Crawlable URLs
Crawl the website, sitemaps, menus, internal links, orphan pages, parameters, PDFs, images and useful public resources.
Guides
Change design, CMS, URLs, domain or hosting without throwing away pages, links and signals built over time. Method, checks and responsibilities before, during and after launch.
In brief
No one can guarantee unchanged rankings after a redesign. Avoidable risk can, however, be reduced: measure the current website, inventory every useful URL, define an equivalent destination, implement permanent redirects, keep content and technical signals consistent, test the new website and monitor Google and Bing after launch.
Yes—if “without losing” means protecting what exists and reducing controllable mistakes, rather than promising that Google will freeze every ranking.
Before estimating risk, timing and checks, establish what actually changes for users and crawlers.
| Project | What changes | Main risk | Decisive check |
|---|---|---|---|
| Redesign | Visuals, components and journeys; URLs may remain unchanged | Reduced content, removed links, worse HTML or performance | Page-by-page comparison of old and new output |
| CMS or technology change | Publishing system, templates, markup and often URLs | Slugs, metadata, schema and status codes rewritten without a map | Complete inventory and final-response testing |
| New URL architecture | Paths, categories, languages or parameters change | 404s, chains and non-equivalent destinations | One-to-one map and server-side permanent redirects |
| Domain change | The host changes for all or many pages | Signals split across properties, variants and subdomains | Redirects, verified properties and Change of Address where applicable |
| Hosting or CDN change | Infrastructure and DNS, without changing visible URLs | Downtime, DNS, TLS, headers, WAF or insufficient capacity | New-origin tests, TTL, monitoring and the old server kept available |
Prudent rule
Google recommends changing one thing at a time where possible. Combining domain, CMS, architecture, design and content changes at the same moment makes transferring signals and identifying the cause of a problem more difficult.
Without a starting point, you cannot distinguish a migration loss from seasonality or a page that was already producing nothing.
Crawl the website, sitemaps, menus, internal links, orphan pages, parameters, PDFs, images and useful public resources.
Search Console and Bing show indexed and excluded URLs, selected canonicals, errors, redirects and blocks that must be understood before they are copied.
Export at least pages, queries, clicks, impressions, countries and devices over a period long enough to identify seasonality.
Identify URLs receiving external links, referrals, mentions, downloads or direct traffic: they are not always the most visited pages.
Save main copy, titles, descriptions, headings, images, alt text, structured data, canonicals, hreflang and editorial dates.
Enquiries, sales, calls, sign-ups and lead quality prevent a useful page from being sacrificed merely because its design looks old.
Every old resource needs an explicit decision, a destination and a verifiable technical outcome.
One row for every URL
The map connects source URL, current status, traffic, backlinks, content, proposed destination, redirect type, expected canonical and responsible person.
It is not generated from the sitemap alone: historic URLs can exist in Search Console, backlinks, logs, old feeds, images, PDFs, campaigns and users’ saved links.
The destination must answer the same need. Redirecting a specific detail page to a generic homepage does not transfer useful context.
The old URL should reach the definitive page directly, avoiding chains, loops, HTTP hops and intermediate www variants.
Test pages with traffic, links, sales or reputational importance first; then cover the full inventory.
The map must accompany development, testing and monitoring: it cannot remain a forgotten spreadsheet before go-live.
Each tool communicates a different situation. Using one for convenience rather than meaning creates contradictory signals.
| Situation | Tool | Message | Common mistake |
|---|---|---|---|
| The resource has moved permanently | Server-side 301 or 308 | The destination should replace the source as the primary URL | 302, JavaScript or chains when a direct redirect is available |
| The move will genuinely be reversed | 302 or 307 | The source may remain the URL shown in results | Using it for months on a permanent migration |
| Two accessible URLs show identical or very similar content | Consistent rel=canonical | Indicates the preferred version without moving the user | Using it instead of a redirect when the old page is being retired |
| The resource no longer exists and has no equivalent | A genuine 404 or 410 | Communicates that the URL no longer provides the requested content | Returning 200 with an empty page or sending everything to the homepage |
A new page can return 200 and have a perfect redirect, yet lose value if the content and relationships that made it useful have disappeared.
