Site Migration SEO Checklist: Preserving Rankings AND AI Citations

Site Migration SEO Checklist: Preserving Rankings AND AI Citations

Bottom line: A safe site migration seo project rests on three things: a complete old-to-new URL map, 301 redirects on every URL with equity, and identical content, metadata, and schema on the new pages. Do that and rankings usually recover in four to six weeks. But AI citations run on a separate clock, and most checklists ignore them entirely.

Here's the uncomfortable part most migration guides skip: doing everything right still doesn't guarantee a clean move. SALT.agency analyzed domain migrations and found only 27% recovered their organic visibility within 90 days. That's not a redirect problem. It's a trust-rebuilding problem, and in 2026 the trust you're rebuilding lives in two places at once - Google's index and the entity graphs behind ChatGPT, Perplexity, and AI Overviews.

Key Takeaways:

• —301 redirects on every equity-bearing URL are non-negotiable - skipping them is the single biggest cause of permanent traffic loss.
• —Blue-link rankings typically stabilize in 4-6 weeks; AI citations often lag 3-6 months because they depend on recrawl cycles and model updates, not just redirects.
• —Entity and source signals can break independently of your redirects - a page can rank fine and still vanish from AI answers if schema, brand mentions, or external citations point to the wrong place.
• —Benchmark before you move: rankings, GSC data, AND your current AI-citation share. You can't prove recovery against a baseline you never captured.
• —Ahrefs found brand mentions correlate with AI citations far more strongly than backlinks - so migrating a domain without updating your mention footprint quietly starves the AI layer.

What is a site migration in SEO?

A site migration is any change significant enough to force search engines and AI crawlers to re-evaluate your site. That's broader than most people assume. Swapping domains is the obvious one, but a migration also covers moving from HTTP to HTTPS, changing your CMS or platform, restructuring URLs, consolidating multiple sites, or a full redesign that rewrites your HTML and internal linking.

The common thread: you're changing the address, the structure, or the signals that engines used to identify and rank you. The bigger the change, the longer the reassessment. A protocol switch is minor. A domain-plus-platform-plus-redesign move - the kind agencies get called in to rescue - is the hardest variety, because every trust signal moves at once.

The reason site migration seo matters so much is asymmetry of risk. A great migration gets you roughly back to where you were, maybe slightly better. A bad one can cut organic traffic 30-60%, and in more severe cases 30-80%. The upside is capped; the downside is not.

Why do rankings drop after a site migration?

Even a textbook migration usually produces a temporary dip, and understanding why keeps you from panicking in week two and "fixing" things that aren't broken.

Google doesn't instantly transfer ranking signals across a redirect. It crawls the new URL, compares it to the old one, tests it in live results, and gradually recalibrates. Google's John Mueller has repeatedly said ranking stabilization after a migration can run anywhere from a few weeks to a few months depending on scale. During that window, fluctuation is the system working, not failing.

The drops that don't recover almost always trace to a handful of avoidable mistakes:

• —Missing or chained redirects - every hop dilutes equity, and un-redirected URLs simply drop out.
• —Changed content or metadata - if the new page isn't recognizably the same page, Google re-evaluates it from a weaker position.
• —Dropped schema - structured data that vanished in the rebuild removes the machine-readable context engines relied on.
• —Blocked crawling - a robots.txt or noindex left over from staging is the classic launch-day disaster.
• —Internal links still pointing at old URLs - forcing crawlers through redirect chains instead of clean paths.

None of these are exotic. They're the reason Search Engine Land's migration guidance hammers on redirect mapping and pre-launch QA over everything else.

The site migration SEO checklist: 12 steps

Run this in order. The pre-launch work is where migrations are won - the post-launch phase is mostly monitoring and confirming what you already set up correctly.

