into a single guide, redirects each old post to the guide section that covers its question, and removes one post that had no traffic, links or search impressions, returning a 410 for it. The trade-off is that the guide must now answer every original question, and the team reviews it against each old post before launch.

Neither scenario shows real results. They illustrate the decisions that a redirect map forces: exact versus consolidated matches, and redirect versus deliberate removal.

What should you monitor after the switch?

Monitor crawl errors, redirect hits, indexing status and organic traffic by page group for several weeks after launch. Changes in rankings often appear first as shifts in impressions, so compare groups rather than single keywords. Any group that drops sharply should be traced back to its rows in the map.

Keep the old-URL redirects running, and keep watching the Search Console reports for not-found errors. Search engines recrawl at their own pace, so a clean map can take time to show up in the reports. Patience matters, but so does acting on a group that does not recover.

If you are planning a migration and want an engineering review of your URL inventory and redirect rules, you can request a project consultation or read how we approach web development projects. Our work on system integrations and the web and e-commerce engineering guides covers the wider data flows that migrations touch.

Frequently Asked Questions

Should I use 301 or 302 redirects for a permanent site move?

Use 301 redirects for pages that have moved permanently, because that status code tells search engines the new address is the lasting one. Reserve 302 redirects for short-term changes. Using 302 for a permanent migration sends the wrong signal and can slow the transfer of ranking value.

Do I need to redirect every old URL?

Redirect every old URL that has search visibility, traffic, conversions or external links. Pages with none of these can be retired deliberately, and returning a 404 or 410 for them is correct. Make that decision per URL from your inventory data rather than redirecting everything by default.

How long should redirects stay in place?

Keep redirects in place as long as the old URLs still attract traffic or links from other sites. Removing them early can break links that external sites still point to. Review the redirect list periodically, and only retire rules after the logs show that the old URLs are no longer requested.

Can I migrate a website without a redirect map?

You can migrate without a redirect map, but search engines and visitors will then reach removed URLs and lose the signals those pages earned. Every changed URL needs either a destination or an intentional removal status. The map is the record that proves each of those decisions was made deliberately.

Sources

Tags: website migration redirect map 301 redirects SEO URL structure site migration technical SEO web development

Need Help Implementing This?

Our team can help you put these insights into practice. From AI automation to custom software development, we build solutions that deliver real results.

Send a project request