Theory Road.

Website Migration SEO Checklist: How to Keep Your Rankings.

Migrations lose traffic by accident, one category of pages at a time. This is the checklist that keeps rankings through a redesign, a platform change or a new domain.

By Theory RoadSeptember 11, 202610 min read

A website migration is the most common way a healthy business loses a large share of its organic traffic in a single week. Not a penalty, not a competitor, not an algorithm update. A redesign, a platform change from WordPress to Shopify or to a headless build, a domain change, or a URL cleanup that nobody checked. This website migration SEO checklist is the method we use to keep rankings through the move. It is not complicated. It is tedious, and the tedium is the point: every step exists because skipping it has cost someone their traffic.

Why migrations lose traffic.

Search engines rank URLs, not businesses. Every page that ranks today carries a history: links pointing to it, content Google has already evaluated, signals accumulated over years. When the migration changes the URL, removes the page, rewrites the title, drops the structured data or blocks the crawler, that history is either transferred through a correct redirect or it is lost. Most migrations lose it by accident, one category of pages at a time, and nobody notices for a month because traffic still arrives from the pages that happened to survive.

The second cause is that migrations are run by people optimizing for the new site. The designer wants cleaner navigation, the developer wants tidy URLs, the marketing lead wants fewer pages. Each instinct is reasonable and each deletes something that ranks. The checklist puts a person in the room whose job is the old site.

Crawl the old site before anything is built.

Before a single page is designed, run a full crawl of the existing site and keep the export somewhere permanent. This is the record of what exists, and once the old site is gone it cannot be recreated. The crawl should capture every URL with its status code, title, description, H1, canonical, indexability, word count, internal links in, structured data types, images, and hreflang if the site has language versions.

Then combine the crawl with three other sources: the Search Console page-level export for the last sixteen months, the analytics landing page report for the same period, and a backlink export. A URL that appears in any of those four sources is a URL the migration must account for. Sites regularly rank for pages that are in nobody's memory. The exports find them; the redesign brief does not.

Map every URL.

The URL map is the central document of the migration. One row per old URL, a column for the new URL it becomes, a column for the action (keep, redirect, consolidate, remove), and a column for who decided. Nothing gets a blank cell. If a page is being removed, the row says where its traffic and links go instead, which is usually the closest surviving parent, not the homepage.

Build the map from the crawl, not from the new sitemap. Starting from the new site's page list and matching old pages to it guarantees that anything without an obvious new home is forgotten. Start from the old list and force a decision on every row. For large sites with patterned URLs, write the rules first and verify them against a sample of real URLs, because patterns always have exceptions.

Preserve the pages that rank.

Sort the URL map by organic clicks and by referring domains. The pages at the top of either list are the ones that will hurt if they break. For those pages the default action is keep: same URL if at all possible, same title, same H1, same core content, same internal links pointing at them. Design can change. The words that earned the ranking should not, at least not on launch day. Rewrite them later, one at a time, once the migration has settled.

Consolidation is where good intentions do the most damage. Merging five thin pages into one strong page is often correct, and it is a decision that must be made per page with the numbers in view. If a page has clicks or links, it earns a redirect to the page that absorbs it, and the absorbing page has to actually cover the topic or the ranking transfers to nothing.

Redirects: the rules that hold.

  • Every changed or removed URL returns a 301 to its mapped destination. Not a 302, not a JavaScript redirect, not a page that says the content moved.
  • Redirects land on the final URL in one hop. Chains leak signal and time out crawlers, so test for them after the rules are written.
  • Redirect to the equivalent page, never to the homepage in bulk. Mass homepage redirects are treated as soft 404s and the history is lost.
  • Redirects from previous migrations are carried forward and repointed to the newest destination, and protocol and host variants all resolve in one hop to the canonical form.
  • Image and PDF URLs get redirects too if they rank or carry links. Image search traffic is real and almost always forgotten.

Titles, metadata and structured data.

Content management systems regenerate titles and descriptions from templates by default, and a migration is when that default silently overwrites years of tuned metadata. The URL map should carry the old title, description and H1 for every kept page, and the new site should be checked against them before launch. Where the new template must differ, the change is deliberate and logged, not a side effect.

Canonical tags deserve their own check. A new platform may set canonicals to a staging domain, to a parameterized version of the URL, or to the homepage across the board, and any of those will deindex the site within days. Confirm that every indexable page has a self-referencing canonical on the production domain and that the robots meta tag on production says index, follow. The most common catastrophic migration error is a noindex or robots.txt disallow carried from staging to production, and it is caught by reading one line.

