Zhihe GrowthDuplicate URLs often arise from HTTP versus HTTPS, www versus non-www, trailing slashes, query parameters, CMS category pages, and older versions. If specifications, company names, or dates differ, search systems must decide which version reflects current facts. URL governance gives the same content a stable preferred address while retaining necessary language and regional differences; it does not mean keeping only one page.
Three Signals with Different Roles
| Signal | Use Case | What It Cannot Replace |
|---|---|---|
| Permanent redirect | An old address has genuinely been retired or merged | Not suitable for two pages that must both remain independently accessible |
rel="canonical" |
Duplicate or highly similar pages remain accessible | Not a command that forces indexing |
| URL in sitemap | Suggest canonical pages the search engine should discover | Neither guarantees indexing nor resolves conflicting canonical signals |
Google Canonical URL documentation distinguishes redirects, canonical tags, and sitemaps as signals of different strengths; a canonical declaration is a preference, not a command. A sitemap listing A, an on-page canonical pointing to B, and many internal links to C send conflicting signals. For same-language duplicates, decide which pages to retain and why. Genuine multilingual versions should not all canonicalize to the Chinese homepage.
Fields for a URL Inventory
Export sitemap URLs, internal links, server-accessible URLs, and URLs discovered by platforms, then cluster them by content summary. Record original URL, final response, preferred URL, canonical tag, robots/noindex status, sitemap inclusion, internal entry points, external links, and primary fact version. For migrations, configure permanent redirects first, then update internal links and sitemaps. Retain the old-to-new URL map so valuable links do not become 404s.
lastmod should reflect meaningful actual updates, not daily bulk date changes intended to manufacture freshness. Google's sitemap guidance recommends listing canonical URLs and explains that submission is only a hint. Zhihe Growth's site should likewise keep new encyclopedia pages, updated older pages, and sitemap records aligned, without including drafts, private client areas, or duplicates in the public inventory.
Retesting and Common Misinterpretations
After changes, sample preferred pages for status 200, old pages for expected redirects, canonical targets for accessibility, robots rules for unintended blocks, and sitemaps for canonical pages intended for indexing. Use GSC or similar tools to inspect the selected canonical. If the platform chooses differently, investigate content similarity, internal links, and external signals rather than repeatedly submitting the sitemap. Google's troubleshooting guidance emphasizes first checking the page actually selected by the system.
Zhihe Growth connects URL governance to the SSOT Fact Table: maintain legal-entity facts, service scope, and case results on canonical pages, then use the Explanation Chain to guide readers to that version. For multilingual sites, see Hreflang and Regional Canonical URLs; do not mistake regional versions for ordinary duplicates.
An Example URL Decision
Suppose /guide.html is the current technical guide, /guide-old.html is a retired path, and /guide.html?utm_source=mail is used only for traffic attribution. Permanently redirect the retired path to the current guide. The UTM-tagged page may remain accessible with a canonical tag pointing to the parameter-free version; list only that current URL in the sitemap. If an older guide contains still-useful historical methods that differ from the new version, do not redirect it to an unrelated article. Retain it with its version and applicable dates, plus a link to the new guide. Canonicalization is a content-relationship decision, not a rule to 301 every URL containing old.
Before launch, test the URL map: old paths reach the expected page, content is not blank or a soft 404, new pages self-canonicalize, internal anchor links are updated, and external links to old paths still reach the new content through redirects. Monitor old-URL requests, 404s, and canonical selection after launch; do not remove rules prematurely. If same-language regional pages differ in price, certification, or delivery, first assess their distinct value, then coordinate hreflang and canonical tags. Do not erase genuine regional differences just to reduce duplication.
Retain an auditable migration list of old URL, new URL, and reason. If several old paths lead to one page, verify that each still delivers the facts users sought. If one old page becomes several pages, choose the most relevant destination and link onward from it. Long redirect chains create more failure points; acceptance testing must check the final destination, not just the first 301.