Website Migration SEO Checklist: How to Redesign Without Losing Rankings

A website migration done well protects your rankings, preserves your traffic and improves performance at the same time. Done badly, it hands years of compounding to your competitors in a fortnight and the damage is usually invisible until it is expensive.

This guide covers the full process from pre-launch planning through post-launch monitoring, including the parts most checklists skip: what a normal recovery dip looks like, the rollback criteria worth writing down and the platform specific traps. For the bigger picture on building a site worth migrating to, our guide on service page structure and copy that converts covers structure and conversion and the complete SEO guide covers how this fits into the broader strategy.

What counts as a migration

It is broader than most people think. If URLs, content, templates, internal links or platform structure change, there is SEO risk.

If the platform itself is changing, our guide on WordPress vs Webflow vs Shopify vs custom builds covers the trade offs before you commit, because some platform constraints (covered below) are migration decisions in disguise.

Why migrations lose rankings: the seven causes

URLs change without redirects: old URLs 404 and every backlink, listing and bookmark pointing at them leads to a dead end.

Ranking pages get deleted: 'old' pages that did not fit the new design take their rankings to the grave.

Content gets rewritten too aggressively: the new page looks better but no longer answers the query that ranked. Google sees a different page and treats it as one.

Metadata and headings get stripped: titles, descriptions and H1s vanish in the rebuild, taking the signals with them.

Internal links break: menus, breadcrumbs and body links all change and important pages quietly lose their internal authority.

Search engines get blocked: robots.txt rules, noindex tags and staging leftovers prevent indexation invisibly.

Tracking breaks: traffic and leads drop and nobody can measure what happened because analytics died in the switch.

Before you migrate: the SEO safety net

Everything in this section happens before the new site goes live. This is where migrations are won or lost.

1. Crawl and document the existing site

You cannot protect what you have not documented. Export every live URL with its title, meta description, H1, status code, canonical, indexability and internal links (Screaming Frog, Sitebulb or the Ahrefs and Semrush crawlers all do this in minutes). This crawl is the migration's source of truth.

The inventory most migrations miss: PDFs and documents that rank or get linked (price lists, guides, forms all live at URLs too), images earning image search traffic (their URLs change with the rebuild), any subdomains and their content, URL parameters doing real work and the embedded assets other sites hotlink. Each is a URL class needing a redirect decision and each is invisible in a page level crawl until a quarter of the 404 report turns out to be PDFs.

2. Build the benchmark pack

The performance record you will judge the migration against, captured in an afternoon:

• Search Console exports (16 months of queries, pages, impressions, clicks)

• GA4 organic traffic and conversion figures by landing page

• Current rankings for your money terms

• The top pages by traffic and by conversions (they are not the same list and both need protecting)

• Your backlink profile's most linked pages

Screenshot the trends, file the exports, date everything. Post launch, this pack is the difference between 'traffic feels down' and 'the services section is down 40% while everything else held', which is the difference between panic and a fix. Our Google Search Console guide covers reading these reports if they are unfamiliar.

3. Build the redirect map

A spreadsheet, three working columns: every old URL from the crawl, its new destination URL and the decision status (mapped, merged, deliberately retired). The rules that make it work:

• Redirect to the most relevant equivalent page, never bulk redirect everything to the homepage

• Merged pages redirect to the page that absorbed their topic

• Deliberately retired pages with no equivalent can 404 or 410 honestly, as a decision rather than an accident

• Every redirect is a single hop 301, not a chain

For a site of any size, sort the sheet by traffic and links first: the top 100 URLs carry most of the equity and deserve individual attention, the long tail can be handled by pattern rules, carefully.

Redirect implementation: choosing the method

Whichever method: the redirect map spreadsheet remains the source of truth, the implementation is just its deployment. A copy lives somewhere that survives developer handovers, because the person maintaining these rules in three years has not been hired yet.

4. Preserve what ranked

For every page that carries traffic: keep the title tag and H1 or change them deliberately with eyes open, keep the substance that answers the ranking query even as the design and copy improve, keep the heading structure's topic coverage and carry across the schema markup worth keeping. Our guide on how to optimise service pages for SEO covers the structural elements worth preserving.

