SEO migration: switch domains or platforms without losing rankings (and prove it)
Almost everything ranking for "SEO migration" is an agency checklist ending at the 301 map. Nobody designs the measurement. This post treats a migration as an experiment: a baseline before, a sample by URL type, a go-live calendar, and thresholds for telling a real break from normal resettling.
TL;DR
- ✓ A migration is an experiment, not a checklist. The 301 map is one block; everything else is measurement design.
- ✓ Start the baseline 4 weeks ahead. Tracking history is not retroactive — add the keyword after the change and you have nothing to compare against.
- ✓ Measure by URL type, not the whole domain. Home, category, product and guide pages resettle at different speeds.
- ✓ Go-live day needs its own calendar, not a loose date. D-28 through D+56, with thresholds that separate normal resettling from a real error.
- ✓ The proof is not total traffic, it is the ranking URL. Confirm the new URL replaced the old one before you call it a win.
What an SEO migration is, and why rankings drop
An SEO migration is any change that alters the URLs Google already indexed your site under: a domain change, a platform change (WooCommerce to Shopify, Wix to WordPress), a redesign that changes the URL structure, a move from HTTP to HTTPS, or a hosting change that does not touch a single URL but can still shift load times and certificates. Each type carries a different risk, and it is worth knowing which is which before you panic over any of the four.
From lower to higher risk: a hosting change that touches no URLs is nearly invisible to Google; moving from HTTP to HTTPS is low-risk if the 301 is done right (it is the best-documented case and the one with the fewest surprises); changing the URL structure inside the same domain and platform raises the risk because you are touching paths one by one; switching platforms raises it further because the entire structure usually changes at once; and switching domains is the riskiest of the five, because on top of the URLs, the whole Search Console property changes too.
Google says as much in its own documentation on site moves involving a URL change: on a small or medium site, moving most pages can take a few weeks, and large sites take longer. "Temporary fluctuations are normal" while the crawler reprocesses everything. The problem is not that position moves in week one — that happens almost every time — it is not having a way to tell that normal movement apart from a real error that is still there in week four.
The migration as an experiment: the 301 map in one block
This is the part that is already documented everywhere, so it gets three rules and not one more. First: every old URL redirects 1:1 to its new equivalent — never dump everything on the homepage. A mass redirect to the homepage reads to Google as "this no longer exists," and you lose almost the entire signal. Second: no redirect chains — if the old URL was already the target of an earlier redirect, fix the hop so it goes direct. Third: keep the 301s active for at least a year; do not pull them the moment position looks recovered, because traffic and external links still arrive through the old path for a long time.
With the 301 map closed, the real work of this post begins: designing how you are going to know whether it worked. That is daily rank tracking with the ranking URL, and it is the rest of this guide.
Four weeks of baseline before the change
This is the part almost no guide mentions, and it is what makes everything else possible: before you touch a single URL, you need four weeks measuring how your keywords behave when nothing has changed. It is not methodological overkill. A keyword's tracking history starts the day you add it — it is not retroactive — so if you start measuring after go-live, you have nothing to compare against and your whole "verification" turns into an opinion.
Four weeks also buys you something more useful than a starting point: the background noise of each keyword group — how much its median position normally swings from one week to the next with nobody touching anything. Without that number, any movement after the change looks suspicious. With it, you know a one- or two-spot shift is normal noise and a six- or seven-spot drop is not.
Use this window for two more things: export the Search Console history of the old property before the change starts diluting the data, and check Google's Search Status Dashboard to confirm there is no major algorithm update running or freshly announced right in your window. If there is, the sensible move is to push the date back — you will not be able to separate what is yours from what is the market's.
A keyword sample by URL type
Measuring "the domain" is measuring nothing. A migration does not hit the homepage, category pages, product pages and content guides equally — each page type has its own template, its own internal-linking pattern and, almost always, its own resettling pace. That is why the sample is built by group, not as a flat list of keywords.
The ratio that works in practice is 5 to 10 keywords per URL type — home, category, product, guide or blog, plus any other group that matters for your site — which lands between 30 and 50 keywords total for a mid-sized project. Pick them from two sources: the highest-impression queries from Search Console for each page type, and the site scan, which shows which keywords your domain already ranks for with volume, current position and difficulty — useful for filling out groups Search Console does not cover well.
Work out the quota cost before you build the list, not after. Every keyword consumes one unit per country and language, plus one more unit for each city you track it in. Fifty keywords measured at country level fit comfortably inside Pro's 50 keywords; the same sample split across three or four cities is already pushing the limit, so decide upfront whether this migration needs city-level tracking or the country level is enough.
