Technical SEO and migration · Backed by Google for Startups
Website migration without losing rankings, from the redirect map to the weeks after go-live.
Changing domains, platforms or URL structures is the riskiest moment in the life of a website. Netlinks runs the migration with a complete inventory, redirects mapped one by one, validation in a closed staging environment and daily monitoring after launch.
We work in English, Portuguese and Spanish. Headquarters in Belo Horizonte, Brazil.
01Overview
Website migration
Website migration without losing rankings is an engineering project with a method, not a checklist someone opens on launch night. Every live URL carries history, external links and hard-won position. If that URL disappears, moves without notice or answers with the wrong status code, the search engine treats it as brand new and the history goes with it.
Netlinks has run domain, platform, URL structure and internationalization migrations since 2017, across more than 200 companies. The work starts weeks before go-live, with an inventory of everything currently live, and continues for at least 30 days afterward, with traffic, crawling and errors visible in the Netlinks App.
The most recent one is CLM Controller, an international accounting firm: a domain reorganization with 2,013 mapped redirects, three languages on the .com, the .com.br dedicated to Brazil and complete hreflang, live on August 27, 2026, with no drop on the core terms.
- 2,013
- redirects mapped in the CLM Controller migration
- 200+
- companies served since 2017
- 30 days
- of daily monitoring after go-live
- 24h
- of live tracking in the App and on WhatsApp
02Definition
What a website migration is, and why it costs rankings
A migration is any change that alters the address, the technology or the structure of pages that are already indexed. Google does not see your project. It sees URLs. When a URL changes and nothing explains where it went, the crawler has to rediscover everything, and the authority that address accumulated is not transferred on its own.
Post-launch drops almost always come from the same short list:
- Missing or chained redirects. The old page returns a 404, or takes three hops to reach its destination. Every hop wastes crawl budget and dilutes the signal.
- Blanket redirects to the homepage. The single most common mistake in the market. Sending 800 product URLs to the homepage tells the search engine that content no longer exists.
- An indexed staging environment. Staging goes up unprotected, gets crawled, and the new site launches competing against a duplicate of itself.
- Canonicals and hreflang pointing at the old domain. The new site spends months telling search engines that the good version is the one already switched off.
- Lost structured data and metadata. The new platform does not carry over the schema, titles and descriptions that existed, and pages start from zero.
- A speed regression at launch. The new site ships without caching, without image compression and with too much script. Users feel it and Core Web Vitals records it.
None of these are unpredictable. All of them show up in the inventory and in staging before launch, provided somebody actually looks.
03Scope
The six migration types we run
The protocol is identical in every case: inventory, map, test, switch, watch. What changes is where the risk lives.
Domain change
Rebranding, a legal entity change or a merger. The job is transferring authority, filing the change of address and keeping redirects alive for years, not for a quarter.
Platform change
WordPress to Next.js, Magento to Shopify, custom stack to headless. Every platform builds URLs, pagination and filters its own way, and that is where catalogs get lost.
URL structure change
New categories, new silos, the end of query parameters. The architecture improves, as long as every old address knows where to go.
Redesign with content changes
A new site with merged or retired pages. Before deleting anything, we measure how much traffic and how many links that page carries.
Domain merge or split
Two brands becoming one, or one operation splitting by country. It involves canonicals, hreflang and a clear decision about which domain owns which topic.
Internationalization
One domain per country, or one domain with language folders. That was the CLM Controller design: .com in English, Spanish and Chinese, .com.br dedicated to Brazil. Detailed in international SEO.
04Cadence
How we run the migration, phase by phase
A migration done properly starts four to six weeks before the switch date. The schedule below is what we run on mid-sized projects; operations with tens of thousands of URLs get more time in inventory and testing.
- Phase 1
Inventory what exists
Full crawl of the current site, 16 months of Search Console export, the list of URLs with traffic, the list of URLs with backlinks and everything in the sitemap. Nothing enters the map from memory.
- Phase 2
Redirect map
Every old URL gets an explicit destination, one by one, prioritized by traffic and by links earned. Pages without an equivalent get the closest match by topic, never the homepage.
- Phase 3
New site in staging
Staging behind authentication, off the main domain, with no chance of being indexed. We validate titles, descriptions, structured data, canonicals, hreflang, pagination and speed before any switch.
- Phase 4
Dry run
The full redirect set is tested against the real URL list, checking status codes, chains and loops. The report goes to your engineering team with whatever is still missing.
- Phase 5
Supervised go-live
The switch happens in a low-traffic window with the Netlinks team online. New sitemap submitted, change of address filed, robots opened, caches warmed and a sample of URLs tested live.
- Phase 6
Thirty days of watch
Daily crawling, index coverage, real user 404s, priority keyword positions and Core Web Vitals. Whatever surfaces gets fixed that week, not in next month's report.
05Case study
The redirect map: 2,013 lines at CLM Controller
CLM Controller is an accounting firm serving foreign companies operating in Brazil. The site lived on a single domain with Portuguese, English and Spanish content mixed together, and the business needed to separate its international audience from its Brazilian one without losing either track record.