Keep the answers, evidence, data, products and sections that satisfy the search need; improve weak material without hollowing out the page.
The title, H1, description and main content should describe the same page. Rewriting all of them at once makes the migration effect harder to interpret.
Rebuild only markup consistent with what remains visible: organisation, author, products, breadcrumbs, FAQs and relationships must not describe an outdated reality.
Update menus, breadcrumbs, contextual links, footers and feeds directly to the new URLs instead of relying on redirects for internal navigation.
Signal consistency
Google treats redirects and canonicals as strong signals and a sitemap as a weaker one. If they point to different destinations, the search engine has to resolve a contradiction created by the website.
The new URL should be self-canonical, return 200, appear in the sitemap and receive internal links; the old URL should redirect directly to that destination.
The new website should be tested as if it were live, without becoming an indexable public copy or carrying development blocks into launch.
Server-side authentication or network restrictions are preferable to a public robots.txt file for preventing unwanted access.
If you use noindex or disallow during development, keep an explicit checklist for removing them from the public website.
Verify 200s, redirects, 404s, 410s and 5xx responses with HTTP requests, not just by looking at what the browser displays.
Compare old and new inventories, titles, headings, canonicals, hreflang, schema, images, PDFs and internal links.
Every source URL must reach the intended destination in one hop, with no rule creating loops or collisions.
Enquiries, checkout, calls, emails and confirmations must work before traffic is moved: SEO is pointless if the commercial journey breaks.
Layout, images, fonts, JavaScript, focus, keyboard use, errors and assistive technologies must be tested on representative templates.
Schedule new ideas, late copy changes and non-essential functions for later: continuous change prevents reliable testing.
Every language has its own URLs, canonicals, alternatives and content. Translating a template does not automatically rebuild those relationships.
Every old language variant points to the equivalent new page in the same language, not the Italian homepage or a default language.
Translated versions remain self-canonical; canonicalising every language to one page can cause them to be treated as duplicates.
Each page lists itself and every definitive alternative, with the same set of relationships on the other versions.
Sitemaps, language selectors and internal links must use reachable final URLs, without redirects or legacy variants.
For URLs, hreflang, canonicals and localisation, also read the complete multilingual SEO and GEO guide.
The correct outcome depends on whether equivalent content exists and the value that URL still holds for people and search engines.
Keep
Preserve the address and improve the page in measured steps. This is the lowest-risk route when structure and intent remain valid.
Move
Use a direct permanent redirect to the equivalent destination and update every internal link to the new URL.
Merge
The new page must genuinely cover the useful intents of its sources. Automatic redirects to a vague page are not enough.
Remove
Return a real 404 or 410, remove the URL from sitemaps and internal links and provide a useful error page without masking the status.
Do not redirect everything to the homepage
A homepage is not a universal equivalent for deleted products, articles, services and documents. This confuses users, can produce soft 404s and hides which content has genuinely disappeared.
Order, ownership and the ability to restore service reduce the interval in which users and crawlers receive contradictory responses.
Content, the URL map and configuration remain stable throughout the launch window.
Retain the old website, database, configuration, DNS, certificates, exports and a rollback plan proportionate to the project.
Avoid windows in which new URLs exist while old ones return errors or still point to intermediate destinations.
Check noindex, robots.txt, authentication, WAF and headers on the live response after caches, CDN and routing.
Submit the sitemap containing final URLs, preserve verifications and use Change of Address only for supported domain moves.
Update Bing Webmaster Tools and submit added, changed, moved or removed URLs; receipt does not guarantee indexing.
Immediately compare status codes, chains, canonicals, links, hreflang, schema and resources against the approved map.
