Multilingual SEO: How to Rank and Get Cited in Every Language

Multilingual SEO is the work of making a website rank in search engines and appear in AI answers in more than one language. Each language version gets its own keyword research, its own content decisions and its own technical setup. The goal is simple. A buyer who searches in French, German, Italian or Spanish should land on a page written for them, in their language, about the problem they typed.

Note from Maeva: I started my career in translation and localization and built my first agency in Spain. Over 15+ years of running multilingual programs for B2B SaaS companies, I have seen one pattern repeat on every account. Translated pages rank poorly, and pages rebuilt by native speakers around local search behavior rank well.

The clearest example is a program Flying Cat ran for an ecommerce marketing platform. Its translated pages sat in the mid-20s on Google in France. After native writers rebuilt the French content around local search habits, the client tripled its top-3 keyword rankings in France within six months. Organic conversions in France grew 60% in year one.

This guide covers the process Flying Cat uses across English, French, German, Italian and Spanish. It starts with the pages you already have and ends with AI visibility in each market.

What multilingual SEO covers

Multilingual SEO is one part of international SEO. It focuses on language. Multi-regional SEO focuses on countries. The two overlap, but they solve different problems.

A US software company that adds UK and Australian pages is doing multi-regional SEO. Every page is in English, and the work is about local pricing, spelling, examples and country targeting. A company that adds French and German versions of its site is doing multilingual SEO, even if it sells into only one country per language.

Multi-regional SEO and multilingual SEO are both part of international SEO.

Many B2B SaaS programs end up doing both. A French version may need to serve France, Belgium, Switzerland and Quebec. A Spanish version may need separate pages for Spain and Mexico when search terms, regulations or pricing differ. For the broader planning questions, such as which markets to enter first, see Flying Cat's guide to building an international SEO strategy.

Start with an audit of the pages you already have

Every multilingual program Flying Cat runs starts with the English site. Your existing pages already carry proven topics, conversion data and internal links. The audit decides which of them can be localized and which need a new page written from scratch for each market.

Localization works well in two situations. The first is when search intent matches across markets, so the French results for a topic show the same kind of page as the English results. The second is when the topic is less market-specific, such as product features, integrations, templates and how-to guides for your software. Topics tied to local rules, local competitors or local buying habits usually need new pages. Data privacy, invoicing, payroll, tax and compliance topics fall into this group.

The team sorts each English page into one of four outcomes.

To decide, the team checks each target keyword in the local search results, from the local country, and compare the page types that rank. If the English page is a feature guide and the German results show feature guides, localization is a good starting point. If the German results show regulation explainers or local vendor lists, the English page will not compete.

ActiveCampaign is a good example of this split. Its French, Italian, German and Spanish sites had little content. Brand searches drove most of their traffic. Flying Cat localized the best English pages for each language with local examples, images and tone. The team also wrote about twenty new bottom- and mid-funnel pages per language each month for the keywords the English site did not cover. In eight months, the four sites gained 83,000+ more non-branded clicks and 4.1M+ more impressions.

"Our primary goal was to focus on content that met the needs of searchers in our primary markets. Within two quarters, we were able to drive an additional 80k visits to our website." (Joe Baugnet, Global SEO, ActiveCampaign)

Run keyword research in each market with native linguists

Translated keywords are a frequent reason multilingual pages fail on the accounts Flying Cat audits. A translation tool returns a correct word. A native speaker who knows the industry returns the word buyers type into Google. Those two are often different.

Flying Cat's veterinary software client shows how this plays out. In North America, "SOAP notes" is the standard term for clinical records. In the UK and much of Europe, vets more often say "medical records". The French market also responds differently to messaging. Burnout framing that works in North America does not land the same way with French vets. A translated keyword list would have missed both points.

Local search terms can also be more specific than the English ones. For the ecommerce marketing platform, one of the German pages that reached rank 1 targeted "dsgvo double opt in". That query combines the German name for GDPR with a consent process that German marketers search for by name. An English-first keyword list would never produce it.

Flying Cat's process for each market looks like this.

1. A native linguist with SEO training builds the keyword list in the local language, starting from the seed topics in the audit.
2. They check search volume and difficulty for the local country in a keyword tool, then read the local results page for each priority term.
3. They map each keyword to an existing page to localize or to a new page.
4. They flag terms where buyers use English loanwords, because many B2B software categories keep their English names in some markets.

Tools speed up the volume checks. The final keyword choice stays with the native linguist. Flying Cat's international keyword research guide walks through the tool side in more detail.

Build a glossary before anyone writes

A glossary keeps every writer, reviewer and translator on the same terms. Without one, the same feature gets three different names across ten French pages. Readers notice it. Search engines trying to understand what the page is about notice it too.

Each language glossary Flying Cat builds records these decisions.

