Replatforming SEO Checklist 2026: Save Rankings

Replatforming a site (moving from one CMS, e‑commerce platform, or hosting environment to another) is one of the highest-risk technical SEO endeavors. A study of 892 domain migrations found that average recovery took ~523 days, but the fastest recoveries (19–23 days) were achieved through 1:1 redirect accuracy and content parity (DigitalApplied). This checklist gets you to that fast‑recovery end of the curve by covering every essential step for a 2026 migration.

Pre‑Migration: Benchmark and Map Everything

Before touching any infrastructure, capture a full snapshot of your current site.

1. Export priority URL metrics – organic sessions, conversions, and conversion rates from Google Analytics 4 (GA4) and Google Search Console (GSC) for your top pages. Also export keyword rankings, backlink profiles, and current sitemaps (Lumar).
2. Create a complete URL inventory – use Screaming Frog to crawl the live site as Googlebot, collecting every URL and its status code (Edwin Romero).
3. Audit existing structured data – document all schema types (Product, FAQ, Article, BreadcrumbList) and verify that they pass Google’s Rich Results Test (Waravel).
4. Record rendering behavior – note whether content is delivered via SSR, CSR, SSG, or hydration. This will be critical for parity testing later (Naturaily).
5. Benchmark Core Web Vitals at the 75th percentile using field data. Targets for 2026: LCP ≤ 2.5 s, INP ≤ 200 ms, CLS ≤ 0.1 (Naturaily).
6. Set AI citation baselines – use BrightEdge or manual searches to count how often your content appears in AI Overviews, SearchGPT, and Perplexity. The 2025 BrightEdge guidance recommends retaining at least 95% of AI citations within 60 days of migration (BrightEdge).

URL Mapping and Redirects

Every old URL must map to a single, meaningful new URL. This is the single most influential factor for recovery speed.

• Build a 1:1 301 redirect map between old and new URLs. Never bulk‑redirect to the homepage, or search engines may treat it as a soft 404 (Urllo).
• For thousands of similar‑slug URLs, AI‑based mapping can help, but always pair it with manual validation (Reddit discussion).
• Tools like Rapid301 or WISLR can generate and validate bulk redirect maps. Avoid AI‑only tools that produce inconsistent mappings.
• Remove redirect chains – each chain increases latency and may confuse crawlers. Google’s Crawl Stats report counts each request in a chain separately (GSC Help).
• Use Screaming Frog list mode or a custom Python script to validate that every redirect lands on a 200‑status final URL.
• Don’t forget non‑HTML resources – images, PDFs, CSS, and JS files need proper 301s or at least a 410 if intentionally removed.

Canonical Tags and Indexation Controls

• Self‑referencing canonicals on canonical pages. No cross‑domain drift (1Digital, May 2026).
• Canonical tags must be server‑side rendered – Google and most AI crawlers ignore JavaScript‑injected canonicals (Naturaily).
• Use noindex, follow on faceted navigation pages, internal search results, and temporary content during a staged rollout (SEOlogist).
• Never combine robots.txt disallow with a noindex – Google won’t see the noindex tag if the page is blocked (Google Search Central).
• For the staging environment, use HTTP authentication plus a noindex header. Do not rely on robots.txt alone – 50% of migration clients block staging incorrectly (SearchViu).

Robots.txt and Staging Security

• On the production site, include a Sitemap: directive in robots.txt. Limit crawl‑delay to Googlebot via GSC settings, not robots.txt (Moz).
• For staging: protect with HTTP username/password and whitelisted IPs. Keep a “real” robots.txt on staging to simulate production crawling, but never rely on it for security (SearchViu).
• If you want to appear in ChatGPT Search, allow OAI-SearchBot in robots.txt and allow its published IP ranges (Naturaily).

XML Sitemap Management

• Regenerate XML sitemaps containing only 200‑status, canonical URLs. Remove old domain URLs and parameterized filter pages (DigitalPresent).
• Host at /sitemap.xml, reference in robots.txt, and submit to GSC and Bing Webmaster Tools immediately after DNS cutover (Swanky).
• If the platform generates sitemaps automatically (Shopify, Contentstack, CrafterCMS), verify that the output excludes staging URLs and noindexed content.

Internal Link Updates

• Bulk‑update all internal links to point to final URLs, not redirected ones. Maintain anchor text consistency (DigitalPresent).
• Ensure critical pages are within three clicks of the homepage. Rebuild hub pages if link depth increased (Niteco).
• Build internal linking logic into templates and content models. Use breadcrumbs to reflect new hierarchy (Naturaily).
• After launch, run Screaming Frog to identify orphaned pages and broken internal links.

Rendering Parity

• Google recommends SSR or static generation over client‑side rendering for SEO (Contentstack).
• Most AI crawlers (ChatGPT, ClaudeBot) do not execute JavaScript – content must be available in the initial HTML response (Naturaily). The exception is Google’s Gemini, which uses Googlebot.
• In staging, test rendering with ?nocache parameters and verify that titles, meta descriptions, and structured data appear in the HTML source.
• If you’re using client‑side rendering, switch to SSR or SSG for indexable pages. Dynamic rendering is still an option but Google officially prefers SSR/SSG (Google JavaScript SEO).

Structured Data and Schema Migration

• Map every schema type from the old platform to the new one. Product schema must include GTIN, MPN, SKU, and price fields (Presta).
• Use JSON‑LD, render it server‑side, and maintain consistent @id references across the entity graph (1Digital).
• Validate all schema with Google’s Rich Results Test after launch.
• AI citation equity is now a first‑class KPI. Schema loss directly impacts visibility in AI Overviews and Perplexity (BrightEdge). Retain at least 95% of AI citations within 60 days.
• Audit the new environment for noindex on pages that should be indexed and for index on pages intended to be excluded.

