Skip to main content

Canonical vs 301 redirect: which signal Google follows

Google still indexes a URL of mine months after I made it a 301. When to pick a canonical or a 301, and why 1 in 6 sites I checked leaves a duplicate homepage live.

Founder, CometWeb

I build CometWeb Insight, a website audit app that ends with a ranked to-do list. I write about what I measure on real sites: performance, SEO, accessibility and CO₂e.

Published:
16 min read

Google still indexes /tools/obfuscator on cometweb.io as a separate page with its own canonical, even though that address has returned a 301 since I reversed the redirect between its two versions in late August. Googlebot last visited two days before the change and hasn't returned. For canonical vs 301, Google lists a redirect first among canonical signals, but it counts only after Google sees it.

I checked both signals on 471 sites from the Tranco list, from the global top 10,000 to mid-ranking and Polish domains. On 82 of them, about one in six, the homepage answered on a second address without a redirect, and only 17 pointed every copy's canonical to the main version. The research pack has the data and scripts. Below you'll see when to use which and how to check your own site.

What a canonical tag and a 301 redirect each do

A canonical tag is a line in your page's <head> that names the preferred URL for content you serve under several addresses. A 301 redirect sends every visitor and crawler from one URL to another for good. The canonical leaves both pages reachable and asks search engines to show one of them. The 301 removes the old URL from the site, so people and bots only ever reach the new one.

The phrase "canonical redirect" has a second meaning you'll meet in search results. It also names a redirect that sends address variants (http, www, a missing slash) to the one form you chose. WordPress does exactly that with its built-in redirect_canonical() function.

Both meanings meet in the same place, because your canonical tags should name the same address your redirects lead to. For the bare definition, see the glossary entry on what a canonical URL is.

Canonical tag301 redirect
What the visitor seesthe page they asked forthe new URL
Old URL keeps workingyesno, it forwards
How Google treats ita strong signala strong signal, listed first
When Google learns about itwhen it crawls the duplicatewhen it recrawls the old URL
Works for PDFs and imagesonly as an HTTP headeryes
Easy to undoyesyes, but caches and links remember it

Which signal Google follows when canonical and redirect disagree

Google picks the canonical itself, and it says so plainly in its guide to URL canonicalization, where a stated preference is "a hint, not a rule". In its guide on consolidating duplicate URLs, Google ranks the methods by strength. Redirects come first and rel="canonical" second, both called strong signals, while sitemaps come last as a weak one. It also prefers HTTPS over HTTP as the canonical, unless other signals say otherwise.

So a working 301 usually wins. Once you redirect a URL, Google can't fetch the old page at all, so there's no tag left on it to disagree with. Temporary redirects (302 and 307) are different, because Google's redirect guide says it doesn't use them as a signal that the target should be canonical.

The catch is timing, and my own site shows it. Originally, /tools/code-obfuscator redirected to /tools/obfuscator, and then I reversed the direction. Search Console still reports /tools/obfuscator, the address that now redirects, as "Submitted and indexed". Google hasn't changed its chosen canonical either, because its last crawl came before the switch. My mistake wasn't the redirect itself. I changed the direction of an existing redirect and then did nothing to bring Googlebot back, and a redirect only counts from the day Google sees it.

After a redirect change, you can shorten that wait in three steps. Inspect your old URL in URL Inspection and click Request indexing, since that's how Search Console Help says you ask Google to fetch a single URL. Update your internal links and canonical tags so they name only the new address. Then submit a sitemap with the new URLs, as Google's site move guide advises, and a few weeks later inspect your old URL again to check it shows up as a page with redirect.

Google's documentation adds one rule that matters more than the ranking: don't name different URLs as canonical through different methods. A canonical tag, a redirect and a sitemap entry that point three ways force Google to guess.

Use a 301 when the old URL should disappear, and a canonical when both must stay

I decide with one question. If nobody should ever land on the old address again, redirect it. If both addresses have a job to do, keep both and add a canonical.

Use a 301 redirect for these:

  • moving a page to a new slug, merging two articles into one or retiring a product with a clear successor;
  • forcing one version of your domain: https and either www or the bare domain;
  • a site migration, where Google's site move guide asks you to keep redirects for at least a year.

Use a canonical tag for these:

  • URLs with tracking or sorting parameters, such as ?utm_source=newsletter or ?sort=price;
  • the same product listed in two categories with two paths;
  • a printable or AMP version of an article and content you republish on another domain.

Two choices I'd avoid. A canonical from a deleted product to its category is a bad idea, because the two pages aren't duplicates and Google may ignore the tag. A redirect on filter URLs is just as bad, because it breaks the filter for every visitor who uses it. Mass redirects of retired products to one category are risky too. Google's site move guide warns that many URLs sent to one irrelevant target may be treated as a soft 404.

