Discussion starter: an international SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- Indexing
- International SEO
en-US, en-GB and en-AU pages with near-identical content: have you seen Google fold them into one canonical cluster despite hreflang? What fixed it?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
The short answer is yes, and it is expected behaviour rather than a bug. Google’s localised versions documentation says “Localized versions of a page are only considered duplicates if the main content of the page remains untranslated.” en-US, en-GB and en-AU pages with the same body copy are, by that definition, untranslated duplicates. hreflang does not stop Google clustering them. It tells Google which URL in the set belongs to which audience.
The important distinction is between clustering and serving. Google can pick one canonical for the cluster and still use hreflang to show Australian searchers the en-AU URL. So before fixing anything, check whether the problem is real (Australians landing on US pages, wrong currency, AU pages missing from results) or mostly a reporting artefact in Search Console. What fixes the real version is usually a combination of removing conflicting signals and making the main content genuinely different where it matters, not adding more hreflang.
How clustering and hreflang interact
Think of a publisher printing US, UK and Australian editions of the same book with identical text. A library catalogue may keep one master record for the title, because the content is the same. A good librarian still hands the Australian edition to the Australian reader, because the edition list says which copy suits which reader. The catalogue record is the canonical. The edition list is hreflang.
The technical sequence runs roughly like this:
Googlebot crawls each locale URL (mostly from US-based IP addresses)↓Canonicalisation: is the main content effectively the same?↓If yes, the URLs are clustered and one is selected as canonical↓hreflang maps alternates to language and region↓Serving: Google may show the locale-appropriate alternate from the clusterGoogle’s multi-regional sites guidance anticipates exactly this set-up. For similar or duplicate same-language content across regions, it says to “pick a preferred version and use the
rel="canonical"element andhreflangtags to make sure that the correct language or regional URL is served to searchers.” That wording only makes sense if a selected canonical and a served regional URL can differ.When clustering is a real problem and when it is cosmetic
What you see Likely meaning Priority en-AU URLs show “Duplicate, Google chose different canonical than user”, but Australian searchers still see en-AU URLs in results Clustering is working with hreflang swapping. Mostly a reporting issue, as performance data tends to aggregate to the canonical Low, unless snippets show the wrong price or currency Australian searchers see en-US URLs hreflang is not being applied (broken annotations, or signals point elsewhere) High AU results show the right URL but a US title, snippet or price Practitioners commonly report details coming from the canonical version. Not documented by Google, so verify on your own queries Medium to high on commercial pages en-AU pages are missing from AU results entirely Possible hreflang failure plus a canonical choice that sidelines the AU URLs High What I would check first
- Inspect a sample of URLs per template. In URL Inspection, compare the user-declared canonical with the Google-selected canonical for en-US, en-GB and en-AU versions of the same page. Do this across product, category, blog and home page templates, because the pattern often differs by template.
- Check what local searchers actually get. Search from the target country (or with a reliable location tool) for queries those pages rank for. If the en-AU URL appears for Australian users, the cluster is not hurting delivery.
- Validate hreflang mechanics. Google’s documentation is strict: “Each language version must list itself as well as all other language versions”, and “If two pages don’t both point to each other, the tags will be ignored.” Also check codes (
en-GB, noten-UK), and that every hreflang target returns 200, is indexable and is not redirected or canonicalised elsewhere. - Check canonical tags per locale. Each locale should normally carry a self-referencing canonical. A common cause of folding is templates that point en-AU and en-GB canonicals at the en-US URL, or at a URL without the locale path.
- Look at what Googlebot sees, not what you see. Fetch each locale with a crawler using a US IP and no cookies. If currency, prices, shipping or contact details switch by IP or cookie after load, Googlebot may be seeing three identical US pages.
- Compare internal and external links. If navigation, sitemaps, footer links and most backlinks point to the US version, it will usually be the strongest candidate in any cluster.
Watch for this:
Google’s multi-regional guidance says not to use IP analysis to adapt content and to avoid automatically redirecting users between versions, because this can stop users and search engines from seeing every version. Geo-redirects and IP-based currency switching are among the most common reasons locale pages look identical to Google.
Hypothetical example:
A retailer runs
/us/,/uk/and/au/with identical product copy. Prices are rendered by JavaScript from a geolocation lookup, so a US-based crawler sees USD on all three. hreflang is valid, but the AU and UK pages are clustered under the US URL, and Australian results sometimes show USD pricing. Moving currency, GST-inclusive price, shipping and returns information into the server-rendered HTML for each locale changes the inputs to canonicalisation. Valid hreflang alone could not have fixed it, because the pages Google fetched were genuinely the same.The fixes, in the order I would try them
Fix What it changes Effort When it is enough Remove conflicting signals: self-referencing canonicals, complete reciprocal hreflang with correct codes, x-default, locale sitemaps, locale-consistent internal linksStops you telling Google that the US page is preferred Low to medium When the cause was cross-locale canonicals or broken annotations Stop IP-based redirects and content switching; serve each locale at its own URL to every visitor, with a banner suggesting the local version Lets Googlebot see each version as it really is Medium When the crawler was seeing one version everywhere Localise the main content that differs in reality: currency, GST- or VAT-inclusive prices, shipping, returns, stock, payment options, local contact details, local reviews and examples, spelling Gives canonicalisation a real reason to keep the pages apart Medium to high On commercial pages where these differences exist anyway Accept clustering on pages that are truly identical (for example documentation), relying on hreflang for serving Saves effort where nothing is lost Low When local users already see local URLs Consolidate to one English version plus local versions only where commercial content differs Removes the duplicate problem at the source High (migration) When three versions add maintenance cost without local value There is no documented threshold for how different same-language pages must be before Google keeps them apart, and no controlled public test I can point to that establishes one. That is why I would start with signal hygiene and rendered-content checks, which fix a clear mechanism, before rewriting copy for the sake of difference. Converting US spellings to Australian or British ones is a small difference in main content, and I would not expect it to change a canonical decision on its own.
Read literally, Google’s “pick a preferred version” guidance means canonicalising all versions to one URL and relying on hreflang for serving. That can work, but it puts the choice of canonical in your hands at the expense of local indexing, and it depends on Google swapping in the alternate. If the locale pages differ commercially (price, currency, availability), I would keep self-referencing canonicals so each remains an indexable candidate.
Why local signals matter beyond canonicalisation
Local differentiation is not only about keeping pages out of a cluster. Aleyda Solis’s analysis of the May 2026 core update found UK results shifting from US.com retailers towards local-market domains. That is a ranking observation from one dataset, not proof about canonical selection, but it suggests market fit is worth building on regional pages anyway.
How to tell whether the fix worked
Canonical decisions are re-evaluated as pages are recrawled, so expect weeks rather than days on sites with slow recrawl. I would track:
- Google-selected canonical for the same URL sample in URL Inspection, per template
- Count of en-AU and en-GB URLs under “Duplicate, Google chose different canonical than user” in the Page indexing report, using folder-level properties for each locale
- Share of Australian clicks landing on en-AU URLs, and British clicks on en-GB URLs, in Performance reports filtered by country and page
- Correct currency and price in snippets and product results for local searches
- Conversion rate for local visitors, because landing on the wrong locale usually shows up as checkout friction before it shows up in rankings
My practical position
If Australian and British searchers already land on their own URLs, clustering is a reporting inconvenience and I would leave it. If they land on US pages or see US prices, fix the signals first (self-referencing canonicals, complete hreflang, no IP-based switching), then make the commercial content genuinely local in the server-rendered HTML. If you cannot justify local differences on a page type, stop maintaining three versions of it and serve one English page with
x-default.Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.