Skip to content
ROI Synth

Multilingual SEO and hreflang implementation

Multilingual SEO makes each language version of a site findable by the people who read that language, and hreflang is how that is implemented. Hreflang annotations tell search engines which language or regional version of a page to show to which users. Implementation requires every page in a set to list all alternates, including itself, with valid codes, canonical URLs and matching return links. Errors are common and easy to miss.

By ROI SynthUpdated 11 min read

The problem this solves

Hreflang can be implemented in HTML link elements, HTTP headers or XML sitemaps. Google documents all three. For large sites, sitemaps are often easier to maintain than page-level tags.

Typical errors include missing return links, alternates that point to redirected or non-canonical URLs, invalid region codes such as 'en-UK' instead of 'en-GB', and hreflang sets that conflict with canonical tags pointing to another language.

What the work includes

  • Validation of existing hreflang across all markets
  • Choice of implementation method for your platform
  • Specifications for generating annotations from your data
  • Ongoing validation checks

How the work is done

  1. Map equivalents

    Determine which pages are true equivalents across markets.

  2. Choose the method

    Select HTML, header or sitemap implementation.

  3. Implement from data

    Generate annotations from the same data that defines canonical URLs.

  4. Validate continuously

    Check return links, codes and target status codes regularly.

When the wrong version reaches the wrong audience

Close-up of a colorful scratch-off world map showing Europe with focus on planning travel

Hreflang annotations tell search engines that several pages are versions of the same content for different languages or regions, and which version suits which audience. Without them, a search engine may show your American page to British searchers, your Spanish page to Mexican searchers when you have a Mexican version, or treat near-identical English versions as duplicates and pick one arbitrarily.

Hreflang is simple in principle and fiddly in practice. Every page in a set must list every alternate, including itself; every listed URL must be the canonical, indexable version; every alternate must list the others back. One broken link in the chain can cause the whole set to be ignored. On a coded site, the reliable approach is to generate annotations from the content model, so they cannot fall out of step with the pages.

  • Search Console shows impressions from one country landing on another country's pages.
  • Your hreflang tags were added by hand and have not been checked since new pages launched.
  • Language versions exist, but search engines index only one of them.
  • Tags point to URLs that redirect, return errors or carry a different canonical.
  • You are launching new languages or country versions and want annotations right from day one.

The rules a valid hreflang set must follow

Google's documentation sets out the requirements clearly. Most hreflang failures come from breaking one of them, often without noticing because nothing visibly breaks on the page.

  • Return links

    If page A lists page B as an alternate, page B must list page A. One-way annotations are ignored.

  • Self-reference

    Each page lists itself among its alternates, with its own language or region code.

  • Valid codes

    Language codes in ISO 639-1 format, optionally followed by an ISO 3166-1 region code: en, en-GB, es-MX. "en-UK" is a common, invalid mistake.

  • Canonical, indexable targets

    Every URL in a set returns 200, is not noindexed and is its own canonical. Annotations pointing at redirects or non-canonical URLs undermine the set.

  • An x-default where useful

    A fallback version for users whose language or region has no specific version, often a language selector or the main international page.

Tags, headers or sitemaps

Explore the picturesque pedestrian street lined with historic buildings in Ljubljana, Slovenia

Hreflang can be delivered in three ways. Each is valid; the right choice depends on your site's size and how it is built. What matters most is using one method consistently rather than mixing them.

Ways to implement hreflang
MethodBest forWatch out for
Link tags in the page headMost coded sites with a moderate number of alternatesPage weight on sites with many languages
HTTP headersNon-HTML files such as PDFsHarder to inspect and maintain
XML sitemapsLarge sites with many alternates per pageSitemaps must be regenerated whenever pages change

On Next.js, the metadata API can generate alternate language links per route from the same data that knows which translations exist. That keeps annotations and pages in step automatically: when a translation is published, its annotations appear; when one is removed, they disappear.

How hreflang and canonicals must agree

Canonical tags and hreflang annotations work together, and conflicts between them are one of the most common causes of failure. Each language or regional version should have a self-referencing canonical. If the German page's canonical points at the English page, search engines are being told two contradictory things: that the German page is a duplicate of the English one, and that it is a distinct alternate for German speakers.

We check both together, for every page type and locale. On coded sites, the same helper that builds clean canonical URLs builds the hreflang alternates, which removes the chance of the two disagreeing.

A hypothetical hreflang repair

A hypothetical online course provider has English pages for the US and UK and Spanish pages for Spain and Mexico. Search Console shows UK searchers landing on US pages, and the Mexican pages barely appear at all.

The audit finds four problems. Annotations use "en-UK" instead of "en-GB". The Mexican pages carry canonicals pointing at the Spanish pages, so search engines treat them as duplicates. Several course pages exist in only some versions, but annotations list all four regardless, pointing some alternates at 404 pages. And there is no x-default. The fix generates annotations from the course data, listing only versions that actually exist, with correct codes, self-referencing canonicals and an x-default pointing at the course's international page. Within the following weeks, Search Console shows UK impressions shifting to UK pages and the Mexican pages being indexed in their own right.

