Reference

Glossary of website migration SEO terms

The vocabulary that comes up in migration planning meetings, defined plainly, with a note on why each one matters when a site moves.

301 redirect also: Permanent redirect

An HTTP status code telling a client that a resource has permanently moved to a new URL. The 301 is the standard instruction for any permanent URL change and is what passes accumulated ranking signals from the old address to the new one. Browsers cache 301 responses aggressively, which means a redirect pointed at the wrong destination is unusually awkward to unwind — returning visitors may keep following the cached rule after the server has been corrected. Search engines follow a chain of redirects rather than only the first hop, but signals consolidate most reliably when the chain is a single hop from the original URL directly to its final destination.

In a migration This is the workhorse of every migration. Nearly every row in a redirect map is a 301, and its destination has to be a genuinely relevant page or the signal is discarded.

410 Gone

An HTTP status code stating that a resource has been intentionally removed and will not return. It differs from 404 Not Found in intent rather than in effect: 404 means the server cannot find the resource, while 410 means the server knows it is gone on purpose. Search engines generally drop 410 URLs from the index faster than 404s, since the removal is declared rather than inferred. It is the correct response for content deliberately retired with no equivalent destination, and the wrong response for anything that might come back or that has a reasonable replacement elsewhere on the site.

In a migration The honest alternative to a homepage catch-all. Pages with no relevant destination should return 410 as a recorded decision rather than 404 by accident.

Canonical tag also: rel=canonical, Canonical link element

An HTML link element declaring which URL is the preferred version of a page when several URLs serve substantially the same content. It is a strong hint rather than a directive, so search engines may select a different canonical if other signals disagree — internal linking, sitemap inclusion and redirects all contribute to that judgment. A page should always reference itself as canonical unless it genuinely is a duplicate. Canonical tags built from a hardcoded base URL are a frequent source of error, because they survive environment changes that they should not survive.

In a migration Canonicals that still point at the old domain, protocol or host after a move keep signals pinned to a site you are trying to retire. Verify them in production.

Change of Address also: Search Console Change of Address tool

A Google Search Console feature used to notify Google that a site has moved from one domain to another. It requires both the source and destination properties to be verified and a working sitewide redirect to be in place before it can be submitted. The tool announces intent and speeds up processing; it does not transfer anything by itself, and it cannot compensate for redirects that point at irrelevant destinations. It applies only to moves between different domains, and has no role in protocol changes, subdomain consolidations or URL restructures on the same host.

In a migration File it within a day of cutover on a domain change, after the sitewide redirect is verified. Filing before the redirects are live either fails or does nothing.

See also: migrations/domain name change

Crawl budget

The number of URLs a search engine will fetch from a site in a given period, determined by how much load the server can handle and how much demand there is to crawl the content. It matters mainly on large sites: below a few thousand URLs, budget is rarely the limiting factor. Budget spent on redirect chains, faceted duplicates, parameterized variants and low-value archives is budget not spent on the pages that earn revenue. Server log analysis is the only way to see how it is actually being allocated rather than how you assume it is.

In a migration Migrations spend budget heavily as the new URL set is discovered. Chains and index bloat compete for that budget precisely when discovery speed matters most.

Crawl parity also: Parity diff

The condition where a new site presents the same set of indexable URLs, carrying the same signals, as the site it replaces — allowing for changes that were intended. Verifying it means crawling both environments under identical configuration and comparing field by field: titles, meta descriptions, headings, canonicals, meta robots, word counts, internal link counts, structured data and hreflang. The comparison is the point. Individual pages can look correct in a browser while a whole template has quietly lost its canonical tag or half its internal links.

In a migration The single highest-value pre-launch check for any replatform or redesign, and the one most often dropped when the build runs late.

DNS TTL also: Time to live

The duration, in seconds, that a DNS record may be cached by resolvers before it must be looked up again. A high TTL reduces lookup traffic; a low TTL makes changes propagate quickly. Typical production values sit between one hour and one day, which means a record changed at cutover may take that long to reach every resolver. Lowering TTL shortly before a migration shortens propagation to minutes and makes rollback practical. The reduction has to be made at least one full TTL period in advance to take effect, which is why it is scheduled rather than done on the day.

In a migration Lower it to around 300 seconds 48 hours before cutover. Doing it on launch day is too late for the change itself to have propagated.

hreflang also: Language annotation

An annotation declaring the relationship between language or regional versions of the same content, implemented in the HTML head, in XML sitemaps or in HTTP headers. Each version must reference every other version including itself, and that reciprocal confirmation is what validates the cluster. Language codes follow ISO 639-1 and optional region codes follow ISO 3166-1 Alpha 2 — a common error is using uk for the United Kingdom, where the correct region code is gb. An annotation pointing at a redirected or missing URL is ignored, which breaks the cluster for every member.

