What this template is for
Use the map whenever URLs change: a new e-commerce or content platform, a new category structure, a move from one domain to another, HTTP to HTTPS, a subdomain folded into a folder, or a content consolidation that merges or removes pages. A redesign that keeps every URL needs no redirect map, though it still needs the pre-launch checks; a migration that changes URLs needs both.
The map sits among three other documents. The site migrations guide explains the rules — one to one, permanent codes, no chains, resources included, deliberate removals, tested before launch — and the sequence around them. The website migration checklist verifies the whole move, from baseline to post-launch monitoring, and checks the map as one of its items. The technical SEO audit grades a site's health afterwards and will find what the map missed. The map itself is the artefact: the one table developers implement, testers run and the SEO lead signs off.
The map is also where the new site architecture becomes visible. When many old URLs map to one new page, the structure is consolidating; when old URLs have no equivalent, content is being lost. Both can be right, but both should be decided on purpose and recorded in a row, not discovered in the error logs a month after launch.
The template
Build the map as a spreadsheet with one row per old URL and these columns. Keep it under version control or in a shared sheet with change history, because it will be edited until the day of launch.
| Field | What to write |
|---|---|
| Old URL | The full old URL, including pages, files, images and parameter URLs that have links or rankings, and where it was found |
| Signals | Organic clicks and impressions, external linking domains, conversions or revenue, and the priority they give the row |
| New URL | The single equivalent new URL, written in full |
| Match type | Exact (same content at a new address), equivalent (same purpose, changed page), consolidated (several old pages into one) or removed |
| Status code | A permanent redirect, or a gone response for a removal, with the reason for any exception |
| Chain check | Any existing redirect from or to the old URL, resolved so the old URL reaches the final destination in one hop |
| Canonical and sitemap | The new URL's self-referencing canonical, its indexability, and whether it is listed in the new XML sitemap |
| Test result | The status and destination returned on staging and again on the live site, with the date of each test |
| Owner and notes | Who decided the row, any exception and its reason, and links to related tickets |
How to fill each section
Old URL
Build the list from every source, because each one alone misses pages: a full crawl of the old site, the current XML sitemap, server logs showing URLs that receive requests but are not linked, Search Console's pages with impressions, and the backlink tools' list of linked URLs. The migrations guide treats the union of these lists as the inventory. Include images, PDFs and other files that carry links or rankings of their own, and the parameter URLs that rank. Record where each URL was found, because a URL that appears only in the logs is often an old campaign link still in use. Paid campaigns add rows of their own, and the guide to Google Ads during a website migration lists the final URLs, link assets, feed links and conversion pages to export from the ad account and Merchant Center.
Signals
For each row, record what the URL earns: organic clicks and impressions from Google Search Console over a period long enough to cover seasonality, the external domains linking to it, and any conversions or revenue it drives. The signals set the priority. High-signal rows are mapped first, checked by a person rather than a rule, and tested individually. Low-signal rows can be mapped by pattern, but they still need a destination or a deliberate removal, because links from outside never update.
New URL
Write one destination per old URL, chosen because it serves the same purpose. Redirecting many old pages to the homepage or a top category does not transfer their value: as the migrations guide explains, drawing on Google's documentation on site moves listed in the sources, a redirect to a page that does not match is treated much like a missing page. Where the new site has no equivalent, the right destination is the closest page a visitor would accept as the answer, or no destination at all.
Match type
Classify every row. Exact means the same content at a new address; equivalent means the same purpose on a page whose content or template changed; consolidated means several old pages merged into one; removed means the content is going. The classification is the summary the project lead reads: a map that is mostly consolidations or removals describes a different site, and the site architecture guide is where that change should have been decided. Removals need a named decision-maker in the owner column.
Status code
Use a permanent redirectA permanent server redirect telling browsers and search engines a URL has moved for good. (301 or 308) for every row that has a destination, so that signals consolidate on the new URL; Google's page on redirects, listed in the sources, describes how each code is treated. Temporary codes delay consolidation and belong only to genuinely temporary moves. For removals with no equivalent, return a gone or not-found response and remove the URL from internal links and sitemaps. Write the reason beside any exception so a later reviewer does not "fix" it.
Chain check
Look up every existing redirect on the old site, from the server configuration, the platform's redirect manager and any plug-ins. Where an old URL already redirects, or a new destination is itself redirected, rewrite the row so the original URL goes to the final destination in one hop. Chains slow crawling, lose some requests along the way and multiply with every future migration, so resolving them now is the cheapest it will ever be.
