Hreflang Checker Example: 12 Tags Across 4 Languages

Two corrections before anything else, because both misconceptions cause real, wasted effort.

The underscore is a straightforward transcription error — hreflang uses a hyphen, always. The confusion is understandable, because og:locale uses an underscore (en_GB) while lang, hreflang and every other web standard uses a hyphen (en-GB). Same value, two conventions, one attribute away from each other.

But there is a worse locale mistake, and it is far more common. It is this:

There is no country code "UK". The ISO 3166-1 code for the United Kingdom is GB. And "UK" is not merely unrecognised — it is the code for Ukraine's language, Ukrainian, in the language standard. So en-uk is not a typo Google quietly forgives; it is an invalid combination, and Google will ignore the annotation entirely.

The rules, precisely: language is ISO 639-1 (two letters, mandatory); region is ISO 3166-1 Alpha-2 (two letters, optional); joined with a hyphen. Language alone is valid — de targets German speakers everywhere. Region alone is not valid. There is no such thing as hreflang="gb"; that reads as the language "gb", which does not exist.

The two missing return links are flagged as warnings. They deserve to be errors, because of how the rule actually works.

Hreflang is a claim that requires confirmation. Page A saying "the German version is over there" is an unverified assertion — anyone could point a tag at any page on the internet. Google only accepts the relationship if the German page agrees, by pointing back. Without the return link, the claim is unconfirmed and Google discards it.

The consequence is disproportionate. This is not two broken pages — an incomplete cluster can cause Google to disregard the annotations across the set. The nine correctly-configured tags in the table are doing nothing for the pages whose partners never replied. The set is only as good as its weakest member.

This one is worth more than everything else on this page, because it is invisible to most hreflang checkers — the tags all validate, the return links are all present, and the whole thing is being ignored.

The rule: every language variant must have a self-referencing canonical. The German page's canonical points at the German page.

What goes wrong: somebody, reasoning that the English page is the "original", sets the canonical on every translated page to point at the English one. This is a perfectly natural intuition and it is fatal. That canonical says "do not index me, index the English page instead." The hreflang tags simultaneously say "index me, and show me to German speakers."

Canonical and hreflang answer different questions. The canonical says "which URL is the definitive version of this page". Hreflang says "which other pages are alternates of this one". A translated page is not a duplicate of its source — it is a different page, and it canonicalises to itself. If your hreflang is not working and everything appears correct, check the canonicals first.

Pick one and use only that one. Implementing hreflang in both the head and the sitemap gives Google two sources to reconcile, and the moment they disagree — which they will, because one of them will be updated and the other forgotten — you have a conflict rather than a redundancy.

Two final requirements that quietly break sets: the URLs must be absolute, with protocol and domain — a relative path is ignored. And every URL in the set must return a 200. An hreflang pointing at a redirect, a 404 or a noindex page invalidates that entry, and an invalid entry can take the cluster with it. When a language is retired, remove it from every set. Leaving a dead tag behind is not a tidiness problem; it is a live fault.

What are hreflang tags?

Hreflang tags are HTML attributes that tell Google which language and region a page is intended for. They help Google show the correct language version to users in each country, and they help it understand that your variants are alternates of one another rather than unrelated pages.

What is a missing return link?

If page A has a hreflang tag pointing to page B, page B must also have a hreflang tag pointing back to page A. Hreflang is a claim requiring confirmation — anyone could point a tag at any page on the internet, so Google only accepts the relationship if the other page agrees. Without the reciprocal link the claim is unconfirmed and Google discards it, and an incomplete cluster can cause the annotations to be disregarded across the whole set.

What is x-default in hreflang?

x-default is a special hreflang value that tells Google which page to show when no other language variant matches the user's language. It is typically set on the homepage or a language selector page. Missing x-default is not an error but is recommended best practice.

Does hreflang improve my rankings?

No. Hreflang is a serving signal, not a ranking signal — it does not help a page rank higher, it changes which URL is displayed at the position the page has already earned. If your German page does not rank in Germany, hreflang will not fix that; you have a content or authority problem. What it does fix is the specific failure where you are ranking but Google shows the wrong language version to the wrong country.

Is hreflang="en-uk" correct for the United Kingdom?

No, and this is the most common locale error there is. There is no ISO country code "UK" — the code for the United Kingdom is GB, so the correct value is en-gb. Worse, "uk" is the language code for Ukrainian, so en-uk is not a typo Google quietly forgives but an invalid combination that causes the annotation to be ignored. The rules: language is ISO 639-1, region is ISO 3166-1 Alpha-2, joined with a hyphen. Language alone is valid; region alone is not.

My hreflang tags all validate but Google ignores them. Why?

Check your canonicals first. Every language variant must have a self-referencing canonical. If someone has set the canonical on each translated page to point at the English "original" — a natural intuition and a fatal one — then the canonical says "do not index me, index the English page" while the hreflang says "index me and show me to German speakers". Google resolves the contradiction by trusting the canonical, and the entire international setup evaporates while every tag validates perfectly. A translated page is not a duplicate of its source; it is a different page, and it canonicalises to itself.

Should I put hreflang in the head or the sitemap?

Use the sitemap for anything at scale, and use only one method. Head tags mean every page carries every tag — with six languages and ten thousand pages that is sixty thousand tags to keep in sync. The sitemap is a single file, generated programmatically, that cannot drift. Implementing both gives Google two sources to reconcile, and the moment one is updated and the other is forgotten you have a conflict rather than a redundancy. Also ensure every URL in the set is absolute and returns a 200: an hreflang pointing at a redirect, a 404 or a noindex page is an invalid entry, and an invalid entry can invalidate the cluster.