Redesigns should upgrade presentation and conversion. The search relevant substance moves house intact. 'Make it look new without making it read new' is the discipline for the money pages, at least for the first months. Improve the content later as its own project, one variable at a time.

This also applies to E-E-A-T signals: author bios, credentials, review embeds, case studies and trust signals that took years to build need to survive the migration intact. A redesign that strips authorship and proof elements has undermined its own ranking infrastructure.

5. Plan the internal linking

New menus, breadcrumbs and footers should reinforce the same priority pages the old ones did and body links should point directly at final URLs (never at old URLs that redirect: internal redirect chains waste crawl budget and dilute signals). Our guide on how to plan your site architecture for small businesses covers doing it deliberately and a migration is genuinely the right moment to fix structural sins, provided every moved page lands in the redirect map.

Platform specific gotchas worth knowing before you commit

Shopify: forces its URL structure (/pages/, /products/, /collections/), so a migration to Shopify changes almost every URL by definition and the redirect map is the whole game. Its native redirect tool handles the volume, the trap is assuming any old URL can be preserved.

WordPress: permalink settings decide the URL structure sitewide. Set them before content goes in, because changing them later is a second migration. Watch plugin generated noindex settings and staging plugins left active.

Wix, Squarespace and site builders: varying degrees of URL and redirect control. Check that the platform can actually implement your redirect map before choosing it, because a platform that cannot do proper 301s at scale has made the migration decision for you.

Custom and headless builds: total control, total responsibility. The redirects, canonicals and sitemap logic all have to be built, which means they all have to be specified, which means the redirect map goes in the developer brief, not in a launch week conversation.

Pre-launch QA: the staging leftovers sweep

The checks that prevent the classic disasters, run on staging and again at launch:

• robots.txt is not blocking crawlers (staging blocks everything by design, that setting carrying over is the most common migration disaster in existence)

• No leftover noindex tags (the site launches, looks perfect and Google systematically removes every page)

• Canonicals point at final live URLs, not staging or old domain addresses

• The XML sitemap contains only final indexable URLs

• Internal links point directly at final URLs

• Schema carried across where it earns its place

• GA4, Tag Manager and conversion tracking all tested and confirmed before launch, not the week after

The full launch day checklist

Print it, work it top to bottom, tick it honestly. Fifteen items, most of them minutes each, standing between the rebuild investment and the silent killer. Check speed regressions against the thresholds in our Core Web Vitals explainer and run your meta titles and descriptions through a quick check to catch anything stripped during the rebuild.

Launch day: the running order

After launch: the monitoring windows and the normal dip

What normal looks like, so nobody panics into a rollback: some ranking wobble and traffic softness in the first one to four weeks is expected while Google recrawls and re-evaluates. Impressions fluctuate, individual keywords bounce and the graph looks nervous. A well executed migration typically stabilises within four to eight weeks, often returning stronger as the improved site's signals land.

What is not normal: a sustained drop past 20 to 30% that deepens after week two, entire sections vanishing from the index or the 404 report growing daily. Normal is a wobble that flattens. A problem is a slide that steepens.

After the verdict read, run the full 25 point SEO audit checklist against the new site to catch anything the migration monitoring did not cover. Then the improvement projects (content upgrades, speed optimisation) can begin safely, one variable at a time.

The content freeze window

A discipline that saves diagnostic sanity: freeze content changes from two weeks before launch until the recovery verdict at week eight. New pages, rewrites and deletions during the migration window add variables to a system you are trying to read. Queue the content work, it launches into a stable, measured environment two months later and performs better for it.

Redirect map hygiene after launch

The map is not finished at launch, it is versioned. Month one adds the 404 report's discoveries. Month three removes any temporary rules and flattens chains found in re-crawls. Annually, someone confirms the rules still deploy (hosting migrations and CDN changes silently drop redirect layers more often than anyone admits). File the current version wherever the business keeps things that outlive staff, because the redirect map is infrastructure with a decade of working life.

