Canonical URLs, Redirects and Duplicate Content for Developers

A canonical URL is the address you want search engines to treat as the main version of a page when the same or very similar content is available at several addresses. Almost every website creates duplicates by accident: trailing slashes, upper-case letters, tracking parameters, "www" and non-"www" hosts, sorting options and printer views.

Google does not penalise this, but it has to choose one version to index, and it may not choose the one you want. Signals such as links can also be split across copies, and crawl time is wasted on addresses that add nothing. As a developer, you can remove most of this ambiguity with consistent URL generation, permanent redirects for true duplicates and rel=canonical for variants that must stay reachable.

This guide follows Google's documentation on consolidating duplicate URLs and adds the practical patterns we use in PHP and Laravel applications.

Where duplicate URLs come from

Duplicates usually come from the framework, the server configuration or marketing tools rather than from content editors. Listing the sources for your own site is the first step.

Run a crawler over a staging copy of the site with parameters enabled. The list of distinct URLs returning the same page is usually longer than expected.

Server logs are another good source. Search them for requests from Googlebot to URLs with parameters, mixed case or unexpected paths. They show which duplicates Google has already discovered, often through old links, marketing emails or third-party sites, and which ones are worth fixing first.

How Google chooses a canonical URL

Google's article on consolidating duplicate URLs explains that Google groups pages with the same or similar content and selects one as canonical. It uses several signals, and your preferences are among them, not the only factor.

• Redirects: a strong signal that the target should be canonical.
• rel=canonical link element or HTTP header: a strong signal.
• Sitemap inclusion: a weak signal that listed URLs are preferred.
• Other factors: HTTPS over HTTP, internal links pointing to a version and how the URLs look.

When signals conflict, for example a canonical tag pointing to one URL while every internal link points to another, Google may choose a different canonical from the one you declared. Consistency across all signals is the real goal.

It also helps to remember what canonicalisation does not do. It does not hide pages from users, and it does not guarantee that the canonical version will rank. It simply tells Google which address should represent a group of duplicates, so that ranking signals are concentrated on one URL instead of being spread across several.

Redirects for true duplicates

If a duplicate URL has no reason to exist for users, redirect it permanently with a 301 or 308. Google's documentation on redirects describes permanent server-side redirects as the strongest way to signal a change of address. Handle normalisation as early as possible, before routing.

// Normalise host, scheme, trailing slash and case before routing (PHP).
$host = 'example.com';
$uri = parse_url($_SERVER['REQUEST_URI'], PHP_URL_PATH) ?? '/';
$clean = strtolower(rtrim($uri, '/')) ?: '/';
$isHttps = ($_SERVER['HTTPS'] ?? '') === 'on';
if (!$isHttps || $_SERVER['HTTP_HOST'] !== $host || $clean !== $uri) {
    $query = $_SERVER['QUERY_STRING'] ?? '';
    header('Location: https://' . $host . $clean . ($query !== '' ? '?' . $query : ''), true, 301);
    exit;
}

Lower-casing every path is safe only if your routes are case-insensitive by design. If some identifiers are case-sensitive, normalise selectively. In Laravel, the same logic fits naturally into a global middleware that runs before the router.

Avoid chains

Each rule should redirect straight to the final URL. A request for http://www.example.com/Blog/ should reach https://example.com/blog in one hop, not three. Chains slow users down and can stop crawlers following them. Test the rules in combination, because each one may look correct on its own while together they create two or three hops.

rel=canonical for variants users need

Some duplicates must stay reachable: a sorted product list, a URL with campaign parameters or a filtered view shared by a customer. Here, the page should declare the preferred version with a canonical link in the head.

<link rel="canonical" href="https://example.com/services/magento-development">

• Use absolute URLs, including scheme and host, to avoid ambiguity.
• Point to an indexable page that returns 200, not to a redirect or a page marked noindex.
• Give each canonical page a self-referencing canonical, so parameters added later are handled automatically.
• Declare only one canonical per page; conflicting tags from a theme and a plugin are common.
• Keep the main content equivalent. Canonical is for duplicates, not for pointing unrelated pages at a popular one.

Generating canonicals in templates

Build the canonical from the route rather than the request, so stray parameters never leak into it. Allow-list the parameters that genuinely create distinct content, such as page numbers on later pages of a list.

// Laravel: canonical from the current route, keeping only allowed parameters.
$keep = array_intersect_key(request()->query(), array_flip(['page']));
if (($keep['page'] ?? '1') === '1') {
    unset($keep['page']);
}
$canonical = url()->current() . ($keep ? '?' . http_build_query($keep) : '');

Canonical headers for non-HTML files

PDFs and other files cannot carry a link element, but the same signal can be sent as an HTTP header. This is useful when a brochure exists both as a PDF and as an HTML page, or when the same file is served from several paths.

Link: <https://example.com/downloads/price-list.pdf>; rel="canonical"

Set it in the web server configuration or in the controller that streams the file. For files that should not appear in search at all, an X-Robots-Tag: noindex header is the equivalent of a noindex meta tag, but do not combine it with a canonical header on the same response.

Internal links, sitemaps and hreflang

Declaring a canonical URL while linking everywhere to a different version sends mixed signals. Make sure every internal link, menu item, breadcrumb and pagination link uses the canonical form. Sitemaps should list only canonical, indexable URLs that return 200.

On multilingual or multi-regional sites, each language version is a separate page with its own self-referencing canonical, connected to the others with hreflang annotations. Do not point every language version to the English page as canonical; that tells Google the translations are duplicates.

Our technical SEO checklist covers sitemaps and hreflang for online stores, and our guide to faceted navigation explains how filters create thousands of near-duplicate URLs in large catalogues.

Common canonical URL mistakes

• Every page canonicalised to the home page, usually from a template that hard-codes the site root.
• Canonicals pointing to redirected or 404 URLs after a migration.
• HTTP canonicals on an HTTPS site, left over from before the switch.
• Paginated pages all pointing to page one, which hides the products or articles on later pages.
• Using robots.txt to block duplicates, which stops Google seeing the canonical tag at all.
• Using noindex on duplicates while also declaring a canonical, which sends contradictory signals.

Google's documentation specifically advises against using robots.txt or the URL removal tool for canonicalisation, and against mixing noindex with canonical declarations.

Testing and monitoring

Automated tests catch most regressions before they reach production. Add a small suite that requests known duplicate forms and asserts the result.

1. Request the HTTP, "www", trailing-slash and upper-case variants of key pages and assert a single 301 to the canonical form.
2. Request pages with tracking parameters and assert a 200 with the clean canonical in the head.
3. Parse every page in the sitemap and assert that its canonical equals its own URL.
4. After deployment, crawl the site and compare the canonical of each page with its address.

In Search Console, the URL Inspection tool shows both the canonical you declared and the one Google selected. The page indexing report lists pages where "Google chose a different canonical than the user". Those entries deserve investigation. Our guide "Is my website on Google?" explains the other indexing statuses.

Canonicals during migrations

Platform changes and redesigns are where canonical problems cause the most damage. Old URLs must redirect to their new equivalents, new pages must declare the new canonical addresses, and the sitemap must list only new URLs. Leaving old canonical tags in templates, or redirecting everything to the home page, can undo years of search visibility. Our guide to migrating a store without losing SEO covers redirect mapping in detail.

Frequently asked questions

Need a technical SEO review of your URL structure? Our SEO services include canonical and redirect audits, and our developers can implement the fixes.