How does probeo help with a site migration?

The new site was ready. The traffic was on the old one.

They had rebuilt the site and had not cut over. The old site had years of search traffic behind it, and the worry was simple: moving would lose it.

A migration is the one moment every page on a site changes at the same time. Whatever is wrong on the new site goes live everywhere on the same day.

What they could not see

• Which old pages had no matching page on the new site.
• Which pages would lose their title, headings or structured data in the move.
• Which of the pages carrying their search traffic were affected, and which were fine.
• Whether the new site was slower or heavier than the one it replaced.

What probeo showed them

The new site read the way search engines read it, next to the old one, ordered by what each page was worth in search. The problems that would cost traffic came first.

Not a list of every difference. A migration always changes things. The question was which changes would cost them.

What they did with it

They had a list of what to fix before switching instead of after it, with the pages that mattered most at the top.

What is worth comparing before you switch

Five things are checkable on both the current site and the replacement, and each one is a documented failure mode rather than an opinion.

• Indexability. The noindex rule tells Google not to show a page, and it can arrive in an HTTP header as well as in the HTML, so a check that only reads markup misses it.
• Robots rules that cancel each other. A page disallowed in robots.txt never has its indexing rules read, so a noindex on a blocked page is ignored, and where rules conflict the more restrictive one applies.
• Canonicals. A declared canonical is a strong signal, not a binding instruction, and Search Console reports the canonical you declared next to the one Google selected.
• Structured data. Items missing required properties are not eligible for rich results, markup must represent visible content, and a violation can cost a page its rich result without affecting how it ranks.
• Accessibility conformance. It is defined for full pages, cannot be claimed if part of a page is excluded, and where a page is part of a process every page in that process has to conform.

One more that is easy to get wrong: blocking a page in robots.txt is not a way to keep it out of Search, because a disallowed page can still be indexed if other sites link to it.1,2,3,4,5,6

Performance, measured the way it is graded

The Core Web Vitals thresholds are published: largest contentful paint within 2.5 seconds, interaction to next paint at or under 200 milliseconds, cumulative layout shift at or under 0.1, assessed at the 75th percentile of page loads across mobile and desktop.7

A lab run before launch does not establish that. Lab data deliberately limits variables, may not represent the 75th percentile of field data, and interaction latency cannot be measured in the lab at all because there is no user.8

So a pre-cutover performance check is a comparison, not a certificate. The new page against the old one, measured the same way, on the same pages.

After you switch: what settling looks like

The settling period has a documented shape, and knowing it stops a team from treating normal behavior as an emergency.

• Crawling goes up, not down. After a migration Google temporarily crawls the new site more heavily than usual, because crawls of the old site get redirected to it.
• A hosting move with no URL changes does the opposite. A temporary drop in crawl rate right after launch is normal, followed by a steady increase over the following days.
• The two sitemaps cross over. Indexed pages from the old URLs drop toward zero as indexing of the new URLs rises.
• The newest data is provisional. Recent Search Console performance data can still be collected and may change, and it is marked on the graph.

The crawl stats report covers the last 90 days and splits requests by response code and by purpose, separating discovery of new pages from refreshes of known ones, which is where a stalled move shows itself first.9,10,11,12

What is actually broken

The rest is not settling. These are states a page can sit in indefinitely, and each one is reported rather than inferred.

• Redirect errors. Google reports chains that are too long, loops, URLs that exceed the maximum length, and bad or empty URLs in a chain, and its crawlers follow up to ten hops by default.
• Google choosing a different canonical than you did. The inspection tool shows the canonical you declared and the one Google selected, and the indexing report names the case where they disagree.
• Soft 404s. A page that says not found while answering 200, which client-rendered apps produce by default, leading to error pages being indexed.
• Pages crawled and not indexed, or discovered and not crawled. Both are named states with different causes: the second usually means Google expected to overload the site and rescheduled.

Old URLs showing as pages with redirects are not an error. That is Google reporting a non-canonical URL that redirects, which is what you asked it to do.13,14,15,4,3,16

What this cannot tell you

It cannot tell you how Google will treat the new site on the day you switch. Rankings move while Google recrawls and re-evaluates moved pages, and that settling takes weeks.9

What it can tell you is whether a drop afterward is the move settling or something that was broken before you switched.

Google publishes no pre-cutover comparison procedure. Its move guidance starts at launch, so the checks above are documented individually while the idea of running them as a gate is not.9

There is also no documented figure for how much of a dip is normal during a move, and an automated accessibility check cannot establish conformance: W3C is explicit that no tool alone can determine whether a site meets the standard, and that knowledgeable human evaluation is required.17

Google publishes no recrawl interval for a known URL, so nobody can tell you when a specific page will be picked up again. Requesting indexing through the inspection tool typically takes a day or so and can take much longer, and Google does not guarantee that a page will be crawled, indexed or served at all.4,18

Questions