Rollback criteria: write them down before launch

The conversation nobody wants to have at 9pm on launch day goes better decided in advance. Reasonable triggers for rolling back:

• The redirect layer fundamentally failing at scale (not individual misses, which get fixed forward)

• Site wide indexation blocking that cannot be fixed within hours

• Core functionality (forms, checkout, booking) broken with no immediate fix

Reasonable triggers for fixing forward instead: individual redirect misses, cosmetic issues, a normal looking ranking wobble and almost everything else, because rollbacks carry their own SEO cost (another round of URL confusion) and the fix forward is usually faster.

The special cases

Domain changes: add the Change of Address tool in Search Console on top of everything above and expect the longest re-evaluation timeline of any migration type: brand signals, links and history all transfer slower than page level equity. Keep the old domain's redirects live for years, not months.

Ecommerce migrations: multiply everything by the catalogue: product and collection URLs, faceted navigation, out of stock handling and review data all need their own mapping decisions and revenue makes the monitoring windows tighter. Our SEO for ecommerce industry page covers the vertical's wider strategy.

Consolidating multiple sites: this is a migration per site. Each old domain gets its own crawl, benchmark and redirect map into the merged structure, sequenced rather than simultaneous so problems are attributable.

Local SEO through a migration

The layer local businesses forget:

• The Google Business Profile website link needs updating the day the domain changes (and its UTM tags rebuilt)

• Location and suburb pages are usually the URLs that change most in a restructure and their redirects protect the local rankings that took years to earn

• Citations across directories need the new domain rolled through (start with the majors, the long tail can follow)

• Schema markup on location pages (addresses, hours, LocalBusiness types) needs to survive the rebuild intact

A migration that protects the blog but breaks the suburb pages has protected the wrong asset for most service businesses. Our guide on how near me searches work across Melbourne explains why local page equity is worth protecting. Use our local SEO readiness checker pre and post migration to catch local signals that dropped.

Who does what: the migration responsibility map

One person can hold multiple roles, no role can be held by nobody. Every migration horror story has an empty cell in this table, usually the third row.

The realistic timeline

Compressing this into a fortnight is possible and occasionally necessary. Compressing it into a weekend is how the seven causes happen. The SEO work adds days to a rebuild project measured in months and it is the days that protect the investment the months represent.

What migration SEO work costs (and what skipping it costs)

For a typical small business site, the safety net (crawl, benchmark, redirect map, QA, monitoring) represents roughly 15 to 40 hours of specialist work depending on site size: commonly $2,000 to $6,000 as a project line, more for large or ecommerce sites.

Against it: the rebuild itself usually costs $5,000 to $30,000+ and a botched migration commonly costs 30 to 50% of organic leads for a quarter or more plus remediation. The migration SEO line is the insurance premium on everything else in the budget. Our guide on how much a website costs in Melbourne covers the broader budget picture and our guide on hidden website costs covers the line items that surprise people.

Our guide on how long SEO takes to work explains the compounding that is actually at stake: a migration that loses two years of accumulated equity does not cost a bad month, it costs the two years. Any SEO services Melbourne engagement should include migration support as standard if a rebuild is on the horizon.

A worked migration: the Elev8d relaunch

Our own June 2026 rebuild as the honest case study: full restructure onto root level URLs, meaning nearly every address changed, on a site whose organic visibility was the business's main channel. The safety net ran exactly as this guide describes: complete crawl and benchmark pack filed before design sign off, a redirect map covering every legacy URL one to one into the new structure, staging QA including the noindex sweep and the bulk redirect test in list mode before and after the switch.

What the monitoring windows showed: the first fortnight produced exactly the nervous graph the normal dip section describes. Impressions wobbling as recrawling worked through the map, individual terms bouncing, nothing structural in the 404 report beyond a handful of long tail misses added to the rules within days. By the following month the curve had flattened and turned and the quarter after the migration became the site's strongest on record for impressions, because the new structure the migration protected was also genuinely better.

The lesson we would underline from the inside: the daily check against a written baseline is what made the wobble boring instead of frightening. And boring is precisely what launch month should feel like.