1. Benchmark everything. Export current rankings, GSC impressions and clicks by URL, GA4 organic sessions and conversions, and - critically - your current AI-citation footprint. This is your recovery baseline.
2. Crawl the old site fully. Use Screaming Frog to capture every indexable URL, its metadata, and its internal links.
3. Build the URL map. Map every old URL to its exact new equivalent. No lazy redirects to the homepage - match intent one-to-one.
4. Preserve on-page content and metadata. Keep titles, H1s, body content, and canonical logic as close to identical as the redesign allows.
5. Port every schema type. Article, Product, FAQ, Organization, Breadcrumb - whatever the old pages had, the new ones need too.
6. Implement 301 redirects. Server-side, single-hop, old URL to new URL. Never 302 for permanent moves.
7. Fix internal links. Update them to point at final destination URLs, not through redirects.
8. Stage and QA privately. Password-protect staging so it never gets indexed. Check for stray noindex tags.
9. Launch, then immediately verify crawlability. Confirm robots.txt allows crawling and no noindex shipped to production.
10. Submit the new XML sitemap in Google Search Console, and file a Change of Address for domain moves.
11. Update external signals. Refresh Google Business Profile, major citations, social profiles, and request updates on high-value backlinks pointing to old URLs.
12. Monitor weekly for 8-12 weeks. Watch GSC coverage, rankings, redirect health, and AI-citation share against your baseline.

Steps 1, 5, and 11 are the ones generic checklists treat as afterthoughts - and they're exactly the ones that protect your AI visibility. That's where the next section lives.

What most migration checklists miss: the AI-citation layer

Every checklist you'll find online optimizes for one thing: passing link equity through redirects so blue-link rankings survive. That's necessary. It's also incomplete, because AI engines don't cite you the way Google ranks you - and they can drop you even when your rankings hold.

Here's the mechanism most guides never explain. AI citation depends on entity recognition and source trust, and both can break independently of your redirects.

How entity signals break during a migration

An AI engine doesn't just see a URL. It sees an entity - your brand as a node in a knowledge graph, connected to the topics you're known for, the mentions of you across the web, and the structured data that spells out who you are. A 301 redirect moves the URL. It does not automatically move the entity.

Three things quietly detach during a careless migration:

• —Structured-data drift. If Organization schema, sameAs links, or entity markup change or disappear in the rebuild, AI models lose the machine-readable thread connecting the new page to your known entity. Koozai's analysis of migrations in the AI-search era notes AI Overviews are more sensitive to structure, entities, and technical clarity than classic rankings are.
• —Mention-footprint lag. Ahrefs' brand-mention research found brand web mentions correlate with AI citation rates roughly three times more strongly than backlinks do. Those mentions - in articles, forums, roundups, Wikipedia - still name your old domain. Redirects don't rewrite third-party text.
• —Training-data inertia. Models learn from crawled snapshots. Even after they recrawl your new site, the association between your brand and your old URLs persists in prior training until the next model update - which is why AI-citation recovery lags blue-link recovery by months, not weeks.

How to verify AI-citation recovery

You can't manage what you don't measure, and almost nobody baselines AI visibility before a migration. Do these four things:

1. Pre-migration: Record which prompts currently surface your brand in ChatGPT, Perplexity, and Google AI Overviews, and note the exact URLs being cited.
2. Post-migration: Re-run the same prompts weekly. Watch for two failure modes - you stop being cited, or you're cited but the link 404s or redirects.
3. Audit the entity chain: Confirm your new pages carry the same (or richer) Organization and Article schema, and that sameAs points to live profiles. This is the fastest technical fix, and you can flag gaps automatically with SEO Magics' AI SEO Audit.
4. Track citation share over time rather than checking once. Recovery here is a slope, not a switch - a topic we go deep on in how to track your brand's AI citation share.

The takeaway an SEO lead needs: treat your entity as a migration asset in its own right. Redirect the URLs, yes - but also re-assert the entity through schema, refresh the highest-authority mentions, and measure citations on their own timeline. This is the layer that separates a migration that "kept its rankings" from one that also kept its AI visibility.

How long does it take to recover rankings and AI citations?

Recovery isn't one timeline - it's two, running at different speeds. Blue links move on Google's recrawl-and-recalibrate cycle. AI citations move on model recrawl plus training-data cycles, which are slower and less predictable.

Two numbers keep the timeline honest. On the blue-link side, only 27% of domain migrations recover inside 90 days - so if you're not fully back by month one, you're in normal territory, not crisis territory. On the AI side, expect a longer tail: the citation layer sits on top of traditional SEO and rebuilds last.

One practical rule we hold clients to: if traffic hasn't started climbing back by week four, investigate immediately. Drops caught inside the first four weeks - while the algorithm is still processing the move - are the ones that get corrected fastest.

What tools do you need to run a migration safely?

