Technical SEO & URL Architecture / Base2026 Syndication
Canonical tag vs redirect: Keep the page or move it?
Use a canonical preference when duplicate or very similar pages need to stay available. Investigate a permanent redirect when an old address has been retired and a real replacement exists. A canonical does not move the visitor.
Choose the behavior, then the mechanism
A canonical preference keeps a useful alternate page available while identifying a preferred URL. A redirect changes the request destination. Choose based on what the visitor should experience, not only on which URL you want Google to show. [1] [2]
| Situation | Direction to investigate | Check first |
|---|---|---|
| A duplicate print view remains useful | Keep it available and declare a canonical preference | Does it contain the same main information as the preferred page? |
| An old address has permanently moved | A permanent server-side redirect, such as 301 or 308 | Does the destination genuinely replace the old page? |
| The move is temporary | An appropriate temporary redirect | When and how will the original URL resume? |
| The pages answer different questions | Keep their distinct purposes clear | Are shared keywords hiding different reader jobs? |
“A strong signal that the specified URL should become canonical.”
— Google Search Central [1]
Keep a print view; move a retired address
Fictional teaching example: a site retires https://example.com/boiler-service-old/ in favor of https://example.com/boiler-servicing/. A print version at https://example.com/boiler-servicing/print/ remains useful and contains the same main information.
The print page can put an HTML link element with rel=canonical and the absolute href https://example.com/boiler-servicing/ in its head. The browser still displays the print page. The annotation identifies a preferred URL; it does not navigate the reader. Google recommends an absolute URL for this declaration. [1]
For the retired address, the server might return HTTP/1.1 301 Moved Permanently with Location: https://example.com/boiler-servicing/ and Content-Length: 0. This is a teaching response, not a captured live response. The status and Location belong to the HTTP response, not an HTML tag. [2] [5]
A browser following that response requests the new address. Before implementing the move, inspect the destination. Sending a retired service URL to a vaguely related homepage is not the same as taking the visitor to replacement service information. [2]
Credit: OpenAI Media Service API / gpt-image 2.0; contextual teaching illustration [6]
Keep three observations separate
| Observation | What it establishes |
|---|---|
| Canonical found in supplied HTML | What that source capture declares. |
| Actual response and final destination | Whether the request moves and where the visitor arrives. |
| Google's reported canonical | The selection shown in the relevant indexed information. |
Page Source Check helps with the first row. It reads supplied HTML and optional URL context, not the live redirect chain or Google's choice. Use Search Console's indexed information separately; a live URL test does not replace the recorded canonical report. [3] [4]
Keep related signals consistent. A sitemap that prefers one URL and a canonical that names another deserve investigation. Do not use noindex as a shortcut for choosing a preferred page among duplicates: it concerns exclusion from Search, not visitor movement. [1]
Frequently Asked Questions
Does a canonical tag redirect the browser?
Does Google have to select my declared canonical?
When should I use a temporary redirect instead?
When the move is temporary and the original address is expected to resume. Choose the response behavior for that situation rather than describing a temporary move as permanent. [2]
Should different articles about one keyword share a canonical?
Not merely because they share a keyword. Compare their main content and reader jobs; canonical consolidation concerns duplicate or very similar pages. [1]
Write a decision someone can implement
For a real pair, record the URL being reviewed, candidate destination, content relationship, visitor requirement, duration, source observation and any actual response or Search Console observation. An uncertain replacement is a reason to inspect the content, not a reason to redirect faster. [2] [3]
For the fictional old service URL, the decision record is: visitor need, reach the replacement service page; content relationship, the new address replaces the retired information in this teaching setup; proposed behavior, permanent redirect; required verification, actual response, final URL and replacement content; separate observation, Google's selected canonical when available. [2] [3]
Decide what should happen to the visitor before deciding what to tell a search engine. Keep a useful duplicate available with a consistent canonical preference; move a genuinely retired address to its real replacement. If the pages serve different jobs or the move is temporary, reconsider consolidation or permanence. [1] [2]
Check the page's HTML at https://base2026.dev/tools/page-readiness/, then verify any redirect as a separate response-level change. The URLs and response in this article are teaching examples. [4]
Sources and checks
Links support the statements cited above. Different URLs do not automatically mean independent evidence.
- Google Search Central: canonicalization Checked .
- Google Search Central: redirects Checked .
- Search Console URL Inspection Checked .
- Base2026 Page Source Check Checked .
- RFC 9110: 301 Moved Permanently Checked .
- A06 contextual body figure Checked .
Answer-Ready Service Pages
Structure replacement service pages so visitors and AI search systems find clear, unambiguous answers to buyer questions.
AI Visibility Diagnostic Audit
Independent audit of public entity footprint, crawlability, and technical discovery before undergoing major URL migrations.
Start with public evidence
Unsure whether to keep or move key business URLs?
A Free AI Visibility Snapshot examines the observable public footprint of your business website, identifying canonical conflicts, crawl barriers, and delivery gaps.