Submit test forms, complete representative checkouts and check emails, events and confirmations in the real environment.
Systems that search and cite the Web depend on reachable URLs, consistent content and updated indexes. There is, however, no universal button that transfers every citation.
Redirects, canonicals, sitemaps and links help Google recognise the new page as it crawls old and new URLs.
Bing Webmaster Tools helps observe redirects, indexed URLs, errors and signals that also support grounded experiences.
It can quickly notify added, updated, moved or deleted URLs; it does not replace sitemaps, crawling or page quality.
Robots.txt, WAF and CDN should allow only the intended access to new URLs, without inheriting accidental staging blocks.
An AI answer may retain an old URL or outdated passage for a while. A redirect protects the visit but cannot guarantee when the source will be refreshed.
Name, author, organisation, services, products, contact details and structured data should describe the same reality before and after migration.
For visibility and citation, read how to appear in ChatGPT and AI answers ; for access and bots, read llms.txt, robots.txt and AI crawlers.
Comparisons by page, query, error type and time interval distinguish normal settling from problems that require correction.
| When | Check | Signal | Action |
|---|---|---|---|
| First 24 hours | Uptime, DNS, TLS, status codes, redirects, robots, noindex, forms | Immediate technical errors and broken journeys | Fix immediately or apply the planned rollback |
| First week | Logs, crawls, 404s, chains, sitemap, URL Inspection | Discovery and response of priority URLs | Repair the map and inconsistent signals |
| First 4–6 weeks | Indexing, clicks, impressions, queries, pages and conversions | Progressive transfer and persistent changes | Analyse groups of URLs rather than reacting to the total alone |
| Following quarter | Trends, updated backlinks, redirects still active, lead quality | Stability of the new asset and editorial opportunities | Optimise content without confusing migration data |
These problems are often invisible in a visual preview but obvious to crawlers, users, search engines and measurement systems.
“We no longer need the old URLs”
Changing every slug without an inventory breaks links, history, bookmarks and signals built over time.
“We will redirect everything to the homepage”
The destination is not equivalent, the user loses context and Google may interpret the response as a soft 404.
“The plugin will handle the SEO migration”
A tool can apply rules; it cannot decide which content is equivalent, which pages convert or which signals must be preserved.
“We will change everything together”
Simultaneous domain, CMS, URL, design and copy changes increase risk and make the cause of a drop difficult to isolate.
“The new website looks good, so it is better”
Successful visuals do not compensate for reduced content, poor HTML, worse performance, broken forms or neglected accessibility.
“The canonical replaces the redirect”
A canonical does not move the user and remains a hint. If the old resource is retired, an appropriate permanent redirect is required.
“After launch, we just wait”
Fluctuations may be normal, but errors, blocks and chains must be found immediately: waiting does not fix incorrect configuration.
“We can switch off the old domain immediately”
Redirects, verifications and historic backlinks need continuity. Keep the domain and minimum infrastructure for as long as necessary.
Ten checks that must produce evidence, not simple verbal assurances.
Crawls, sitemaps, Search Console, Bing, logs, backlinks, PDFs, images and campaigns have all contributed to the URL list.
Every source has a decision, a relevant destination, a status code and an owner.
Priority URLs, followed by the full list, reach the destination directly without loops or chains.
Canonicals, sitemaps, internal links, hreflang and redirects identify the same final URLs.
Intent, information, evidence, products, images, metadata and schema are present and current.
The launch plan removes temporary authentication, noindex and disallow rules without exposing private areas.
Responsive layout, browsers, keyboard, assistive technologies, performance, errors and resources have been tested on representative cases.
Forms, emails, calls, checkout, payments, consent and confirmations have passed end-to-end tests.
Search Console, Bing Webmaster Tools, proportionate analytics, logs and notifications have access and baselines available.
It is clear who decides, who intervenes, what evidence is retained and how service is restored after a serious error.
