Indexing, Canonicals, Redirects, Duplicates, and HTTP Status Codes

Direct answer

Indexability decides whether a retrievable page may enter an index. Canonicalization decides which URL may represent a group of duplicate or very similar pages. A canonical tag does not make a page indexable, force Google to select it, or redirect users.

Choose the intended state first, then align the HTTP response, redirect, index directive, canonical, internal links, sitemap, and hreflang. Google describes redirects and rel="canonical" as strong canonical signals and sitemap inclusion as weaker. Conflicting signals create ambiguity, and Google can select a different canonical.

For removed content, return a meaningful 404 or 410 when no replacement exists. Redirect only to a relevant replacement. A 200 page that says “not found” is not a harmless design choice. It can become a soft 404 and waste diagnosis time.

What you will be able to do

By the end, you will be able to:

• distinguish indexability, indexing, canonical declaration, and canonical selection;
• predict how common HTTP states affect crawling and index eligibility;
• choose between permanent redirect, temporary redirect, canonical, noindex, 404, and 410;
• reconcile a duplicate URL cluster;
• diagnose “Google chose different canonical” without adding contradictory directives;
• classify technical combinations as invalid, contradictory, weak, or acceptable.

Four states that teams confuse

Indexable

The page has no known technical directive preventing index eligibility. It might still be excluded for duplication, quality, access, or system choice.

Indexed

The search system currently stores the page or its canonical representative.

Declared canonical

The site owner indicates a preferred representative through rel="canonical" in HTML or a supported HTTP Link header.

Selected canonical

The search system chooses a representative for a duplicate cluster based on multiple signals.

A page can declare itself canonical while Google selects another URL. The tag is a signal, not a command.

Why duplicate clusters exist

One resource can appear through:

• HTTP and HTTPS;
• www and non-www;
• tracking parameters;
• filters and sorting;
• print or PDF versions;
• mobile alternatives;
• session or tenant parameters;
• copied or translated content;
• old and new route structures.

Not every duplicate is a penalty. Canonicalization helps systems consolidate signals and choose a representative. The problem is uncontrolled inventory, conflicting ownership, wasted crawling, and the wrong URL appearing.

Canonical signals and their strength

Google's canonicalization documentation describes:

• redirects as a strong signal that the target should be canonical;
• rel="canonical" as a strong signal;
• sitemap inclusion as a weak signal;
• consistent internal links as supporting evidence.

Signals can stack. A permanent redirect to URL A, a self-canonical on A, internal links to A, and sitemap inclusion for A all communicate one state.

Signals can also conflict. A page that canonicals to A while links, sitemap, and hreflang support B creates avoidable ambiguity.

When to use each mechanism

Permanent redirect

Use 301 or 308 when users and crawlers should use the replacement URL going forward.

Examples:

• HTTP to HTTPS;
• old slug to new slug;
• merged page to the closest equivalent.

Google's redirect guidance treats server-side permanent redirects as strong canonical signals.

Temporary redirect

Use 302, 303, or 307 when the original URL remains the long-term owner and the move is temporary.

Do not leave a “temporary” state indefinitely without reviewing whether the intended ownership changed.

Canonical

Use a canonical when multiple accessible variants need to remain available but one should represent the cluster in search.

Examples:

• tracking parameters;
• print-friendly HTML;
• syndication or supported product variants.

Do not use canonical as a substitute for an appropriate redirect when the alternate should no longer be visited.

Noindex

Use noindex when a retrievable page should not appear in search. It does not consolidate the page's signals into another URL in the same way a redirect or canonical can.

Google's noindex guidance requires the crawler to access the page and read the directive.

404 or 410

Use 404 Not Found or 410 Gone when no resource and no relevant replacement exists. Do not redirect an obsolete page to an unrelated home page merely to avoid an error.

HTTP states and directives

Google's crawler status-code documentation explains how its crawlers process common responses. The table below is a practical summary, not a timing guarantee.

Google generally follows up to ten redirect hops, but production internal links should point directly to the final URL. A chain adds latency, failure points, and maintenance cost.