URL patterns by platform: Shopify, WooCommerce, Tiendanube, VTEX, Wix and the old Mercado Shops
Every store platform has its own URL grammar, and knowing it upfront is what lets you build the 301 map without guessing route by route. This table covers the prefixes you will run into most often on LATAM projects; always verify against your own store before calling the map final, since some platforms let you customize part of the path.
| Platform | Products | Categories / other | Note |
|---|---|---|---|
| Shopify | /products/{handle} | /collections/{handle}, /pages/, /blogs/{blog}/{article} | Fixed prefixes, not configurable |
| WooCommerce | /product/ (or /producto/ on Spanish installs) | /product-category/ (or /categoria-producto/) | Configurable in permalinks |
| Tiendanube | /productos/{slug}/ | Categories at the domain root | Verify each category's slug |
| VTEX | /{slug}/p | /{department}/{category} | Common in large LATAM retail |
| Wix | /product-page/{slug} | /post/{slug} for blog posts | — |
| Old Mercado Shops | Listing-ID-based URLs, in some cases ending in -_JM | Platform shut down Dec 31, 2025; confirm your store's actual pattern | |
"WooCommerce to Shopify" is the most common case in LATAM agency searches, and the one that leaves the least room to improvise: almost no WooCommerce URL matches Shopify's fixed structure, so the 301 map gets built route by route, no shortcuts. "Wix to WordPress" tends to be friendlier, since WordPress lets you configure permalinks to get closer to the old structure — though they rarely line up exactly.
On the old Mercado Shops: we mention it as context, not as a hook. The platform shut down on December 31, 2025, and plenty of stores in the region were forced to migrate in the months before that, mostly to Tiendanube or Shopify. If that is your situation, the whole protocol in this post still applies — the only difference is you probably no longer have room to build a baseline before the change, so start it today, even though the move already happened.
Go-live day and the 6 to 8 weeks after: calendar and thresholds
The part no 301 checklist writes down: what to watch each week and when to actually worry.
Migration measurement calendar
- D-28: start the baseline with the sample by URL type, and export the Search Console history of the current property.
- D-28 to D-1: calculate each group's background noise — how much its median swings from one week to the next with no changes.
- D-14: close the 1:1 301 map and check Google's Search Status Dashboard; if a core update is running, push the date back.
- D-7: freeze content, titles and internal links. Do not change the URL and the template on the same day.
- Go-live day: ship the change, turn on the 301s and the new sitemap; if you changed domains, submit Change of Address in Search Console. Re-check the full sample by hand in Rankiamos — each keyword can be manually re-measured every 6 hours, within your daily quota.
- D+1 to D+14: daily review — is the "ranking URL" already the new one? Did any group drop more than double its background noise?
- D+15 to D+56: weekly review of each group's median; from week 3 on, a group still declining is treated as an error until proven otherwise.
- D+56: before/after report by URL type, and a call on whether to close intensive monitoring or extend it two more weeks.
Thresholds are what turn this calendar into a criterion instead of a gut feeling. Use these four, defined with your own background noise, not a number you read somewhere else:
A drop bigger than double the background noise
If a group normally swings one or two spots week to week and suddenly drops six, that is not noise anymore — check it before another week passes.
A group still declining in week 3
Almost any drop is normal in weeks one and two. If a whole group — not one stray keyword — is still falling in week three, the diagnosis is no longer "fluctuation," it is an error until proven otherwise.
The old URL still ranking in weeks 3-4
If the "ranking URL" still shows the old path three or four weeks in, the 301 is not being served correctly, there is a redirect chain, or the sitemap still points at the old path.
A keyword that suddenly reads "30+"
Rankiamos reads the first 30 organic results; anything past position 31 shows as "30+". If a keyword that was steady in the top 30 suddenly falls to "30+", it is almost always a 404, an accidental noindex, or a robots.txt block on the new URL — check it that same day, do not wait for the weekly review.
Confirming the new URL replaced the old one in Google
Position alone proves nothing if you do not know which URL is generating it. For every keyword in your sample, the signal that actually matters is the "ranking URL" — available on Free and Pro: if it still shows the old path, the 301 has not fully taken hold with Google yet, even if position looks stable. Seeing both URLs alternate in a keyword's history, the old one one day and the new one the next, is exactly the transition phase the calendar above describes; what is not normal is seeing it still happen past week four.
If your migration kept the same domain — you changed platform or URL structure, not domain — the Pro plan's "possible cannibalization" flag hands you an extra signal without you asking for it: since it scans each keyword's last 10 measurements and flags it when different URLs from your own domain show up, a keyword still alternating between the old and the new path can end up flagged there too — which is exactly a sign that the replacement has not closed. It reads from the position history you are already tracking, not from Search Console, so it looks at the same data from a different angle. This is the point to monitor the old and new URL in parallel using the full method we already published, rather than repeating it here — the "Moment 1-4" template from that guide applies just as well to a migration as it does to a content merge.
Changing domains without losing SEO: Change of Address in Search Console, and running two series
On Google's side, when you switch domains or subdomains you use the Change of Address tool in Search Console: it explicitly tells Google you moved the site, and it only makes sense for that specific case. You need verified ownership of both properties and your 301s already live before you submit it. Keep the old property verified after the move — you keep seeing its data while Google works through transferring authority to the new domain.