Canonical, 301, noindex or robots.txt: one decision table

In audits I often see noindex or a robots.txt block where a canonical would fit better. You can pick by answering three questions: should people see the page, should it be indexed, and does the crawler need to fetch it?

OptionVisitor sees the pagePage in the indexSignals move to another URLTypical use
301 redirectnono, the target replaces ityes, most stronglychanged URL, merged pages, migration
canonicalyesusually no, the target is indexedyes, as a hintparameters, variants, syndication
noindexyesnonocart, thank you page
robots.txt blockyesmay be indexed without a descriptionnofilters and internal search, when you want to save crawling

The most common mistake from this table is a robots.txt block on a page that carries a canonical or noindex. The crawler stops at the block and never sees either signal, and Google's duplicate guide advises against robots.txt for canonicalization.

One in six sites I checked leaves a second homepage live

On most sites I checked the redirects already do the job, and the canonical only repeats what they say. But on about one in six sites I checked, the same homepage still answered on a second address, usually the www or the http version, without any redirect.

1 in 6 sites I checked serve the same homepage on a second address without a redirect ?
17 of 82 of those sites point every live copy to the main version with a canonical ?

That's the situation the canonical exists for, a duplicate you didn't remove. It caught the duplicate on only a small minority of those sites. Most had no canonical anywhere on the homepage, and a few had a canonical that pointed each copy to itself.

That last group is the one I'd look for on your own site. A template that builds the canonical from the address you requested will always "agree" with the page, so every audit of a single URL passes. But open the www and the bare version side by side, and each one names itself as the original, which tells Google nothing. Hardcode your domain and protocol in the canonical template, then add the redirect anyway.

A canonical that points to a redirect usually has the wrong host

The audit warning "canonical points to redirect" means your tag names a URL that answers with a 3xx instead of the page. It was rare in my sample, about 1 in 50 pages that carry a canonical (the breakdown is in the pack). But it comes from the template, so on most affected sites it repeated across several of the pages I checked.

On 13 of those 16 pages the redirect led straight back to the page carrying the tag. The tag says "the original is http://example.com/", the server answers "that's moved to https://www.example.com/", and Google gets two opinions from one site. That pattern points to a base URL in the CMS or SEO plugin that no longer matches the host and format your server redirects to.

The fix is one setting, so it's worth doing even though Google usually works it out. Point the canonical at the final URL, the one that answers 200, and check it with the canonical checker. I also found a handful of canonicals pointing at pages that return 404, and the fix there is the same: name the live page.

Five mistakes that make Google ignore your canonical

When something doesn't add up, Google picks the canonical itself. Its webmaster blog and troubleshooting guide list these five causes:

  1. Several canonical tags on one page, say from your theme and an SEO plugin, and Google will likely ignore them all.
  2. A canonical in the <body> instead of the <head>, usually from a script that adds the tag too late or broken HTML that closes the <head> early.
  3. A canonical pointing to a page with noindex, a 404 or a redirect, which names an original that isn't in the index.
  4. A relative URL or a typo in the host, so always write the full address with the protocol.
  5. A canonical that disagrees with your other signals, when internal links or the sitemap name a different URL than the tag.

Sitemaps are the third signal that often disagrees

A sitemap is the weakest of the three signals and the easiest to forget after adding redirects. In more than a quarter of the sitemaps I sampled, at least one URL from a small random sample redirected or carried a canonical pointing elsewhere. One such URL won't hurt you, but a sitemap full of them keeps telling Google that your old addresses still matter.

After any move, regenerate the sitemap from the final URLs only, and I cover the same clean-up for links in the guide to fixing redirect chains.

How to check canonicals and redirects on your site

On a small site the check takes you about 15 minutes with free tools.

  1. Paste the four versions of your homepage (http and https, with and without www) into the redirect checker. Three of them should end on the fourth in one step.
  2. Run the canonical checker on the homepage and one page of each template, such as an article, a product and a category. The canonical should be absolute and equal to the URL that answers 200.
  3. Compare the host and protocol in the canonical with where your redirects lead. If they differ even by www, fix the canonical template.
  4. Open your sitemap and paste a few of its URLs into the redirect checker. Each one should answer 200 straight away.
  5. In Search Console, open Indexing > Pages and look at "Page with redirect", "Alternate page with proper canonical tag" and "Duplicate, Google chose different canonical than user". The first two are usually fine. The third means Google disagreed with your tag.
  6. Inspect a few changed URLs with URL Inspection and compare "User-declared canonical" with "Google-selected canonical". Remember that only the indexed result shows Google's choice, and the live test doesn't.
  7. After you change a redirect, click Request indexing for the old URL in URL Inspection and inspect it again a few weeks later. It should move to "Page with redirect".