Structured data is the third thing that disappears. Product, Article, LocalBusiness, FAQ and BreadcrumbList markup are often generated by a plugin or theme on the old platform and do not exist on the new one. Inventory the types in the old crawl, rebuild them on the new templates, and validate a sample before launch.

Search Console and analytics continuity.

Search Console is both the early warning system and the record of what happened, so it is set up before launch, not after. If the domain is changing, verify the new property as a domain property before the switch, keep the old property, and use the change of address tool once redirects are live. If the domain is the same, the work is submitting the new sitemap and removing the old one. Bing Webmaster Tools gets the same treatment.

Analytics continuity has the same deadline. GA4 and Google Tag Manager have to be installed on the new site with the same measurement ID and the same conversion events firing on the same actions, verified in debug mode on staging. Call tracking number insertion, the Meta pixel and any other ad pixels, and any server-side tagging must be carried over and tested. A migration that breaks conversion tracking for two weeks also breaks every smart bidding campaign feeding on that data.

The website migration SEO checklist for launch day and the day after.

Crawl staging against the URL map.
Every old URL should resolve through the staging redirect rules to its mapped destination in one hop with a 200. Fix the exceptions before anyone sees the new site.
Confirm indexability on production.
Check robots.txt, the robots meta tag and canonicals on production within the first hour. This is the check that prevents the worst outcome.
Submit sitemaps and request indexing on top pages.
New XML sitemap in Search Console and Bing Webmaster Tools, old sitemap removed, manual indexing requests on the highest-value pages.
Verify tracking end to end.
Fire a real conversion on the live site and confirm it lands in GA4, the ad platforms and call tracking.
Re-crawl production the next morning.
Crawl the new site and the old URL list. Compare against the pre-migration export for status codes, titles, canonicals, structured data and internal link counts.
Watch Search Console daily for two weeks.
Coverage errors, soft 404s and redirect errors. Then page-level clicks against the baseline, weekly, for three months.
Before and after: the checks that catch most migration damage
CheckBefore launch (staging)Day after launch (production)
Old URL listEvery URL resolves in one hop to its mapped destinationSame crawl against production; zero chains, zero 404s on ranking pages
IndexabilityStaging is noindexed and blocked from crawlersrobots.txt open, meta robots index, self-referencing canonicals
Titles and H1sKept pages match the old crawl unless a change was loggedSpot check top pages by clicks; no template overwrites
Structured dataEvery type from the old crawl present and valid on new templatesRich results test on a sample; Search Console enhancement reports
TrackingGA4, GTM, pixels and call tracking fire in debug modeA real conversion recorded in every system
Search ConsoleNew property verified, sitemap ready, change of address preparedSitemap submitted, old removed, coverage report clean

Common ways migrations destroy traffic.

These keep happening because each one is owned by a different person who did not know it was theirs. Staging noindex shipped to production. Redirects to the homepage. Trailing slashes or case changed without redirects. Category pages removed because the new navigation did not need them. A blog moved to a subdomain. Hreflang dropped. Internal links still pointing through a chain. Structured data lost with the old theme. Content rewritten on launch day on the pages that ranked. Conversion tracking silently broken, so the drop is discovered from revenue rather than from a report.

None of these are hard to prevent. Each is hard to notice after the fact, because the symptom is a slow decline across many pages rather than a single alarm. The checklist works because it assigns every one of them to a check with a date. If you are planning a move and want it run this way, our SEO and web design and development teams run migrations together, with the old site as the shared brief. You can also reach us for a pre-migration review of an existing plan.

A migration is a promise to the search engine that every page it trusted still exists somewhere. The redirect map is the written form of that promise, and the day-after crawl is how you find out whether you kept it.

How long does it take rankings to recover after a website migration?

With redirects in place, metadata preserved and no indexability errors, most sites see fluctuation for a few weeks while the index updates and then settle near the previous level. Recovery takes far longer when redirects were added late.

Do we need to keep the old domain after a domain change?

Yes, indefinitely. The redirects from the old domain to the new one carry the link history and must keep working. Letting it lapse breaks every link that still points to it.

Should we change content and URLs at the same time as the redesign?

Change as little as possible on launch day. Keep URLs where you can, keep titles and core content on ranking pages, and schedule content improvements for after the migration has stabilized so any drop can be attributed to one cause.

What is the minimum website migration SEO checklist for a small site?

Crawl and export the old site, map every URL, write and test 301 redirects, confirm production is indexable with correct canonicals, submit the new sitemap in Search Console, verify tracking, and re-crawl the day after. Small sites lose traffic for the same reasons large ones do.

Work with us

Let’s talk about what’s next.

A short note on where the business is and where it needs to go. A senior partner replies within one business day.

t@theoryroad.com · (512) 222-7805