Soft 404s

A soft 404 happens when the response looks successful but the content communicates that the resource is absent or offers too little value to be treated as a real page.

Common patterns:

• 200 with “product not found”;
• empty category page with generic navigation;
• all retired pages redirected to the home page;
• client-side router shows an error but the server always returns 200;
• localized page contains no localized content.

Fix the state, not the label in a tool. Restore substantive content, return the correct absence status, or redirect to a truly equivalent resource.

A canonical reconciliation procedure

1. Define the canonical owner

Write the exact preferred URL, language, audience, content state, and reason it represents the cluster.

2. Inventory every variant

Include redirects, parameters, alternate protocols, print files, locale URLs, mobile routes, campaign links, and old slugs.

3. Fetch every response

Record:

• starting and final status;
• redirect chain;
• headers;
• source canonical;
• rendered canonical;
• robots and index directives;
• content similarity.

4. Map link and sitemap signals

Count internal links by target. Inspect sitemap inclusion, hreflang, structured data, feeds, and navigation.

5. Assign an intended state

Each URL should be:

• canonical owner;
• accessible duplicate with canonical;
• redirected alternate;
• non-indexable public utility;
• missing;
• private.

6. Remove contradictions

Align response, canonical, internal links, sitemap, and hreflang. Do not combine a canonical to one page with noindex merely to “make sure.”

7. Verify selected state

Use URL Inspection samples, index reporting, logs, and performance data. Google's selected canonical reflects processed signals, not your latest source file instantly.

8. Monitor regressions

Test templates, route rules, sitemap generation, and canonical data in continuous integration.

Worked example: a six-URL lesson cluster

The intended owner is:

https://www.seoryon.com/academy/canonicals/

URL 1: HTTP version

http://www.seoryon.com/academy/canonicals/

Desired state: direct 301 or 308 to the HTTPS owner. Remove all HTTP internal links and sitemap references.

URL 2: campaign parameter

https://www.seoryon.com/academy/canonicals/?utm_source=email

Desired state: the page can remain accessible, but declare the clean canonical and never use the tracked form in persistent internal navigation or the sitemap.

URL 3: printable PDF

https://www.seoryon.com/academy/canonicals.pdf

Decision depends on purpose. If the PDF is a useful separate printable asset, keep it accessible and decide whether it should be indexed. If the HTML lesson should own search, use a supported HTTP Link canonical or X-Robots-Tag policy according to the intended state. Do not put conflicting HTML markup inside a PDF.

URL 4: French lesson

https://www.seoryon.com/fr/academy/canoniques/

This is not a duplicate merely because it covers the same topic. If it is a true French translation, use its own self-canonical and reciprocal hreflang with the English version. Do not canonicalize French to English and then claim both locales are equivalent alternates.

URL 5: discontinued old slug

https://www.seoryon.com/academy/canonical-tags/

If the new lesson is the direct replacement, use a permanent redirect to the canonical owner. Update all internal links.

URL 6: accidental soft 404

https://www.seoryon.com/academy/missing-lesson/

The router returns 200 with “lesson not found.” Change the server response to 404, remove it from sitemaps, and ensure the error page still helps users navigate.

Why this can fail

The HTML can be perfect while a CDN redirect rule sends mobile users to a different host. Or the sitemap generator can reintroduce parameter URLs. Canonicalization is a system property, not one tag.

SEOryon canonical reconciler

Download the Indexability and Canonical Signal Reconciler.

For each URL, record:

• response state;
• crawl access;
• index directive;
• declared canonical;
• internal-link target;
• sitemap state;
• intended owner;
• conflict class;
• required action;
• verification source.

Use four conflict labels:

• Invalid: syntax or state cannot work as intended.
• Contradictory: strong signals point to different owners.
• Weak: direction is plausible but incomplete.
• Acceptable: signals and intended state align.

Failure modes

Canonical to a blocked URL

The crawler cannot reliably process the target content. Keep canonical owners accessible.