How we test annotations

Because hreflang errors are invisible on the page, testing has to be deliberate. We crawl the site and extract every annotation, then check each set for return links, self-references, valid codes and indexable, canonical targets. Representative URLs are inspected in Search Console to confirm Google sees the annotations. On coded sites, where possible, we add a test to the build that fails if an annotation points at a page that does not exist.

  • All annotations extracted and grouped into sets by crawl
  • Every set checked for return links and self-reference
  • Every code validated against the language and region standards
  • Every target checked for status, indexability and canonical
  • Sitemaps, if used, checked against the live pages

How the work runs

Hreflang work is delivered within the SEO offer. Audit and implementation fit in the first month for most sites; confirmation in search takes the weeks that follow as pages are recrawled.

Typical sequence
WhenWhat happensWhat you see
Week 1Annotations extracted and validated; canonicals checked; Search Console reviewed by countryA list of broken sets and their causes
Week 2Generation logic designed from the content model; x-default chosenA short specification for the implementation
Weeks 3–4Implementation in templates or sitemaps; automated checks addedCode changes and a clean validation crawl
Following weeksCountry-level performance watched as pages are recrawledCorrect versions appearing in each market

How success is measured

The clearest measure is whether searchers in each market see the right version. Search Console's performance report filtered by country shows which pages receive impressions from where. When hreflang works, impressions from the UK land on UK pages, from Mexico on Mexican pages, and so on. We report those shifts per market, alongside the validation results showing that annotations remain error-free as the site changes.

What we need from you

We need to know which markets each version is meant for and what the fallback should be for everyone else. We also need repository access to implement generation where pages are rendered, Search Console access for every relevant property, and a list of planned translations so the content model can represent versions that do not exist yet without annotating them prematurely.

The hreflang errors we find most often

Hreflang has a reputation for being difficult mainly because small errors silently invalidate whole sets of annotations. These are the errors that appear again and again in audits.

Common hreflang errors and fixes
ErrorEffectFix
Invalid codes such as en-UK or es-LAAnnotations ignored for those versionsUse valid ISO codes such as en-GB, or a language code alone
Missing return linksPairs ignored because they do not confirm each otherGenerate annotations for all versions from one source
Pointing at redirectsSearch engines distrust the setAnnotate final, canonical URLs only
Annotating versions that do not existLinks to 404 pages break the setList only translations that are actually published
Canonical to another languageContradictory signals; versions treated as duplicatesSelf-referencing canonicals on every version
Mixing methods inconsistentlyConflicting annotations from tags and sitemapsChoose one method and use it everywhere

When not every page exists in every language

Very few multilingual sites translate every page. Product pages may exist in five languages while the blog exists in two and legal pages in one. Hreflang handles this well, as long as annotations reflect reality: each page lists only the versions that genuinely exist for that page.

On coded sites the content model is the natural place to record which translations exist. The template reads that record and generates annotations for those versions only. When a new translation is published, it is annotated automatically on every page in its set; when one is withdrawn, it disappears from them. Without that link between content and annotations, partial translation is exactly where hand-maintained hreflang breaks first.

The deliverables, listed

At the end of the work you have annotations that are correct now and that stay correct as the site changes.

  • An audit of existing annotations with every broken set and its cause.
  • Generation logic in your templates or sitemaps, driven by the content model.
  • Self-referencing canonicals and a deliberate x-default for every page type.
  • An automated check, where your workflow supports it, that fails on annotations to missing pages.
  • A short note for your team on adding new languages and regions correctly.

The language switcher and automatic redirects

Hreflang works alongside how visitors move between versions. A visible language or country switcher, built from ordinary links to the equivalent page in each version, helps people who arrive on the wrong one and gives crawlers another path between alternates. Linking each version to the matching page, rather than to the homepage of the other language, keeps people where they were.

Automatic redirects based on IP address or browser language are riskier. Crawlers often request pages from a single location and without the language preferences a visitor would send, so forced redirects can stop them from ever seeing some versions. A gentle suggestion — a banner offering the version that seems to fit — respects both the visitor's choice and the crawler's need to reach every page.

Keeping annotations correct as the site grows

Hreflang rarely breaks at launch; it breaks later. A new section launches in two languages instead of four. A page is renamed in one language and its slug changes. A translation is withdrawn because it was out of date. A redesign changes how the head of the page is generated. Each change can quietly invalidate a set of annotations, and nothing on the page looks wrong.

That is why we prefer generation from the content model over hand-maintained tags, and why we add checks where the workflow allows. A build-time test that compares every annotation with the list of published pages catches most problems before they reach production. Inside an ongoing SEO engagement, a monthly crawl validates every set and the report lists anything that has drifted, with the fix.

Search Console provides a second view: impressions by country and the pages shown in each. A sudden shift — UK searchers starting to land on US pages again — is often the first sign that something changed in the annotations, and it points straight to where to look.

Using sitemaps for large multilingual sites

When a page has many alternates — a product available in fifteen country versions, for example — adding every annotation to the head of every page adds weight to each response. Listing alternates in XML sitemaps instead keeps pages lean and puts all the annotations in one place that is easy to generate and validate.