Canonical and sitemap
Confirm that each new destination is indexable, carries a Canonical TagTells search engines the preferred URL. pointing to itself rather than to an old URL or a parameter version, and appears in the new XML sitemapA file listing the URLs a site wants search engines to discover and crawl.. A redirect to a page that canonicalises elsewhere, or that is blocked or noindexed, sends mixed signals. The crawling and indexing guide explains how these signals combine. Update internal linksA link from one page to another page on the same website. to the new URLs as well, rather than relying on redirects inside the site.
Test result
Test every row on staging with a script that requests the old URL and records the status code and final destination, then compare both with the map. Repeat on the live site immediately after launch. Record the date and the result per row, so the sign-off is evidence rather than a feeling. For a domain move, the launch also includes Search Console's Change of Address tool, listed in the sources; the migration checklist places it in the launch sequence.
Owner and notes
Name the person who decided each row, particularly for consolidations, removals and exceptions. Link any ticket the row depends on, such as a new page still being written. After launch, the notes column records fixes for URLs found in the logs that the map missed, which become rows in their own right.
Filled example
Below, a section of the map is completed for an invented business. The URLs, figures and names are illustrative.
Harwick Tiles: replatforming with a new category structure
Harwick Tiles is moving from an old shop platform, where categories sat under /shop/category/ with
product IDs in the URL, to a new platform that groups tiles by room and material. The inventory found 3,840
old URLs; 412 had organic clicks in the last 16 months or at least one external linking domain. The rows
below are a sample.
| Old URL | Signals | New URL | Match type | Status code | Chain check | Canonical and sitemap | Test result | Owner and notes |
|---|---|---|---|---|---|---|---|---|
| /shop/category/metro-tiles | 18,200 clicks, 34 linking domains, £41,000 revenue: priority A | /tiles/wall/metro | Exact | 301 | None | Self-canonical, in sitemap | Staging 301 to correct URL, 14 May; live 301, 2 June | Imogen Hart (SEO) |
| /shop/category/kitchen-wall-tiles and /shop/category/splashback-tiles | 6,100 and 1,900 clicks: priority A | /tiles/kitchen | Consolidated | 301 (both) | Splashback URL already redirected from /splashbacks; that rule rewritten to /tiles/kitchen | Self-canonical, in sitemap | Staging pass both, 14 May; live pass, 2 June | Imogen Hart; merchandising agreed the merge 2 May |
| /shop/product/4471-carrara-hexagon | 2,300 clicks, 6 linking domains: priority A | /tiles/floor/marble/carrara-hexagon | Exact | 301 | None | Self-canonical, in sitemap | Staging pass, 15 May; live pass, 2 June | Imogen Hart |
| /shop/category/metro-tiles?colour=sage | 940 clicks: priority B | /tiles/wall/metro/sage | Equivalent | 301 | Parameter rule tested separately for all 11 colours | Self-canonical, in sitemap | Staging fail: rule sent to /tiles/wall/metro, fixed 16 May; retest pass; live pass, 2 June | Dev ticket WEB-288 |
| /guides/how-to-grout-tiles.pdf | 410 clicks, 12 linking domains: priority B | /guides/how-to-grout-tiles | Equivalent | 301 | None | Self-canonical, in sitemap | Staging pass, 15 May; live pass, 2 June | PDF content rebuilt as a page; Tom Briar (content) |
| /shop/category/clearance-2019 | 0 clicks, 0 linking domains: priority C | none | Removed | 410 | None | Removed from sitemap and internal links | Staging 410, 15 May; live 410, 2 June | Imogen Hart: no equivalent, range discontinued |
In the first week after launch, server logs showed 37 old URLs returning not-found responses that were not in the map, mostly old email campaign links; they were added as rows and redirected within two days.
Common mistakes
- Inventory from one source. A crawl misses unlinked pages, and logs miss pages nobody visited this month. Combine the crawl, sitemap, logs, Search Console and backlink data.
- Everything to the homepage. Mass redirects to a top page discard the value of every URL that did not match. Map one to one, to the equivalent page.
- Temporary codes. Temporary redirects delay consolidation of signals. Use permanent codes for permanent moves.
- Chains left in place. A redirect from an earlier migration plus a new one makes a chain. Resolve every row to its final destination.
- Files forgotten. PDFs and images with links or rankings lose them if their URLs change unmapped. Include them in the inventory.
- Removals by accident. A URL with no row is a removal nobody decided. Every removal needs an owner and a reason.
- Destinations that canonicalise elsewhere. A redirect to a page with a wrong canonical, a noindex or no sitemap entry sends mixed signals. Check the destination as well as the redirect.
- Testing a sample. Every row should be requested on staging and again on the live site. Record the results in the map.
- Switching the old redirects off. External links never update. Keep the redirects live long after the launch, as the migration guide describes.
The other planning documents are in the templates library.