When you need the bigger picture, my earlier guide to canonical, sitemap and robots.txt signals shows how all indexing signals fit together.

How to explain canonical vs 301 to a client

I use one comparison. A 301 is a forwarding address at the post office, so letters stop reaching the old flat. A canonical is a note on the door of a second flat that says "the owner lives next door", and people can still walk in.

When they ask why they need both, I say the redirect cleans up duplicates you control. The canonical protects the ones you can't remove, like campaign links with parameters. Then I show them the four variants of their own homepage in the redirect checker, because one screen of results convinces faster than any explanation.

Check canonicals and redirects across the whole site

The free tools check one URL at a time, which is enough for templates and a homepage. To find problems deeper in your site, CometWeb Insight, our website audit app, checks the whole project in one audit:

  • it crawls your pages and shows the response code, redirects, indexability and metadata of each one;
  • it puts the fixes on one list in priority order, so a template error shows up above a single 404.

The free plan includes one project, one Standard audit a month and a crawl of up to 20 pages. You can create a free account and scan your own domain today.

Final thoughts on canonical vs 301

The easy part is choosing: redirect what should disappear and canonicalise what has to stay. The harder part is keeping redirects, canonical tags, the sitemap and internal links on the same address. Each one lives in a different place and changes on a different day.

Tomorrow morning, paste your four homepage variants into the redirect checker and read the canonical on the page they end on. If those two agree, most of your site probably does too.

Frequently asked questions

Can I use a canonical tag and a 301 redirect on the same URL?

A URL that answers with a server-side 301 never shows its HTML to visitors or Google, so a canonical on it has no effect. If you redirect with a meta refresh or JavaScript instead, make the canonical on that page name the address the redirect leads to. Put a self-referencing canonical on the page the redirect ends on, so both signals name the same address.

Does a 302 redirect work as a canonical signal?

Google says it doesn't use temporary redirects as a signal that the target should be canonical, and it may keep the old URL in results. Use 301 or 308 when the move is permanent. I explain the difference between permanent and temporary codes in the guide to redirect chains.

Should I use a cross-domain canonical or a 301 when I move content to another domain?

Use a 301 when the content leaves the old domain for good, because visitors and links should follow it. A cross-domain canonical fits syndication, where the copy stays online on another site and names your page as the original.

How long should I keep a 301 redirect?

Google's site move guide recommends keeping redirects for as long as possible, generally at least a year. I keep them indefinitely, because old links in emails and on other sites keep working for years.

Does a 301 redirect lose some link equity?

Older guides claimed a 301 loses some link value, but Gary Illyes of Google said plainly that 3xx redirects don't lose PageRank anymore. Real losses come from elsewhere, from a redirect to a page on a different topic or from long chains.

Sources

  1. Google Search Central: How to specify a canonical URL with rel="canonical" and other methods. The order of canonicalization signals, HTTPS preference and the rule against conflicting methods.
  2. Google Search Central: What is URL canonicalization. Why a stated canonical is a hint and what Google takes into account.
  3. Google Search Central: Redirects and Google Search. Permanent and temporary redirects as canonical signals, meta refresh and JavaScript redirects.
  4. Google Search Central: Site moves with URL changes. How long to keep redirects and updating canonicals, sitemaps and links.
  5. Search Console Help: Page indexing report. What "Page with redirect" and the duplicate statuses mean.
  6. Search Console Help: URL Inspection tool. User-declared and Google-selected canonical in indexed results, and Request indexing.
  7. RFC 6596: The Canonical Link Relation. The specification of rel="canonical".
  8. WordPress Developer Resources: redirect_canonical(). The built-in redirect to the preferred URL form.
  9. Tranco: a research-oriented top sites ranking. The list the sample was drawn from.
  10. Google Search Central Blog: 5 common mistakes with rel=canonical. Multiple tags on a page, the tag in <body>, relative URLs.
  11. Google Search Central: Troubleshoot canonicalization. Why Google picks a different canonical than the one you declared.
  12. Gary Illyes on X: 3xx redirects and PageRank. A Google employee saying redirects don't lose PageRank.
  13. CometWeb: canonical tags and redirects on 471 sites. Data, scripts and limits of the check.

Disclosure: CometWeb Insight and the free tools under /tools are CometWeb products, and CometWeb publishes this blog.

Check your site in Insight

Free gives you one project and 60 credits a month. That covers a Standard audit with a priority list.

No credit card Report with date and sources Free account in the app