The trade-off is that sitemaps must stay exactly in step with the live pages. On coded sites that is manageable when the sitemap route reads from the same content data as the templates: each URL entry lists its alternates, including itself, and the entries are regenerated on every build. We validate the generated sitemap against the live site after release, because a sitemap that lists pages which no longer exist undermines the annotations as surely as broken tags would.

Several versions in one language

Sites with several versions in the same language — English for the US, UK, Australia and Canada, or Spanish for Spain and Mexico — depend on hreflang more than most, because the pages are often very similar. Without annotations, search engines may treat them as duplicates and choose one version to show everywhere. With correct annotations and genuine local differences in prices, spelling, contact details and legal notes, each version can serve its own market. We make sure both parts are in place: the annotations, and enough local substance for each version to deserve its separate page.

How hreflang connects to other work

Hreflang implements the decisions made in an international SEO strategy: which languages and countries you target and how the site is structured. It depends on correct canonicals, covered by indexation management, and it must be rebuilt during any site migration that changes international URLs.

Which offer includes hreflang

Hreflang implementation is part of the SEO offer, from $200 per month, and of All-in-One SEO + AEO + GEO, from $300 per month. The number of languages and regions affects the scope. A new multilingual site from the Next.js + TypeScript Website Build, from $1,000 one time, can generate annotations from its content model from launch.

Questions about hreflang

The questions we hear most.

  • "Does hreflang improve rankings?"

    It helps search engines show the right version to the right people. It is not a ranking boost on its own.

  • "Do we need it with only two languages?"

    Yes, if both versions target audiences that could see either. It is simple to do right for small sets.

  • "Do other search engines use hreflang?"

    Search engines document their own approaches to language targeting and they differ. We check current documentation for the engines that matter in your markets and implement what each supports.

  • "Can a plugin handle it?"

    Sometimes. On coded sites, generating annotations from your content model is usually more reliable.

Example deliverable

How progress is evaluated

  • Hreflang errors in validation crawls
  • Correct regional URLs shown in each market
  • Traffic to wrong-market pages

Who this is for

  • Multi-language and multi-region sites
  • Sites showing the wrong country version in results
  • Sites adding new markets

Limits and honest caveats

Common questions

What is multilingual SEO and hreflang implementation?

Multilingual SEO makes each language version of a site findable by the people who read that language, and hreflang is how that is implemented. Hreflang annotations tell search engines which language or regional version of a page to show to which users.

What does a multilingual SEO and hreflang implementation engagement include?

The work includes validation of existing hreflang across all markets, choice of implementation method for your platform, specifications for generating annotations from your data and ongoing validation checks.

What problems does multilingual SEO and hreflang implementation solve?

Typical errors include missing return links, alternates that point to redirected or non-canonical URLs, invalid region codes such as 'en-UK' instead of 'en-GB', and hreflang sets that conflict with canonical tags pointing to another language.

How is multilingual SEO and hreflang implementation carried out?

It runs in 4 stages: map equivalents, choose the method, implement from data and validate continuously. Determine which pages are true equivalents across markets.

Who is multilingual SEO and hreflang implementation a good fit for?

It suits multi-language and multi-region sites, sites showing the wrong country version in results and sites adding new markets.

How do you measure the results of multilingual SEO and hreflang implementation?

Progress is measured with hreflang errors in validation crawls, correct regional URLs shown in each market and traffic to wrong-market pages.

What are the limits of multilingual SEO and hreflang implementation?

Hreflang is a signal for choosing between versions, not a ranking boost. Some search engines use other signals, such as content-language, instead.

Can ROI Synth carry out the multilingual SEO and hreflang implementation work, or only recommend it?

Implementation is included. Within the agreed scope we fix templates, metadata, internal links and redirects, and build the landing pages ourselves, on your existing platform or on a Next.js and TypeScript build when a rebuild is part of the project, starting with steps such as “Validation of existing hreflang across all markets”. Every completed item is recorded with evidence in a weekly progress sheet.

How long does multilingual SEO and hreflang implementation take?

The 4 stages (map equivalents, choose the method, implement from data and validate continuously) fit the proposal's four weekly phases. When the agreed scope can be delivered in a month, the proposal commits to completing it in the first month, subject to timely access and approvals. Search engines then take weeks or months to re-evaluate the changes.

Is multilingual SEO and hreflang implementation guaranteed to bring more traffic or leads?

No. Rankings, traffic, leads and sales depend on search engines, competition and your market, so they are not guaranteed. What is committed is the agreed work, completed and evidenced, and measurement of hreflang errors in validation crawls and correct regional URLs shown in each market.

Sources and further reading

Contact us about Multilingual SEO and hreflang implementation

Tell us about your website and goals. We will say plainly whether this is the right offer, what it would include, and what it can and cannot achieve.

Prefer the full contact page?

Fields marked with an asterisk are required.

Your website's situation, the services and markets you want to grow, and any deadlines. Up to 5,000 characters.

Your details are used only to respond to this general enquiry. See the privacy policy.

Discuss multilingual SEO and hreflang implementation for your business

Tell us about your website and goals. We will confirm whether this work is the right priority, what it would include and how it would be delivered.