Multilingual keyword research means finding the terms people actually search in each language and country, starting from local seed terms and local SERPs, rather than translating your English keyword list. Translated keywords often have little or no volume, miss the loanwords and slang locals use, and point to a different search intent. The workflow is: brief a native researcher, gather local seeds, expand them in tools set to the right country and language, check the live SERPs, validate intent, then map keywords to pages.
Below is that workflow step by step, with the tools to use and the mistakes that most often waste translation budgets.
Why translating your keyword list fails
A translated keyword is a guess about how someone else talks. Typical problems:
- Loanwords. Many Europeans search English terms for digital and adult-industry topics ("cam", "dating app", "VPN", "affiliate"). A correct translation may get a fraction of the searches.
- Different phrasing. A literal translation of "best cam sites" may be grammatical, while locals search a different construction entirely.
- Inflection and compounds. Polish, Czech, Finnish and Hungarian express one idea in many word forms; German joins words into compounds. The form people type is not always the dictionary form.
- Intent shifts. The same phrase can be informational in one market and commercial in another, which changes the content that ranks.
- Variant differences. Dutch in Belgium, German in Switzerland and Portuguese in Portugal differ from their larger neighbours in vocabulary that affects search.
Machine translation of content has the same weakness, which is why our comparison of machine vs human translation for SEO puts native keyword briefs at the centre of any workflow.
The international keyword research workflow
Step 1: brief a native researcher
Ideally the person doing the research is a native speaker of the target language living in, or closely familiar with, the target country. Give them your English topic map (not your keyword list), your niche, the offers you promote, and the page types you plan to create. Ask them to tell you how they would search for each topic before they open any tool.
Step 2: gather local seed terms
Collect seeds from local sources, not from your English list:
- Search autocomplete in the local language, with the browser and search settings set to that country.
- Related searches and "People also ask" boxes on local SERPs.
- Navigation, category names and headings used by leading local sites in your niche.
- Local forums, review sites and social platforms where your audience talks.
- Merchant landing pages for that country, which show the terms advertisers use.
Step 3: expand seeds in keyword tools
Run your seeds through keyword tools set to the target country and language. Most major tools let you choose a country database; Google Keyword Planner lets you set both location and language. Check:
- Search volume and trend for each term.
- Variations, including English loanwords and mixed-language queries.
- Question keywords for informational content.
- Keyword difficulty or competition, read as a relative signal rather than an absolute truth.
Step 4: check the local SERPs
Volume tells you nothing about what ranks. For each priority keyword, look at the top results in that country:
- What type of page ranks: reviews, lists, merchant pages, news, forums?
- Who ranks: local publishers, international brands, translated pages?
- Which SERP features appear: video, local results, shopping, AI-generated answers?
- Are results filtered? For adult queries, SafeSearch settings and local rules can change what appears.
Step 5: validate intent
Classify each keyword as informational, commercial, transactional or navigational, based on what ranks rather than on the words. Group keywords that share intent into one page target.
Step 6: map keywords to pages
Build a keyword map per language with one primary keyword per page, supporting terms, and the equivalent English page (if one exists) so you can connect the versions with hreflang. Some local keywords will have no English equivalent; they may justify pages that only exist in that language.
Tools for native keyword research
| Tool type | Examples | Use it for |
|---|---|---|
| Search engine data | Google Keyword Planner, Google Search Console, Bing Webmaster Tools | Volumes by location and language; your actual queries by country |
| Commercial SEO suites | Ahrefs, Semrush, Sistrix and similar | Country databases, competitor keywords, difficulty |
| Trend and demand tools | Google Trends | Comparing terms and seasonality per country |
| SERP checking | Rank trackers or SERP tools with country settings, local VPN | Seeing what really ranks in each market |
| Question and autocomplete tools | Autocomplete scrapers, question tools | Long-tail and informational queries |
Coverage varies by market. Large-language databases are usually rich; smaller languages can have thin data, so lean more on Search Console data, autocomplete and SERP checks there. Also remember that Google is not the only engine everywhere: Seznam remains relevant in Czechia, and Bing's share varies by country. See European search engine market share before deciding which engines to research for.
