To create a redirect map, collect every old URL from several data sources, score each one by traffic, links and conversions, choose the closest new page for it by hand, then test the full map on staging before launch.
By Kristijan GjorgjioskiIndependent SEO consultantPublished 18 October 2026 · 8 min read
Most migration disasters I'm asked to fix share one cause.
The redirect map was built in launch week, from the new sitemap, by someone who never saw the old site's data.
That's backwards. A good map starts with the old site.
Below is how to create a redirect map step by step, using the process behind my SEO migration projects. Let's get into it.
A redirect map is a spreadsheet with one row per old URL and exactly one destination per row.
That's the core.
The columns around it are what make the map usable by a developer and checkable by you. Here's the layout I use:
| Column | What it holds | Why you need it |
|---|---|---|
| Old URL | The full old address, protocol included | The source of each rule |
| Clicks | Organic clicks or sessions over 12 months | Sets the priority |
| Referring domains | Sites linking to that URL | Shows the links at stake |
| Conversions | Enquiries or sales that started on the page | Shows the revenue at stake |
| Action | Keep, merge, retire or new | Records the decision |
| New URL | The final destination | The target of the rule |
| Status | 301, 308 or 410 | Tells the developer what to build |
| Notes | Why this destination was chosen | Settles arguments later |
| Tested | Pass or fail on staging | Proves the rule works |
Before you start, get access to:
Keep the map in one shared sheet.
Copies sent by email drift apart within days.
Your goal here is one list of every old URL that has ever mattered.
No single source holds them all. So you combine several:
Google's own guidance on site moves suggests a similar mix: sitemaps, top URLs from logs or analytics, and pages from the Links report.
Paste everything into one column.
Then normalise it: same protocol, same host, consistent trailing slashes. Remove exact duplicates.
Don't skip the long tail either. PDFs, linked images, old campaign pages and parameter URLs all collect links over the years.
Pro tip: If the site has been rebuilt before, export its existing redirect rules too. Those old source URLs must now point at the new destinations, or you'll create chains.
Not every row deserves the same attention.
This step tells you where to spend your time.
Next to each URL, add three numbers: clicks, referring domains and conversions.
Then sort. In most audits I run, a small slice of URLs carries most of the traffic and links.
Those rows get mapped by hand and checked twice.
Everything else still gets a decision. It just doesn't need a debate.
Here's the simple tiering I work with:
| Tier | Rule of thumb | How it gets mapped |
|---|---|---|
| 1 | Real traffic, links or conversions | Chosen by hand, reviewed by a second person |
| 2 | Some impressions or a few links | Chosen by hand, spot-checked |
| 3 | Nothing in any source for a year | Pattern rule or a deliberate 410 |
One warning: no traffic doesn't mean no value.
A page with zero visits but a link from a trade body is still worth a redirect. My guide on judging backlink quality shows which old links are worth protecting.
And if you don't trust the old site's tracking, check it before you sort by it.
A GA4 audit checklist will show you whether broken tags are hiding pages that convert.
You can't map to pages you haven't listed.
Crawl the staging site and export every indexable URL with its title and main heading. That becomes your destination list.
While you're there, check two things.
First, staging must be blocked from search engines. Password protection is the safest option, because a robots.txt rule alone won't stop a linked URL from being indexed.
Second, the URL structure must be final.
If folder names are still changing, your matches will need redoing.
No staging yet? Use the agreed sitemap or content plan as your destination list, and mark the map as provisional.
This is also a good moment to run your usual technical SEO checklist against the new templates.
This is the step that decides whether your rankings survive.
For every tier 1 and tier 2 row, ask one question: which new page would satisfy the person who landed on the old one?
Usually the answer is obvious. Sometimes it isn't:
When a call is hard, open the old URL's queries in Search Console.
The new page built for those same searches is the right home.
Merging old pages that chased one query also clears up keyword cannibalisation, so note those merges deliberately.
Write a one-line reason in the Notes column for every match that isn't obvious. When someone questions it three months later, you'll have the answer.