I do not repair the old theme or transfer plugins and technical debt: I analyse the asset, design the new system and bring across only what has value.
GLP method
The migration audit separates diagnosis from the sale: it records the current website, identifies risks and priorities and produces the plan used to estimate the rebuild.
| Service | Scope | Price before tax |
|---|---|---|
| Migration audit | Analysis, risks, preliminary map, action plan and fixed quotation | €175 |
| Essential | Essential bespoke rebuild according to the approved scope | €1,000 |
| Pro | More complex project, content, components and extended migration | from €2,500 |
| Essential e-commerce | Catalogue, checkout and commercial URLs within the migration plan | from €4,500 |
The final price depends on the assessed scope. See the complete route away from WordPress. No intervention guarantees unchanged rankings, traffic or citations.
The technical recommendations derive from Google, Search Console, Bing and IndexNow documentation; the operational and commercial method is an explicitly stated professional synthesis.
Preparation, URL mapping, redirects, fluctuations, Search Console and migration monitoring.
DNS, TTL, infrastructure, removal of temporary blocks and monitoring the new server.
Differences between permanent and temporary redirects and the preference for server-side redirects.
Redirects, rel=canonical and sitemaps as signals of different strength for consolidating duplicate URLs.
When to use it for a domain change, plus requirements, limitations and preliminary checks.
Search Console, time comparisons, seasonality, updates and migration problems.
Correct use of 404 and 410 when a resource no longer exists and has no replacement.
Language URLs, reciprocal hreflang, absolute URLs and x-default on multilingual websites.
Notification of added, updated, moved or deleted URLs and use after migrations and redesigns.
Permanent redirects, removals, crawling, rendering and continuity of citations and grounding.
Sources checked on 7 August 2026. Interfaces, report names and search-engine support can change: always recheck official documentation before a migration.
Direct answers about rankings, redirects, domains, WordPress, timing and responsibilities.
No. Rankings change continuously and a migration requires new crawling and assessment. Professional work preserves useful signals, eliminates avoidable errors and monitors changes; it cannot freeze the algorithm or competitors.
Google states that 301s and other permanent redirects do not cause PageRank loss. The destination must still be relevant, and the migration must avoid chains, errors and contradictory signals.
When a URL is clear, stable and still consistent, keeping it is often sensible. If the architecture must change, every useful old URL needs an equivalent destination and a permanent redirect.
No. The homepage is rarely an equivalent substitute. Merged pages should point to the relevant resource; pages with no replacement should return a genuine 404 or 410.
There is no fixed timeframe. Google says a medium-sized website can take a few weeks for most pages to be moved in its index; size, server speed, links, sitemaps and crawl frequency influence the process.
After moving and redirecting a website from one domain or subdomain to another, with both properties verified. It is not used for simple internal path changes or an HTTP-to-HTTPS move alone.
It is technically possible but increases risk and ambiguity. Google recommends separating changes where possible. If they must coincide, a stronger baseline, complete tests and longer monitoring are required.
Protect it with authentication or server-side restrictions. Noindex and robots.txt can be additional layers, but they must be removed reliably from the public site, and robots.txt does not protect confidential content.
Not with my method. I analyse and recover the necessary content, media, URLs and functions, then build a bespoke website without transferring the theme, plugins and technical debt that prompted the migration.
No. IndexNow quickly notifies changed URLs; the sitemap maintains the broader canonical inventory. They can work together after a migration, but neither guarantees indexing.
Keep them for a long time and often without a useful expiry, especially for domain changes, backlinks, bookmarks and historic campaigns. Removing them early recreates errors for users and crawlers.
The process starts with a migration audit from €175. Rebuilding follows the approved scope: Essential €1,000, Pro from €2,500 or Essential e-commerce from €4,500, before tax. Complexity, content and integrations determine the final quotation.
Next step
We start with the audit: I record the current website’s URLs, content, signals and risks and prepare a verifiable plan before estimating the rebuild.