• Product and feature names: whether each one stays in English or gets a local name.
• Category terms: the exact local phrase for what the product is, based on keyword research.
• Form of address: formal or informal, such as vous or tu in French, Sie or du in German, and usted or tú in Spanish.
• Regulatory terms: the local name for laws and standards, such as DSGVO for GDPR in German.
• Terms to avoid: literal translations that sound wrong, or words that belong to a competitor.

The glossary feeds the workflow. For the ecommerce marketing platform, outlines are written in English so the client's core team can review structure and positioning in one language. Native writers then write the final drafts directly in each local language. Once an outline is approved, its structure is locked, so the market drafts stay aligned on the topics and links the strategy depends on.

Write natively, then review in market with SEO context

Native writing is the step that moved the ecommerce marketing platform from page three to page one in France. The writers built every article around local search habits, retail calendars, regulations and customer examples. The budget covered about two new articles per market each month. Company policy ruled out link building. Legal also blocked comparison articles that named competitors. The results still came, because each page answered the local search better than the competing pages did. French organic impressions grew 506% in year one, and ranked keywords grew 6.8x. When the client launched its German blog from zero, German blog traffic grew 108% in the first three months.

Machine translation has a limited place in this process. Flying Cat uses it to sort keyword lists, to draft low-stakes pages that a native editor then rewrites, and to check consistency against the glossary. Publishing raw machine translation at scale carries a policy risk. Google's spam policies list automated transformations such as translating among the examples of scaled content abuse when the pages add little value for users. Flying Cat's SEO translation service exists for teams that want translated pages done by native linguists who also know SEO.

Local proof also outperforms generic content. For an enterprise learning platform, India-focused content grew clicks 76.5% in a single reporting period while its Latin America content declined. The pages that won were the regulation-specific localized pages. Generic translations underperformed.

Use in-market reviewers who understand SEO

Every draft for the ecommerce marketing platform gets a sign-off from a reviewer in each market, across the UK, France, Germany, Spain and Italy. In-market review catches the details a remote writer can miss, such as an outdated regulation, a local competitor's name or a term that changed meaning.

In-market reviewers need SEO context, or they can undo the work. A sales manager in Paris may rewrite a heading to sound better and remove the target keyword in the process. The Flying Cat team gives reviewers a short brief with the target keyword, the headings that cannot change and the internal links that must stay. Their job is to check accuracy and tone. The SEO structure is already settled.

Flying Cat's article on localization versus translation for SaaS goes deeper into where each approach fits.

Localize internal links and navigation

Internal links should stay inside one language. A French page should link to other French pages, with French anchor text. When a French reader clicks a link and lands on an English page, they often leave. The French page also loses the internal authority that should have gone to its French neighbors.

For multilingual SEO, link pages from the same language to each other.

This rule applies beyond the blog. Navigation menus, footers, CTAs, pricing pages and demo forms all need local versions. One of the challenges on the ActiveCampaign program was exactly this. Sending readers from local pages to English pages would have hurt the user experience and caused indexing problems.

Two more rules come straight from Google's guidance on multi-regional sites. Give each language version its own URL instead of switching languages with cookies or browser settings. Avoid redirecting visitors automatically based on their browser language or IP address. Googlebot usually crawls from the US, so automatic redirects can stop it from reaching your other language versions. A visible language switcher solves the problem for both people and crawlers.

Set up hreflang correctly

Hreflang tags tell search engines which language and region each version of a page targets. They also connect the versions to each other, so Google can show the French page to French searchers and the German page to German searchers.

The different components of a hreflang tag.

Google's hreflang documentation sets out the rules Flying Cat checks on every audit.

• Every version lists itself and every other version, so the full set of tags is identical on each page.
• Every link must be returned. If the English page points to the French page, the French page must point back, or Google may ignore the pair.
• Language codes follow ISO 639-1 and optional region codes follow ISO 3166-1 Alpha-2. A region code cannot stand alone, so "be" reads as a language code. Google ignores reserved codes such as "UK" and does not support codes like "es-419".
• An x-default tag points to the fallback page for visitors whose language you do not cover.
• One implementation method is enough. HTML tags, HTTP headers and XML sitemaps all work the same way.

Valid tags can still sit on top of broken pages. During a content audit for an AI tax research platform, Flying Cat found that its Canada pages drew almost no traffic. A JavaScript-rendered crawl showed 99.9% of the Canadian answer pages redirecting in the browser to their US versions. The pages returned a 200 status and carried hreflang tags, so a plain crawl would have missed the problem.

A re-platform caused a similar issue for Flying Cat's veterinary software client. After the site moved to a JavaScript app, the team checked every French page it had published. Every one served Googlebot the same empty shell with an English title. The French content was invisible to Google until the client's team fixed the rendering.

Both cases point to the same habit. Crawl with JavaScript rendering, check each language version as Googlebot sees it, and recheck after every migration. Flying Cat's complete guide to hreflang tags covers implementation and debugging step by step.

Choose subfolders, subdomains or ccTLDs case by case

The URL structure decides where each language version lives. All three common options work for multilingual SEO. The right choice depends on your domain, your CMS, your team and your legal setup in each country.