Spreadsheet lookups can do a first pass on matching slugs.
Treat their output as suggestions, though. A matching slug with different content is a wrong match.
Some old URLs won't have a sensible new home.
That's fine, as long as you decide on purpose. Here are your options:
| Situation | What to use | Why |
|---|---|---|
| A real equivalent exists | 301 or 308 redirect | Moves the old URL's signals to the new one |
| Content removed, nothing similar | 410 or 404 | Says the page is gone instead of faking a match |
| A clean pattern, such as /blog/2019/post/ to /blog/post/ | Pattern rule | Saves time where pages truly correspond |
| A move you will reverse | 302 or 307 | Only for temporary changes |
What you shouldn't do is send every unmatched URL to the homepage.
Google says that pointing many old URLs at one irrelevant page, such as the home page, may be treated as a soft 404. You lose the redirect and confuse visitors.
Use permanent redirects.
Google's redirect documentation explains that a 301 or 308 tells its indexing systems the target should be canonical. A 302 or 307 doesn't.
Then flatten chains.
If /a already redirects to /b, and /b now moves to /c, update /a to go straight to /c. Google advises redirecting to the final destination directly.
For pages you're thinking of cutting, my content pruning guide covers how to choose between merging, improving and removing.
An untested map is a wish list.
Load your old URLs into the crawler in list mode, with the hostname swapped to staging. Then check every row for:
Fix the failures, retest and repeat until every tier 1 row passes.
On launch day, run the same list against production.
Rules that worked on staging break more often than you'd expect: a different server config, a CDN rule or a plugin that overrides the redirect file.
After launch, point your navigation and body links straight at the new URLs. My internal linking strategy guide covers that clean-up.
Google recommends keeping redirects for at least a year. I keep them indefinitely.
The wider go-live checks, from noindex tags to tracking, are in my website migration SEO checklist.
These turn up again and again when I review maps written by someone else:
None of these are hard to avoid.
They happen because the map gets left until the end, when nobody has time to check it.
A rebuild onto an entirely new set of URLs is survivable when the map is done properly. This hazmat storage relaunch changed every address on the site and still regained its rankings.
Server rules, rendering and crawl problems outside the map belong to technical SEO.
The best time to build a redirect map is while the new sitemap is still being argued about.
Steps 1 and 2 don't need the new site at all. And they often change which pages it should keep.
So here's your first action: take your old site's 50 most valuable URLs and write down where each will land.
Any URL you can't answer for is where your map starts.
If you'd rather have the inventory, mapping and launch testing run alongside your developers, that's the core of my migration support.
A brochure site with a few dozen pages can be mapped in an afternoon. A shop or publisher with thousands of URLs takes days, and most of that time goes on the long tail: old campaign pages, PDFs and parameter URLs. Pattern rules speed up page types that genuinely match, but every high-value URL still needs a person to pick its destination.
No. Google says that sending many old URLs to one irrelevant destination, such as the home page, can be treated as a soft 404, so the redirect does nothing for you. Send each old page to its closest equivalent. If nothing on the new site comes close, let the old URL return a 404 or 410 instead.
Only for moves you plan to reverse. Google treats 301 and 308 redirects as a signal that the target URL should be canonical, while 302 and 307 redirects are not used that way, so the old URL can linger in search results. For a permanent migration, a redirect map should specify permanent redirects throughout.
Not a full one. A redesign that keeps every URL needs no redirect rules, though you should still compare templates, copy and internal links against the old pages. The moment anything in the address changes, including the domain, protocol, folder names or trailing slashes, you need a map for every affected URL.
About the author
Kristijan Gjorgjioski
Independent SEO consultant · 6+ years in SEO · 60+ client websites worked on
I'm an independent SEO consultant with 6+ years in SEO. I run audits, migrations and content programmes for businesses in Europe, the UK and the USA, and I build the automation behind my own research.
Book a free 30-minute call. I look at your site first and tell you plainly what I would fix, and in what order.