In a migration Annotations must be regenerated to point at final destination URLs, and all locales must update in the same deploy or return tags stop reciprocating.

Index bloat

A condition where a large number of low-value URLs are eligible for indexing, diluting crawl budget and making a site appear thinner in aggregate than the quality of its main content would suggest. Typical sources are faceted navigation, pagination parameters, sort orders, session identifiers, tag archives and internal search result pages. It is diagnosed by comparing the count of indexable URLs against the count of URLs that genuinely deserve to rank. Remedies include canonical tags, robots rules, parameter handling and routing changes that prevent the URLs from being generated at all.

In a migration New platforms frequently generate URL patterns the old one did not. Compare indexable counts before and after — a large unexplained rise is bloat, not growth.

Log file analysis also: Server log analysis

Examination of raw server access logs to see which URLs crawlers actually request, how often, and what status codes they receive. It is the only data source that reports crawler behavior directly rather than inferring it from indexation outcomes. Analysis typically groups requests by user agent, site section and status code to reveal where crawl budget is being spent and which sections are effectively invisible. Access is often the obstacle: logs may be short-lived, held by a CDN rather than the origin, or require a request to a team that does not usually field them.

In a migration After launch, the shift in crawler requests from old URLs to new ones is the only direct evidence of how discovery is progressing.

Meta robots also: noindex tag, X-Robots-Tag

A directive controlling whether a page may be indexed and whether its links may be followed, expressed either as a meta tag in the HTML head or as an X-Robots-Tag HTTP response header. Unlike robots.txt, which governs crawling, meta robots governs indexing — and it can only be obeyed if the page is actually crawled. The header form is invisible in page source, so a page can carry a noindex directive that no amount of viewing the markup will reveal. Both forms must be checked in production after a release, not inferred from staging.

In a migration A staging noindex reaching production can start removing pages within days. It is the most expensive thirty-second check on any launch runbook.

Redirect chain also: Redirect hops

A sequence in which one URL redirects to another which redirects again before reaching a final destination. Googlebot follows up to ten hops in a chain before abandoning it, so most chains still resolve — but every hop consumes crawl budget and adds latency for users arriving from external links. Chains usually form by accident, when two independently correct rules run in sequence: one normalizing the protocol, another normalizing the host or trailing slash. A redirect loop, where the chain returns to a URL it has already visited, never resolves at all.

In a migration Aim for exactly one hop and never ship more than two. Chains also form after launch when later releases add rules that run before the migration rules.

Redirect map also: URL mapping document

A row-per-URL document recording where each old URL goes on the new site, which status code is used, its priority tier, and the reasoning wherever the mapping is not one to one. It is the artifact that makes redirect coverage checkable — without one, nobody can confirm whether a specific URL was handled, including the person who wrote the redirect rules. A working map includes an owner column for rows requiring a content decision and a verification date recording when each row was last tested against the target environment.

In a migration The central deliverable of any migration where URLs change. Its completeness is the single strongest predictor of whether traffic survives.

Rendering also: JavaScript rendering, Client-side rendering

The process of executing a page's JavaScript to produce the final DOM a user sees. Search engines fetch the raw HTML first and queue pages for rendering separately, so content that exists only after JavaScript execution is discovered later than content present in the initial response — and internal links that only appear after hydration are followed slowly if at all. The practical test is to view source rather than inspect element: the inspector shows the rendered DOM, which is exactly what the raw response may not contain.

In a migration Moving to a JavaScript framework can silently shift body copy, internal links and canonical tags out of the raw HTML. Verify per template before launch.

robots.txt

A plain text file at the root of a host giving crawlers instructions about which paths they may fetch. It governs crawling only. A URL that is disallowed can still be indexed if it is linked from elsewhere, typically appearing in results without a snippet because the page was never fetched. Blocking a URL also prevents any noindex directive on that page from being seen, which makes robots.txt actively counterproductive as a way of keeping content out of the index. It is the wrong tool for protecting staging environments; authentication is the right one.

In a migration Check it in production immediately after cutover. A staging robots.txt shipping to production is one of two classic launch-day failures.

Term missing?

If something came up in your migration planning that is not defined here, send it over. The glossary grows from real questions rather than a keyword list.

Email us about your migration Message on WhatsApp

Reply: Typically within one business day.
Hours: Monday to Friday, 9am–6pm US Eastern.
First call: Scoping, not a pitch. No deck.