Subfolders keep every language on one domain, such as example.com/fr/. They share the main domain's authority, run on one CMS and need the least maintenance. They also make performance easy to report on, since everything sits in one Search Console property with folder filters.

Subdomains put each language on its own host, such as fr.example.com. They make it easier to run a different CMS, server or team for one market. Search engines tend to evaluate a subdomain more independently from the main domain, so each one may need to build its own authority.

Country-code top-level domains (ccTLDs) use a separate domain per country, such as example.de. They send the clearest country signal to users and search engines. Each domain builds authority on its own. Each one also needs its own purchase, hosting and maintenance. Some countries also restrict who can register their ccTLD.

Google lists a fourth option, URL parameters such as ?lang=fr, and recommends against it in its multi-regional guidance.

Flying Cat decides the structure with each client by weighing the authority of the main domain, whether a local entity or team needs control of its market's site, how well the CMS handles language folders and hreflang, and whether the markets differ by language only or also by country. A company with a strong .com and one central marketing team usually does best with subfolders. A company with established local domains and separate country teams may do better keeping them. Migrating a ranking ccTLD into a subfolder carries risk, so it needs a clear reason.

Track AI visibility in each market

Multilingual SEO now includes AI answers. Google expanded AI Mode to over 200 countries and territories and more than 35 new languages in October 2025. Buyers in Paris, Munich, Milan and Madrid now ask ChatGPT, Perplexity, Gemini and Google AI Mode the same vendor questions that US buyers ask.

AI assistants do not behave the same way in every language. Peec AI analyzed over 10 million ChatGPT prompts and found that 43% of the background searches for non-English prompts ran in English. About 78% of non-English sessions included at least one English sub-search. In practice, English pages and English vendor lists can shape the answer a German buyer reads, and local vendors can drop out of it.

The sources also shift by language. Profound studied 3.25 billion AI citations across 14 countries and found that query language changes which sources AI assistants cite more than geography does. ChatGPT's share of social citations fell from about 10% in English to 3 to 5% in every non-English market measured. YouTube made up 50.5% of social citations in Spanish AI Overviews, against 38% in English.

These findings change how Flying Cat measures AI visibility for multilingual clients.

1. The team builds a separate prompt set for each market, written by native speakers in the local language. The prompts reflect how local buyers ask, which differs from a translation of the English set.
2. The team runs each market's prompts from that market's country, so the answers reflect what a local buyer sees.
3. The team tracks presence, citations and share of voice per market against the local competitors, alongside the global ones.
4. The team finds the local third-party pages that AI assistants cite in each market, such as local review sites, industry publications and vendor roundups. Those become targets for placements and content updates.

Flying Cat's AEO guide to prompt research explains how the team builds prompt sets, and its GEO and AEO service covers the full program.

The results follow the same pattern as search. The ecommerce marketing platform ranks first in AI answer presence among tracked competitors in all five of its markets. Over three months to May 2026, its AI referral users grew 522% in France, 281% in Spain and 150% in the UK. In August 2026, its German ChatGPT sessions grew more than 35-fold and began producing signups. Flying Cat's veterinary software client ranks first in French AI share of voice, at about twice its nearest competitor. It also earned roughly three times as many citations as the tracked competitors combined.

Measure each market on its own terms

Each language version needs its own baseline and its own report. Combining markets hides the ones that are struggling behind the ones that are growing.

For multilingual clients, Flying Cat reports these metrics per market.

• Non-branded clicks and impressions from Search Console, filtered by language folder or domain
• Page-one keyword visibility against the main local competitors
• Conversions, meaning signups and demo requests, from organic and AI traffic
• AI presence, citations and share of voice from the local prompt set

For the ecommerce marketing platform, Flying Cat moved reporting in September 2025 from raw impressions to clicks, conversions and page-one visibility. Monthly reports compare month over month. Quarterly reviews compare year over year. In January 2026, clicks across its five EMEA markets grew 47.4% year over year. Ranked keywords grew 187.3% in the same comparison. By August 2026, "email marketing" in France had climbed from rank 48 to rank 1. The same month brought rank 1 positions in the UK, Spain, Germany and Italy.

Where to start with multilingual SEO

Start with one market where you already have demand, such as signups, sales conversations or branded searches from that country. Audit your English pages for that market, build the keyword list and glossary with a native linguist, and publish a focused cluster before expanding. The ecommerce marketing platform proved the model in France for nine months before it launched Germany, Italy and Spain. A small budget works when it finishes one cluster at a time.

Flying Cat runs multilingual programs in French, German, Italian and Spanish, alongside English. Maeva Cifuentes, Flying Cat's founder, leads the strategy on every account, with native freelance specialists who know each client's vertical. If you are comparing partners, our list of multilingual SEO agencies covers who writes the content in each language at each agency. If you want to see what a program would look like for your markets, the international SEO service page explains how the team works.