Canonical plus noindex without a documented reason

One says consolidate with another representative while the other excludes this page. Choose the actual desired state.

Redirecting every removed page to home

The home page is not a replacement for every resource. Return absence when appropriate.

Chains and loops

Link directly to the final URL and test route rules before release.

Temporary redirects left forever

Review whether the move is still temporary. Align the status with ownership.

Hreflang and canonical conflict

Each true language version should normally be self-canonical and reference equivalent alternates.

Exercise: diagnose eight combinations

Classify each as invalid, contradictory, weak, or acceptable, then choose a fix.

1. 200, index allowed, self-canonical, internal links and sitemap agree.
2. 200, noindex, robots disallow.
3. 301 to B, while all internal links still point to A.
4. 200 parameter page canonicals to clean URL, clean URL self-canonicals.
5. 200 “not found” page with self-canonical and sitemap inclusion.
6. French page self-canonicals, English page self-canonicals, reciprocal hreflang.
7. 302 used for a two-year permanent migration.
8. 404 deleted page with no replacement, removed from sitemap and links.

Answer key

1. Acceptable.
2. Contradictory because crawl blocking can hide noindex. Allow crawling until the directive is processed, unless another state is intended.
3. Weak and inconsistent. Update links to B and remove A from sitemap.
4. Acceptable when content is equivalent and parameters have no distinct purpose.
5. Invalid intended state and soft-404 risk. Restore the page or return 404.
6. Acceptable for true equivalent translations.
7. Weak or contradictory to the permanent intent. Use an appropriate permanent redirect after validating the migration.
8. Acceptable.

Final checklist

• The canonical owner is defined before implementing tags.
• Every URL variant has an intended state.
• Redirects are direct and point to relevant replacements.
• Permanent and temporary status match the real ownership.
• Missing pages return 404 or 410.
• Public content pages do not return misleading 200 errors.
• noindex pages remain crawlable until the directive is processed.
• Canonical targets are accessible and index eligible.
• Source and rendered canonicals agree.
• Internal links point to canonical final URLs.
• Sitemaps contain canonical public URLs only.
• Hreflang and canonical signals do not conflict.
• URL Inspection distinguishes declared and selected canonical.
• Logs and performance data monitor the resulting owner.
• Template and routing tests prevent regressions.

Frequently asked questions

Why did Google choose a different canonical?

Google combines content similarity, redirects, canonical declarations, internal links, sitemaps, protocol preference, and other signals. Inspect the whole cluster rather than repeating the same tag.

Does a canonical redirect users?

No. It communicates a preferred representative to supporting systems. Use an HTTP redirect when users should go elsewhere.

Is duplicate content a penalty?

Duplicate URLs often create clustering and selection problems rather than a manual penalty. Spam policies can apply to manipulative scaled duplication, but ordinary technical duplicates need signal consolidation.

Is 410 faster than 404?

Both represent absence. Do not promise a universal removal time. Use the status that truthfully describes the resource.

Should every page self-canonical?

Self-canonicals are a useful consistency practice for canonical indexable pages, but they do not replace correct redirects, links, sitemaps, or content decisions.

Sources and methodology

Community research showed that “Google chose a different canonical,” old URL structures, parameter duplicates, and canonical errors remain difficult to explain in plain language. Those questions shaped the worked example. Community advice was not used as technical authority.

1. Google Search Central, Consolidate duplicate URLs, updated 10 July 2026 and checked 28 July 2026. Official canonical-signal guidance.
2. Google Search Central, Block indexing with noindex, updated 10 December 2025 and checked 28 July 2026. Official index-directive requirements.
3. Google Search Central, Redirects and Google Search, updated 14 April 2026 and checked 28 July 2026. Official permanent and temporary redirect guidance.
4. Google Crawling Infrastructure, HTTP status codes, updated 4 February 2026 and checked 28 July 2026. Official crawler response behavior. Timing remains non-guaranteed.

Previous: Crawlability and Discovery
Next: JavaScript SEO and Agent-Readable Interfaces