The approved design put English, Spanish and Chinese on the .com and dedicated the .com.br to Brazil, with complete hreflang between versions. It took 2,013 mapped redirects and 20 ported pages, including the calculators that generated inbound contacts, and it went live on August 27, 2026.
What made the map work was the order of operations: first the list of URLs with traffic and backlinks, then the destination decision topic by topic, and only then implementation. A redirect written on switch day is a redirect guessed.
- One line per URL, with source, destination, reason and owner
- Priority set by organic traffic and by external links earned
- Zero redirect chains: every old address points straight to the final destination
- Automated testing of the entire map before and after the switch
06Risk
What usually breaks, and how we block each case
This is the table we open in the first migration meeting. It exists so your engineering team knows exactly what will be checked in every phase.
| Risk | How it shows up | How we block it |
|---|---|---|
| URL with no destination | A 404 on a page that had traffic and links | Cross-referenced inventory of crawl, Search Console and backlinks before the map |
| Redirect chains | Three hops to the destination, slow recrawling | A single final destination rule, verified by automated testing |
| Indexed staging | Duplicate content competing with the new site | Staging off the main domain and behind authentication |
| Lost structured data | Rich results disappear after the switch | Markup ported and validated in staging, with breakage alerts after launch |
| Generic metadata | Default platform titles and descriptions | Old metadata exported and rewritten page by page before launch |
| Speed regression | Core Web Vitals drop and conversion drops with it | PageSpeed measured in staging and fixed before the date, not after |
| Wrong sitemap and robots | The new site is blocked or only partly discovered | A go-live checklist covering robots, sitemap and change of address |
07Go-live
The switch-day checklist
Switch day has a scheduled hour, an owner per item and an execution order. Nothing is checked from memory. This is the checklist the Netlinks team runs alongside your engineering team, with every line ticked live:
Once the last line is ticked, the watch phase begins. The first report goes out the next day, listing every URL that answered something other than expected.
- Full backup of the old site, with the URL list frozen
- Redirect rules live and tested against the real list, one sample per page type
- robots.txt opened and no noindex inherited from staging
- New sitemap submitted, old sitemap kept for a few weeks to speed up discovery
- Change of address filed in Search Console whenever the domain changes
- Canonicals, hreflang and pagination verified on a sample of every page type
- Structured data validated on the live site
- Analytics, conversion tags and form integrations tested with a real submission
- PageSpeed measured before and after, on mobile and desktop
08Post-launch
The 30 days after the migration
A migration does not end at launch. It ends when the traffic curve returns to its previous level and the index reflects the new structure. Through that period, what you follow in the Netlinks App updates without waiting for a month-end close:
- Index coverage: how many new URLs are in, how many old ones are out, and which are stuck in between.
- Real user errors: 404s that happened in actual sessions, not only in crawls, with the referring source.
- Priority keyword positions: compared against the baseline captured the week before the switch.
- Core Web Vitals and PageSpeed: by template, because regressions are rarely site-wide, they belong to one page type.
- Backlinks still pointing at old URLs: the strongest ones become outreach targets to update the link at the source, handled by the authority practice.
Two weeks of volatility is normal while the engine reprocesses the site. Sustained decline after the first month is not. When that happens there is almost always a localized technical cause, and it surfaces in continuous diagnostics instead of turning into a mystery.
The wider technical work behind that tracking lives in technical SEO and SEO audit.
09Infrastructure
When the server moves too
Many of the migrations that reach Netlinks combine two changes on the same day: the site changes shape and also changes home. That is the case for companies leaving slow shared hosting for Google Cloud, or those that need separate staging and production environments.
Our engineering practice owns that part: provisioning, certificates, DNS with planned propagation, caching, image compression and uptime monitoring. SP Labor, a laboratory equipment e-commerce, went through exactly this path: a move to Google Cloud plus a technical audit that closed with a PageSpeed of 97 on mobile and 99 on desktop for the homepage.
When the server move is bundled in, sequence matters: bring up the new infrastructure with the old site running, then switch the content. Two large changes in the same minute make it impossible to tell which one caused a problem.
10The deliverable
What you receive
Finished work, approved by you in the App and shipped. Not a report that only describes the problem.
URL inventory
A sheet with every live URL, 16 months of traffic, backlinks earned and the decision made for each one.
301 redirect map
Source, destination, reason and owner, line by line, ready for your engineering team or for us to implement together.
Staging validation report
Titles, descriptions, structured data, canonicals, hreflang and speed, with everything that has to be fixed before the date.
Signed go-live checklist
Item by item, with an owner and a time. Executed live with your team on switch day.
Ranking baseline
A snapshot of priority keywords the week before launch, so the comparison afterward is honest.
Daily report in the App
Coverage, errors, positions and speed through the first four weeks, with what is fixed and what is queued.
11How it works
From the audit to the page going live
The same path on every project, with an owner and a date on each step.
- 01
Diagnosis call
We map why you are migrating, the target date, who is building the new site and what has already been decided. No commitment.
- 02
Inventory and risk
Crawl of the current site plus traffic and link analysis. It produces the picture of what is at stake and the real size of the project.