When not to migrate

Some migrations should not happen or should not happen yet. Signals to pause:

• The only motivation is aesthetic boredom (a template refresh carries no URL risk and costs a fraction)

• Peak season is inside the next two months (launch after the season, not into it)

• Nobody can produce the redirect map (the migration is not planned yet, whatever the timeline says)

• The new platform cannot support what currently ranks (losing the blog, the location pages or the URL control is a rankings decision wearing a design costume)

Redesigns are worth doing and worth doing properly. The ones that go wrong are almost always the ones that were really just impatience with a launch date.

What we recommend at Elev8d

We rebuilt and migrated our own site in June 2026 (new structure, root level URLs, full redirect layer), so this checklist is not theoretical. The benchmark pack, the one to one redirect map and the daily monitoring window are exactly what we ran and the traffic curve did what the normal dip section describes: wobbled, flattened, then improved on the stronger foundation.

It is also the process we run for clients, where web design Melbourne projects and SEO planning happen as one workstream rather than a redesign SEO gets bolted onto afterwards, because the entire lesson of this article is that the bolt on version is where the losses come from.

Two tools worth using before any rebuild: our website cost estimator sets the budget conversation on real numbers with the redirect and SEO work costed in and our SEO ROI calculator shows what the organic traffic at stake is actually worth, which is the number that justifies the safety net line item.

FAQs

Answers to the questions we hear most about website migrations.

Our staging site got indexed by Google. How bad is it?

Common and fixable: password protect or noindex the staging environment, request removal of the staging URLs through Search Console if they are showing for real searches and make sure the live site's canonicals point at itself. The lasting lesson is preventative: staging environments get protected on day one of every project.

Do 301 redirects pass all the ranking equity?

Google has said for years that 301s pass PageRank without loss and in practice a clean, relevant, single hop redirect preserves equity about as well as anything can. What loses equity is everything around the redirect: irrelevant destinations (the homepage dump), chains, delays in implementation and the content at the destination no longer matching what earned the links.

How long do 301 redirects need to stay in place?

Years. Google recommends keeping them for at least a year for signal transfer, but the practical answer is longer: backlinks, bookmarks and old listings keep sending real people through them indefinitely. Treat redirects as permanent infrastructure.

Will my rankings definitely drop after migrating?

Not definitely. Expect fluctuation for one to four weeks while re-evaluation happens, stabilisation by four to eight and often improvement beyond, since the new site usually is genuinely better. Sustained losses come from the seven causes above, all preventable. The businesses that report migration disasters almost never had a redirect map.

Can I improve content during the migration or should everything move as is?

Move the money pages substantially as is, improve them afterwards as a separate project. Changing URLs, templates and content simultaneously means a post-launch problem has three suspects and no witnesses. One variable at a time is the diagnostic gift you give future you.

My developer says redirects will slow the site down. True?

A properly implemented server level redirect adds milliseconds to old URL requests only. New URLs are untouched. What genuinely hurts is redirect chains and internal links pointing at redirects, both of which this checklist eliminates. If a developer resists implementing the redirect map, that is a capability conversation, not a performance one.

The old site had pages I am embarrassed by. Can I just delete them?

Check the benchmark pack first: embarrassing and valuable are not mutually exclusive and a dated 2019 article quietly ranking for a commercial term deserves a rewrite at its URL, not deletion. Pages with genuinely nothing (no traffic, no links, no rankings) can be retired honestly with a 410 or 404. The crime is not deleting pages, it is deleting them without looking.

Next steps: pick your path

Rebuild coming up? Start the crawl and benchmark pack this week, before design begins. Put the redirect map in every brief and contract.

Mid project without a plan? Pause the launch date, not the project. The safety net takes days to build and the launch only happens once.

Already migrated and the traffic dropped? Talk to us before you rebuild anything else: send the launch date and Search Console access and we will diagnose which of the seven causes it was. Most post-migration losses are recoverable when the diagnosis is right and speed matters.

Sources and further reading

• Google Search Central: Site Moves and Migrations. Google's own documentation for URL changing moves.

• Google Search Central: Redirects and Google Search. How redirect types are treated and why 301s carry the signals.

General information only. Rules vary by situation, particularly around advertising claims, privacy, reviews and consumer law. If you are unsure about compliance, get professional advice.