Why is Search Console showing the wrong URL for Hreflang and Canonical?

Why does Google group distinct hreflang pages?

Google does not treat each hreflang page as a completely isolated entity. When content is deemed sufficiently similar among different language or geographical variants, the engine may decide to group them into a cluster and select a unique canonical URL.

This grouping does not mean that other versions are ignored. The engine retains knowledge of all declared hreflang variants. It simply uses a single URL as a reference point for calculating signals (popularity, authority, etc.).

What does the user actually see in search results?

Despite the technical grouping and the selection of a unique canonical, the URL served in the SERPs remains correct. A user logged in from Switzerland will see the Swiss URL, even if Google has chosen the German URL as the internal canonical.

This represents a decoupling between internal technical management (indexing, relevance calculation) and user display. The engine respects the hreflang declaration at the moment of serving the result. The issue is that this logic remains invisible in Search Console.

Why does Search Console show the German URL instead of the Swiss URL?

Search Console reports data on the canonical URL chosen by Google, not on the URL displayed in the SERP. If the engine has decided that the German URL is the canonical of the cluster, all impressions and clicks from the Swiss version are counted on that German URL.

For an SEO tracking performance country by country, this can be confusing. The numbers may seem “wrong” when they simply reflect Google’s internal mechanics. Let’s be honest: this is a limitation of the tool, not a bug in the algorithm.

• Google groups similar hreflang pages under a single canonical to simplify indexing
• User display remains correct: each country sees its localized version in the results
• Search Console reports metrics on the chosen canonical, not on the URL actually displayed
• This behavior is normal and documented: don’t panic if your GSC data points to another language
• The real issue is the lack of fine country/language segmentation in GSC reports

Is this statement consistent with real-world observations?

Yes, absolutely. This phenomenon is regularly observed on multilingual sites where the content is translated but structurally identical (same layout, same blocks, only the text changes). Google consolidates these variants under a single URL to avoid diluting signals.

The concern is that this consolidation is neither predictable nor well-documented at a granular level. It's impossible to know what similarity threshold triggers grouping. [To be verified]: there are no official metrics to anticipate this behavior.

What nuances should be added to this statement?

Mueller's statement concerns cases where everything works correctly. However, if your hreflang tags are misconfigured (loops, inconsistencies, lack of reciprocity), Google may choose an errant canonical or display the wrong version in SERPs.

Further, Mueller speaks of a scenario where “the user display is correct,” but does not specify how to verify this assertion without being physically present in each country. Geo-simulation tools (VPNs, manual geolocation settings) do not always reflect what Google actually serves. Testing with native IPs is necessary for validation.

In what cases does this rule not apply?

If your hreflang pages are sufficiently different in content, structure, or length, Google may decide to index them separately and not group them. The engine evaluates similarity, not just the presence of hreflang annotations.

Similarly, if you use conflicting self-referential canonicals (a Swiss page pointing to the German version in canonical while a hreflang declares it as a variant), Google may ignore the hreflang and respect only the canonical. And this is where it gets complicated: the user display becomes unpredictable.

What practical steps can be taken to avoid confusion?

First, accept that Search Console will never be the source of truth for your performance by language or country. The data is grouped according to Google’s internal logic, not according to your business segmentation. Use GSC to detect technical anomalies, not to finely manage your international KPIs.

Next, set up a regular geo-localized SERP monitoring. Test your strategic queries from each target country (using tools like BrightLocal, Semrush with geolocation, or native proxy servers) to verify that the right URL is displayed. This is the only way to validate that hreflang works on the user side.

What mistakes should be avoided in hreflang configuration?

Never declare hreflang without reciprocity. If your Swiss page points to the German version, the German version must point back to Switzerland. Loops or unilateral declarations undermine Google’s understanding and can trigger chaotic grouping.

Avoid conflicting canonicals as well. If your Swiss page declares a canonical to Germany AND a hreflang to Switzerland, Google may arbitrarily decide between them in unpredictable ways. Keep self-referential canonicals and let hreflang manage the relationship between variants.

How to verify that your international architecture is sound?

Regularly audit your hreflang tags with a dedicated tool (Screaming Frog, Sitebulb, OnCrawl). Look for reciprocity errors, loops, incorrect language codes. A single missing link can break the entire chain.

Then cross-reference the Search Console data (to identify the canonicals chosen by Google) with your server logs (to see which URL is actually crawled and served). If you find that one language version accounts for 100% of traffic while you have 5 markets, it’s a warning signal. Dig deeper into canonicalization and hreflang.

• Audit the complete reciprocity of your hreflang annotations across all your language variants
• Ensure that each page has a self-referential canonical and does not conflict with hreflang declarations
• Establish a monthly geo-localized SERP monitoring to validate that the correct URL is displayed in each country
• Cross-reference Search Console data with your analytics to identify traffic inconsistencies by language
• Manually test your strategic pages from native IPs or geo-localized SERP simulation tools
• Document the canonical choices made by Google in a tracking table to anticipate changes