- 03
Map and specification
The redirect map plus a technical spec of what the new site must carry: metadata, schema, canonicals, hreflang and sitemap.
- 04
Staging testing
We validate the new site in a closed environment, round after round, until the report comes back clean.
- 05
Supervised switch
Go-live in the agreed window, with the Netlinks team online and the checklist ticked in real time.
- 06
Watch and adjust
Thirty days of daily tracking in the App, weekly fixes and an executive read at the end of the period.
12Who leads
Who leads your project
The three of them own the process and manage the team on every account. No senior who shows up at kickoff and disappears: they sign off on the plan, review the work and talk to your team.
Digo Garcia
Founder and CAIO
Over 20 years in digital. Built and scaled Letras.mus.br, then worked at Méliuz and in the joint venture with Globo.com. Owns AI, technical SEO and engineering inside every project.
Letras.mus.br · Méliuz · Globo.com
André Mousinho
Chief SEO Officer
Comes from Rock Content, Tray/Locaweb and LWSA. Designs the keyword mapping, the content architecture and the GEO strategy that turns a brand into the answer.
Rock Content · Tray/Locaweb · LWSA
Gabriel Raposo
Chief Authority Officer
Comes from Ambev and Agibank. Owns link building, digital PR and the authority that holds up both rankings and citations inside AI answers.
Ambev · Agibank
And the team behind them: Nathália Barcala (cofounder and COO, governance), Roberto Murta (CTO), Marcelo Ligere (senior advisor, 30+ years, formerly Grupo Globo) and Alfred, the artificial intelligence that runs the day to day: delivery queue, reporting and 24/7 support on WhatsApp.
13Who it is for
Where this work pays off the most
Companies replatforming
Moving off WordPress, Magento, Shopify or a custom stack. The risk lives in the catalog and in the filters.
Companies changing brand or domain
Rebranding, merger or a split by market. The job is transferring authority without breaking what already ranks.
Companies that already migrated and lost traffic
The switch happened, the curve dropped and nobody knows why. We run the post-migration diagnosis and the recovery, starting with what costs the most revenue.
14Case studies
Results with a name, a number and a date
Real projects, with what was done and what changed. No generic charts and no promises of a position.
FAQ
Frequently asked questions
How do you migrate a website without losing rankings?
With a complete inventory first, a 301 redirect map built one URL at a time, validation in a closed staging environment, a checklist executed on switch day and daily monitoring for at least 30 days. What costs rankings is rarely the migration itself; it is the URL left without a destination and the staging environment that got indexed by accident.
How far ahead of launch do you need to start?
Four to six weeks before the date is ideal, with the new site already available in staging. Projects with tens of thousands of URLs need longer for inventory. We can start with less lead time, but the risk rises and the map stops being checked line by line.
Is a traffic drop after migration normal?
Two to three weeks of volatility is expected while the engine reprocesses the new structure. Sustained decline after the first month points to a technical cause: a missing redirect, a wrong canonical, a noindex inherited from staging or lost content. In that case the diagnosis happens the same week.
Can I redirect old pages to the homepage?
Not safely. Bulk redirects to the homepage signal that the content no longer exists, and the history of those URLs is lost. Pages without a direct equivalent get the closest match by topic, and only those with no counterpart at all get proper end-of-life treatment.
Do you work alongside my agency or my in-house team?
We do, and it is the most common setup. Netlinks delivers the inventory, the map and the specification, follows development and owns the go-live checklist. When the new site is ours too, web development and SEO run on the same schedule.
What if the migration already happened and traffic fell?
It starts with a post-migration diagnosis: old and new URL lists compared, a map of everything now returning 404, and checks on canonicals, hreflang, robots and structured data. Fixes are ordered by revenue lost. Recovering is slower than preventing, but it works in most cases.
Does the migration include moving hosting and servers?
It does when the project calls for it. Our engineering practice handles provisioning, DNS, certificates, caching and monitoring, as we did for SP Labor when moving that e-commerce operation to Google Cloud. The rule is to separate the two changes in the schedule so any effect has a clear source.
Is an international migration different?
It is more delicate, because it means deciding which domain serves which audience and implementing hreflang across every version. That was the CLM Controller design: 2,013 redirects, .com in English, Spanish and Chinese, .com.br dedicated to Brazil. The full practice is in international SEO.
How long do redirects need to stay live?
At least a year, and in practice we recommend keeping them permanently for URLs that receive external links. Keeping a redirect rule costs almost nothing; switching it off early costs the authority that took years to build.
Keep reading
Related services
Next step
Get a diagnosis of your case
Tell us what your company sells and to whom. One of the three looks at your scenario, points out what is holding organic back today and shows where we would start.
- A review of your site and your market, with no commitment
- An answer from a senior specialist within 24 business hours
- A direct conversation: what can be done, in what order and how deep
Prefer WhatsApp?
Chat directly with a Netlinks Expert on WhatsApp.
Rather talk now? WhatsApp is answered by the team, with Alfred covering the rest of the day.
Talk to an Expert
Fill out the form and we will get back to you within 24 business hours.