You don't need a huge stack, but you need coverage across three jobs: crawl, redirect QA, and visibility tracking.

• —Screaming Frog - full crawls before and after launch, redirect-chain detection, metadata diffing.
• —Google Search Console - indexation status, Change of Address, coverage errors, real ranking data.
• —Ahrefs or Semrush - ranking benchmarks, backlink profiles, and finding which external mentions still point at old URLs.
• —An AI-citation checker - to baseline and track whether AI engines still surface you. Running an AI SEO Audit surfaces broken entity signals - missing schema, dropped sameAs, stale mentions - that a redirect audit alone will never catch.

The gap in most stacks is that last one. Teams QA their redirects obsessively and never check whether they still exist inside an AI answer. If AI-referred traffic matters to your pipeline, that blind spot is expensive - which is the whole thesis behind our AI SEO service and the deeper technical work in our technical SEO audit breakdown.

How We Assessed This

The framework in this article comes from running migrations the way an agency does under retainer, not from a one-time checklist. We build recovery baselines across three layers - classic rankings and GSC data, technical crawl state via Screaming Frog, and AI-citation share tested through direct prompting of ChatGPT, Perplexity, and Google AI Overviews. During a migration we monitor all three weekly against that baseline rather than trusting a single traffic graph, because blue-link and AI signals recover on different clocks and a healthy ranking chart can hide a collapsed citation footprint. The external figures here are drawn from published research - SALT.agency's domain-migration recovery analysis, Ahrefs' brand-mention correlation study, and Google's own guidance from John Mueller - and we've flagged where recovery timelines are ranges, not guarantees. Our view on the AI-citation layer reflects a pattern we see repeatedly on growth-stage sites: the entity signals break quietly, weeks after the redirects test clean, and only show up if you were measuring them before the move.

Frequently Asked Questions

Do 301 redirects preserve SEO rankings?

Largely, yes - 301 redirects pass the majority of ranking signals from old URLs to new ones, and they're the single most important technical step in any migration. But "pass" isn't instant or total. Google recrawls and recalibrates over weeks, a temporary dip is normal, and redirects alone don't rebuild the entity and mention signals AI engines rely on.

How long does a site migration take to recover?

For a well-executed migration, blue-link rankings usually stabilize in four to six weeks. AI citations take longer - often three to six months - because they depend on recrawl cycles and model updates. Data shows only about a quarter of domain migrations fully recover within 90 days, so a slower ramp isn't necessarily a failure.

Will a site migration hurt my Google rankings?

Temporarily, almost always - a short dip while Google reassesses is expected even when everything is done correctly. Lasting damage comes from mistakes: missing redirects, changed content, dropped schema, or blocked crawling. Get those right and you recover; get them wrong and 30-60% traffic loss is a real risk.

How do I keep my AI citations after moving domains?

Redirect the URLs, then re-assert your entity: port all Organization and Article schema, keep sameAs links live, and refresh the highest-authority third-party mentions that name your old domain. Baseline your AI-citation share before the move and re-test the same prompts weekly so you can prove recovery instead of guessing.

What's the biggest mistake in a site migration?

Redirecting everything to the homepage instead of mapping each old URL to its true equivalent. It's fast, and it destroys relevance signals - the new homepage can't inherit the ranking context of fifty different deep pages. A close second is shipping a leftover noindex from staging to production on launch day.

Should I migrate my site if I rank well now?

Only with a full plan and a reason worth the risk - the upside of a migration is modest and the downside is large. If you must move, treat it as a project with a benchmark, a redirect map, schema parity, and AI-citation tracking, not a launch you QA afterward. Google's own migration documentation and a disciplined checklist are the difference between a routine move and a recovery project.

Planning a migration? Get the entity layer right before you move

Most migrations are lost in the two weeks before launch - in the redirect map that wasn't complete and the schema that wasn't ported. The AI-citation layer is lost even more quietly, because almost nobody baselines it.

If you want a second opinion before you pull the trigger - or you've already migrated and traffic hasn't come back - start by running your site through the SEO Magics AI SEO Audit to catch broken entity and citation signals, then read more migration and AI-search playbooks in the SEO Magics journal. When you're ready for hands-on help protecting rankings and AI visibility through the move, book a strategy call and we'll pressure-test your migration plan before it costs you traffic.