Faceted Navigation

• Apply noindex, follow to all filter and sort parameter URLs. Use canonical tags pointing to the base category page (SEOlogist).
• In Search Console, set URL parameters to “No URLs” for faceted parameters (Google Search Central).
• If the new platform introduces JavaScript‑based filtering that doesn’t create new URLs, test that filter results are crawlable via direct links.

Analytics and Search Console Continuity

• Keep the same GA4 property and GSC property. If moving domains, use GSC’s Change of Address tool (requires verified ownership on both domains) (Lumar).
• Set up anomaly detection in GA4 using relative metrics (conversion rate) and exclude low‑traffic hours. Use 90/95/99% thresholds for seasonal traffic (Adobe Analytics Community).
• Schedule daily GSC coverage checks for the first two weeks, then weekly through day 90. Monitor Crawl Stats for 404s and redirect efficiency (Presta).
• If possible, access server logs to verify Googlebot’s crawl pattern. Log file analysis provides the quickest feedback loop (GSQi).

Post‑Launch QA and Monitoring

Immediately after DNS cutover:

1. Verify that all 301s work via curl -I or Screaming Frog. Use ?nocache parameters to bypass server cache.
2. Submit new XML sitemap to GSC and Bing.
3. Use the URL Inspection Tool to request indexing for top 20 URLs (Google Search Console Help).
4. Check GSC Crawl Stats – ideally 100% of requests to old domain URLs should return 301s. 4% 404s is a red flag (GSQi).

30‑day focus: Fix top regressions – broken redirects, missing schema, slow pages. Ensure tracking continuity.

60‑day milestone: Evaluate AI citation counts. Make targeted content and technical improvements.

90‑day: Full recovery is expected for well‑executed migrations. If organic traffic is still >20% below baseline, audit for deeper issues (e.g., rendering parity, internal link equity loss).

Rollback Plan

• Before launch, define a rollback trigger – e.g., organic sessions drop >30% for two consecutive days (Niteco).
• For commercial sites, the window to restore is typically four hours or less.
• Keep the old infrastructure live (but redirected) for at least 14 days. Have a complete DNS reversion script ready and tested.

New in 2026: AI Citation Equity

• AI citations are measurable. Bing Webmaster Tools added AI performance reporting in February 2026, showing total citations and grounding queries (Naturaily).
• To preserve AI citations, structured data must be server‑side rendered and FAQ schema must have corresponding on‑page content.
• Blocking AI crawlers via robots.txt will exclude you from those platforms. If you opt in, allow the relevant bots (OAI-SearchBot, GPTBot, etc.) and keep your HTML clean.

Six Non‑Negotiables (1Digital, May 2026)

1. SEO as a release blocker – do not go live without SEO sign‑off.
2. One‑hop 301 redirects – no chains or loops.
3. Schema parity with consistent @id references.
4. Canonical continuity – self‑referencing, same domain.
5. Sitemap regeneration before DNS cutover.
6. 30+ days of daily indexation monitoring via GSC and Screaming Frog.

Other Pitfalls to Avoid

• Letting platform defaults override intended URL structure – e.g., Shopify may add /products/ or /collections/. Explicitly map every pattern.
• Migrating during peak season – schedule during low‑traffic windows and lower DNS TTL to 300 seconds 24–48 hours before launch (DigitalPresent).
• Forgetting the legacy redirect stack – if you already had redirects from an older site, they must be carried forward.
• Using meta‑refresh or JavaScript redirects – only server‑side 301s preserve link equity.

Migration Timeline Summary

Phase	Timing	Key Actions
Pre‑migration	T‑60 to T‑30	Benchmark, URL inventory, redirect map, schema audit
Staging validation	T‑30 to T‑7	Test rendering/redirects/schema, set up HTTP auth
Pre‑launch	T‑7 to T‑1	DNS TTL to 300s, finalize redirects, submit new sitemap
Launch day	T‑0	DNS cutover, verify 301s, submit sitemaps, check Crawl Stats
Post‑migration	T+1 to T+90	Daily GSC (30 days), weekly crawls (30–90 days), AI citation check at day 60
Recovery	T+90+	Content optimization, performance tuning

FAQ

Q: How long does it take for rankings to recover after a replatforming?
A: For a well‑executed migration with 1:1 redirects and content parity, expect normalization within 6–8 weeks. The fastest recoveries in a study of 892 migrations took 19–23 days (DigitalApplied). A temporary 10–20% traffic dip is common (Yotpo).

Q: Should I block AI crawlers during migration?
A: No – you want AI systems to discover your new content as soon as possible. Instead, ensure that clean HTML and structured data are server‑side rendered. Only block specific bots if you have a clear strategic reason.

Q: Do I need to worry about AI citations separately from traditional SEO?
A: Yes. AI search features (Overview, SearchGPT, Perplexity) rely heavily on structured data and initial HTML. Schema loss during migration has been shown to reduce AI visibility directly (BrightEdge). Treat AI citations as a recovery KPI alongside organic traffic.

Q: What’s the biggest mistake agencies make?
A: Letting platform defaults override the intended URL structure. For example, a WooCommerce to Shopify migration that doesn’t map every /product/... path will generate thousands of 301s to homepages – a soft 404 trap. Always map manually or with validated tools.

Q: Should I migrate during a seasonal low?
A: Absolutely. Launch on a Tuesday‑Thursday in a low‑traffic month. Lower DNS TTL to 300 seconds 24–48 hours before to speed propagation (DigitalPresent).

For deeper dives on specific migration steps, visit the SEO1 Library Technical Guide and Migration Case Studies.