Use your own data first: if your English site already receives visits from other countries, Search Console shows the exact queries people used, including in other languages. That is the cheapest seed list you will ever get.
Finding local search terms: worked example
The following is a hypothetical illustration of the process for a publisher reviewing dating sites who wants to launch in Germany.
- The English page targets "best dating sites". A literal German translation is the obvious first idea.
- The native researcher's autocomplete checks and local SERPs show several alternative phrasings, including compounds and English-German hybrids, and a strong cluster of searches comparing named sites.
- Keyword tool data for Germany shows that one of the alternatives has noticeably higher volume than the literal translation, and that a set of "[site] Erfahrungen" (experiences) queries is large.
- The SERP for the main term is dominated by comparison sites and test-style reviews, which tells the publisher what page format to produce.
- The keyword map ends up with one German comparison page, a set of German review pages targeting "Erfahrungen" queries, and no direct equivalent of two English pages that had no German demand.
The outcome is not a translated site but a German site built on German demand. For a deeper look at this market, see our guide to German keyword research.
Keyword research in other languages: notes by market
- German: check compound and split forms separately ("Dating Seiten" vs "Datingseiten"), and research Austria and Switzerland in their own country databases if you target them.
- French: watch for differences between France, Belgium and Switzerland, and for English terms that French searchers use despite official French equivalents.
- Dutch: research the Netherlands and Flanders separately; vocabulary and the sites that rank often differ.
- Polish, Czech and Hungarian: collect the inflected forms people actually type and check whether tools group them; diacritic and non-diacritic spellings may both carry volume.
- Nordic languages: English queries can be significant alongside Swedish, Danish, Norwegian and Finnish ones; compare both before deciding the language of a page.
- Spanish and Portuguese: always filter to Spain or Portugal, because Latin American and Brazilian data dominate global figures.
- Greek: check Latin-alphabet "Greeklish" spellings as well as Greek script.
Write these quirks into each language's keyword brief so writers and editors handle them the same way across the site.
Prioritising languages and keywords
Research costs time in every language, so prioritise:
- Languages: start where demand, payouts and your production capacity overlap; our guide to which European language to add first covers the decision.
- Keywords: within each language, research your money-page topics first, then supporting informational content.
- Depth: do thorough SERP checks on high-value terms and lighter checks on the long tail.
Common multilingual keyword research mistakes
- Translating the English list and calling it research.
- Using the wrong country database, for example researching Portuguese terms in a Brazilian dataset for a Portugal site.
- Ignoring English loanwords that locals actually search.
- Trusting volume without checking SERPs, leading to pages that do not match intent.
- Researching once. Search behaviour changes, especially in fast-moving niches; refresh key terms at least yearly.
- Briefing writers without the keyword map, so the final content drifts back to translated phrasing.
Frequently asked questions
Can I do keyword research in a language I do not speak?
You can gather data, but you cannot judge intent, nuance or SERP quality reliably. Work with a native speaker for at least the validation and mapping steps.
Are keyword tool volumes accurate for smaller European languages?
They are estimates everywhere and less reliable for smaller markets. Use them to compare terms, and confirm with Search Console data once your pages are live.
How many keywords should I research per language?
Enough to map every page you plan to publish in the next six months, with one primary term and a handful of supporting terms each. Breadth beyond that is rarely used.
Should keyword research come before or after translation?
Before. Research shapes which pages you create and what they target, and writers or post-editors need the keyword map before they start.
Next steps
Pick one target language, brief a native researcher with your topic map, and run the six steps for your top ten money pages. For the fundamentals that apply in every language, see our affiliate keyword research guide, and compare keyword tools and research services in our webmaster software and services directory.