On Rankiamos' side it is worth being just as direct: you cannot change the domain of an existing project. That is not a limitation of this post, it is how the product works today. If you need to track a domain migration with two parallel data series, the way to do it is with two projects: leave the old project tracking the same keywords against the old domain through your 6-to-8-week window — to confirm when its URLs stop showing up — and create a new project on the new domain to measure its own rankings from scratch.
State the cost upfront: the same keyword tracked in both projects consumes double your quota for as long as the parallel monitoring runs. For a 40-keyword country-level sample, that is 80 of Pro's 50 keyword units, temporarily. Once the 6 to 8 weeks pass and the replacement is confirmed, the sensible move is to archive or delete the old project and free up that quota for the new domain.
Rankings after a migration: how to prove the result to the owner or the client
Nobody who signed off on a migration wants to hear "I think it went fine." They want the before/after series, by URL type, with dates attached. In week 2, what you show is the transition phase: URLs alternating, the occasional dip, and the explanation for why that is expected given the calendar. In week 8, what you show is the real comparison: each group's median before go-live against its median over the last two weeks, and the state of the "ranking URL" for every keyword in the sample.
To hand over that result without building the table by hand, Pro exports a one-page PDF with your name and accent color — no logo, this is not a white-label document — showing top 3, top 10, top 30 and total counts, plus a table of up to 28 keywords; and a plain CSV with no branding as a backup of the raw data. Neither one sends itself — you export it and share it yourself, the same way you already do with the monthly report you already hand your client.
One expectation worth lowering before the meeting: a well-executed migration recovers the ground you had, it does not improve it by magic. If a keyword sat at position 8 before the change, the realistic goal is getting back to position 8, not jumping to 3. If the migration also comes with real content improvements, those gains do show up — just separately, and later, never in the first eight weeks.
What it costs to measure a migration, and what a rank tracker does not do
Add up the four weeks of baseline and the six to eight of follow-up monitoring, and the full protocol in this post spans roughly two to three months. In subscription time that is three months of Pro, or the $119.99 annual plan, which also leaves your tracker running the rest of the year instead of canceling it the moment the case closes. Pro's 14-day free trial requires an Apple payment method to activate, and Apple bills the subscription if you do not cancel before it ends — those days only cover the first week after go-live, not the baseline or the follow-up weeks.
And it is worth being just as clear about what the tool does not do. Rankiamos is not a crawler: it does not crawl your site looking for broken links, it does not verify that every 301 returns the right status code, and it does not detect redirect chains. For the 301 map itself you need a dedicated site crawler; what Rankiamos gives you is everything that comes after the map is already live — the measurement that proves whether it worked.
With that said, what is left is putting Rankiamos' migration-monitoring tools to work on your own case: the sample by URL type, the calendar and the thresholds from this post.
How Rankiamos handles it
The product measures the experiment. It does not design your 301 map or make the call for you.
History from the day you add the keyword
A baseline only exists if you start measuring weeks before go-live. History is never retroactive — wait until after the change to add a keyword and there is nothing to compare against.
Daily checks and manual refresh (Pro)
Daily automatic checks on Pro, plus the ability to re-measure any keyword by hand every 6 hours within your daily quota — the resolution you need on go-live day and through the two weeks after.
Ranking URL (Free and Pro)
The proof that the new URL replaced the old one, keyword by keyword, instead of guessing from position alone.
Search Console and site scan (Free)
The two sources for building your sample by URL type, included from the free plan; CSV export (Pro) sits as a backup of the raw data.
Honest note: Rankiamos is not a crawler and does not audit redirects — you build the 301 map with another tool or by hand. Every keyword's history starts the day you add it, so if you did not start the baseline in time, there is no way to reconstruct it after the fact. Cannibalization detection only works within the same domain — if you changed domains, you are the one confirming the replacement by comparing the two projects, not the tool doing it automatically. And Pro's 14-day trial requires an Apple payment method and only covers the first week of the protocol; you pay for the rest.
Free plan: 2 projects, 6 keywords, weekly check · iPhone · iPad · Apple Watch
Frequently asked questions
How long does it take Google to resettle an entire site after a migration?
Do you lose rankings when you change domains?
What does the Change of Address tool in Search Console do, and when should I not use it?
What happens to my URLs if I migrate from WooCommerce to Shopify?
My store was on Mercado Shops and I already migrated — how do I know if I got my rankings back?
Is it a bad idea to migrate during peak season?
Can I measure a migration on Rankiamos' Free plan?
Read also
Keyword cannibalization: detect it and confirm the fix
The other scenario of two URLs fighting each other, with the four-week verification template we reuse in this post.
Agency guideThe monthly SEO report for clients
How to build the report you will show the owner or client once you close out the migration result.