International Keyword Research: Beyond Translation
Literal keyword translation fails because people search with the words, phrases and intent of their own market, and these rarely map one-to-one onto another language. The reliable method is to treat translated terms as seeds, then validate and expand them with local data, local search results and a native speaker's judgment before they shape titles, headings and slugs.
Why does translating keywords literally fail?
Several things differ between markets even when the topic is identical:
- Vocabulary. Google's SEO Starter Guide notes that users with different expertise search with different words, using the example of one audience searching "charcuterie" while another searches "cheese board". The same effect appears across languages and regional variants. Illustrative examples (not from a study): British and American English use different words for some products, and Spanish used in Spain and in Latin America can differ in everyday vocabulary.
- Phrase structure. A natural English two-word query can be a long phrase, a compound word or a question in another language.
- Loanwords and brand terms. Some markets search for the English term even though a translation exists. Others never use the English term.
- Intent. The same topic can be researched, compared or bought in different ways depending on the market, so the page type that satisfies searchers may differ.
- Script and spelling. Languages with several scripts or spellings produce several valid query forms.
Google also says its language matching systems are sophisticated, so exact-match wording is not the goal. The goal is to speak the way your audience does and to serve what they want.
How do I research keywords for each language and market?
Work per market, not per language. "Spanish" is several markets. Define each one as a language plus a country, for example Spanish for Spain or Spanish for Mexico, matching how you plan to use language and region codes.
Step 1: Build a seed list
Start with the topics and page types you want to cover, not with English keywords. For each, collect:
- Your translated terms (from a translator or a draft machine translation).
- Terms from local competitor and marketplace pages that rank in the target country.
- Terms your local sales, support or partner teams already hear from customers.
Step 2: Look at what already exists in search
Search in the target language with the target country's settings, ideally from a browser or tool set to that locale. Note:
- Autocomplete suggestions and related searches.
- Which sites rank: local publishers, marketplaces, forums or international brands.
- Which result types appear: videos, shopping results, maps, answer boxes.
- The words used in titles and headings of top pages.
This is manual, but it shows intent faster than a spreadsheet does.
Step 3: Use data tools, with their limits in mind
| Tool | What it is good for | Limits |
|---|---|---|
| Search Console Performance report | Queries, pages, countries and devices you already get impressions for, with clicks, impressions, CTR and average position | Only shows terms you already appear for |
| Google Keyword Planner | Keyword ideas from words or a site, and estimated monthly searches | Built for advertising; check its language and location options for your market |
| Google Trends | Relative interest over time by region and comparing terms | Scaled 0 to 100, not absolute volume; low-volume terms may show as zero |
| Search results and suggestions | Intent, vocabulary and competition | Manual and personalized; use consistent locale settings |
| Third-party SEO suites | Larger keyword databases per country | Coverage and accuracy vary by language; verify before trusting |
Google describes Trends as a "largely unfiltered sample" of search requests, divided by total searches for a place and time and scaled from 0 to 100, so two regions with the same score can have very different volumes. Use it to compare terms within a market and spot seasonality, not to compare absolute demand between markets.
Step 4: Match search intent
For each candidate keyword, open the results and ask what the top pages are. If they are all product pages and your page is a long guide, you are mismatched. If they are comparison articles and you have a category page, same. Intent labels (informational, commercial, transactional, navigational) are a common practice, not a Google ranking system, but they help you decide which page should target a term.
Step 5: Have a native speaker review the shortlist
Give the reviewer the candidate keyword, the target page and a few top-ranking results. Ask whether the phrase sounds natural, whether it is the term people use, and whether there is a better one. This is the step that turns data into good copy.
Step 6: Map keywords to pages and write for people
Assign one primary keyword and a few related phrases per page, per language. Put the primary term naturally in the title, headings and opening text. Google's guide warns that excessively repeating the same words is tiring for users, and says that words in the URL path have hardly any effect on ranking beyond breadcrumbs. This is why a translated slug should be useful rather than keyword-stuffed.
How do I check what is working after launch?
Once pages are indexed, the Search Console Performance report filtered by page path (such as /de/) and by country shows which queries bring impressions. Look for:
- Queries you did not target, which reveal vocabulary you missed.
- Pages with many impressions but low click-through, which often means the title does not match how people phrase the query.
- Queries in one language landing on pages of another, which can indicate missing or incorrect hreflang annotations.
Update the page copy, then review again after enough time has passed for Google to recrawl. Search Console's default view shows the past three months, so schedule reviews with that window in mind.
Common mistakes in international keyword research
- Researching in English volumes and assuming they carry over.
- Treating a language as a single market when countries differ.
- Using machine-translated keywords without a native review.
- Targeting a keyword with a page type that does not match local intent.
- Ignoring non-Google search surfaces that matter in a given market. Which engines or platforms dominate varies by country, so verify the market rather than assuming.
- Keyword stuffing translated pages to "make up" for weak research.
Build a short per-language glossary of approved terms from this work, and reuse it for product pages, navigation and support content. It keeps wording consistent and makes later updates faster. The multilingual SEO checklist lists where those terms should appear on a page.
Key takeaways
- Treat translated keywords as hypotheses and validate them with local search results and data.
- Define markets as language plus country, since vocabulary and intent vary within a language.
- Combine Search Console, Keyword Planner, Trends and manual review, knowing the limits of each.
- Match the page type to the intent shown in the local results.
- Have a native speaker approve the final keyword list, then monitor queries by country after launch.
Sources
- SEO Starter Guide – user vocabulary, charcuterie versus cheese board example, keyword stuffing, URL words
- Performance report in Search Console – queries, pages, countries, devices and metrics
- About Google Trends data – relative scaled interest, not absolute volume
- About Keyword Planner – keyword ideas and search estimates
- Managing multi-regional and multilingual sites – separate URLs and language versions
Frequently asked questions
Can I just translate my English keywords?
Literal translation is a starting hypothesis, not a result. People in another market may use a different word for the same thing, a different phrase structure, or a different kind of query altogether. Use translated terms as seeds, then validate them against real search data and search results in the target language and country.
Which tools work for keyword research in other languages?
Common choices are Google Search Console's Performance report for terms you already appear for, Google Keyword Planner for ideas and estimates, Google Trends for relative interest, and the search results themselves (suggestions, related searches, competitor pages). Third-party SEO suites also support many locales. Check each tool's language and location coverage before relying on it.
Do I need a native speaker for keyword research?
It is strongly advisable. Tools can show you volumes and suggestions, but a fluent reader judges whether a phrase sounds natural, matches the search intent and fits the market's register. Even a short review of the shortlisted keywords by a native speaker catches problems that data alone misses.
Is search volume the same in every country for the same topic?
No, and you cannot assume it. Volumes, seasonality, competition and search features differ by language and country. Google Trends reports relative interest scaled 0 to 100 rather than absolute volume, so compare markets with care and avoid treating scores from different regions as equivalent counts.