International Keyword Research: How to Do It
Translating your keyword list is not international keyword research. Here is how to build one per market, including markets with thin volume data.
International keyword research means researching what people in a specific country and language actually search for, in their own words, rather than translating a keyword list built for your home market. The two produce different lists, and the gap between them is usually where the easiest ranking opportunities sit.
Building the list is only useful once your site can serve the right version of a page to the right searcher, which is what multilingual SEO as a whole is about; this post is specifically the research step that comes before the content gets written.
On this page
How is international keyword research actually different from translating your list?
Translation preserves meaning; search behavior does not follow meaning, it follows habit. The English phrase a US buyer types is rarely the literal translation of what a French or Japanese buyer types for the same intent — different products get compared by different names, different question phrasings dominate, and entirely different problems get searched first depending on how that market typically discovers a category.
A common real example: a software company's English list is full of "software," but the equivalent German searches skew toward "Programm" or a specific brand-name-as-category term, and a straight translation of "software" ranks it against the wrong competitive set entirely. You only catch this by researching the market directly, not by running your existing list through a translator.
The same gap shows up in question phrasing, not just single terms. English searchers tend to front-load a question with "how" or "what"; other languages often structure the equivalent question differently, or skip the question form entirely in favor of a short, noun-first phrase. A translated FAQ heading can read as grammatically correct and still not match how anyone in that market actually types the question into a search box, which matters because it is exactly the kind of heading an answer engine or a featured snippet looks to match against.
Category naming causes a third, quieter version of the same problem. A product category that has one settled name in your home market often has two or three competing names in another, with search volume split unevenly across them — and no keyword tool will surface that split for you if your starting list only contains the translated version of the name you already use.
Which tools actually give you reliable data by country and language?
None of them evenly. Every major keyword tool draws its volume estimates from search engine data that is denser in some markets than others, and a tool that is excellent for US English data can be thin or unreliable for a smaller market's local language.
| Source | Best for | Weakness |
|---|---|---|
| Google Keyword Planner | Directional volume across most markets Google Ads supports | Broad ranges rather than precise numbers on smaller accounts |
| Google Trends | Relative interest over time and across regions | No absolute volume; needs another source to size the opportunity |
| Semrush / Ahrefs style tools | Deeper keyword and competitor data in well-indexed languages | Coverage and accuracy drop noticeably in smaller or less-indexed languages |
| Native-market interviews or freelancers | Ground truth on phrasing and intent in any market | Slow and not scalable past a handful of terms without a real budget |
Use at least two of these together rather than trusting one. Where Keyword Planner and Trends broadly agree, treat the signal as solid; where they diverge sharply, that disagreement is itself useful information about how thin the data is for that market.
Budget for the gap in tool coverage as a real line item, not an afterthought. A market where every major tool has thin data is not a market where keyword research is optional — it is a market where more of the research has to come from people rather than software, which costs time and sometimes a freelancer's fee rather than a subscription. Treating both as legitimate research spend, rather than assuming a tool subscription alone will cover every market equally, is the difference between a keyword list that quietly weakens in your smaller markets and one that holds up everywhere you actually operate.
How do you build a keyword list for a new market from scratch?
Start from intent, not from your existing terms. Write down the two or three problems your product actually solves, then research how someone in that specific country would describe those problems in their own words — not how your home-market list describes them translated.
List the core problems your product solves, in plain language
Strip out your own brand vocabulary first. Start from the buyer's problem, not your product's feature names.
Pull autocomplete and 'people also ask' for those problems, in-market
Search from a location and language setting matching the target market, not your home base, since results and suggestions both localize.
Check two or three real local competitors' pages and headings
Their on-page language reflects what has already been tested against real local searchers, which is faster than guessing from scratch.
Cross-check the resulting terms against Keyword Planner or Trends
Confirm the phrases you found actually carry search volume before building content around them.
Do the in-market searching from a real changed location and language setting — a VPN, not a visit to your own staging site through an automatic geo IP redirect, which shows you your own assumptions about that market rather than what a real local searcher sees.
The fastest way to catch a bad translation-based keyword list is to read a competitor's actual page in that market out loud. If the terms on our list sound stiff next to it, the list came from a translator, not from research.
What do you do when a language has no reliable volume data?
Treat directional signals as your evidence instead of waiting for a number that may never arrive. Autocomplete suggestions, "related searches," the headings competitors already rank with, and the language used in local forums, marketplaces and reviews all tell you what real people are typing even when no tool will give you a volume figure to attach to it.
Once you have a shortlist built this way, a small paid search test is the fastest way to get real signal: run a handful of the strongest candidate terms through paid ads for a few weeks and watch impressions and click-through rate. That is faster and cheaper than commissioning organic content against a guess, and it works in markets no keyword tool covers well at all.
How do you map keywords to pages without creating cannibalization?
Once the list exists, assign each priority term to exactly one page before any content gets written, the same discipline you would apply in your home market. The risk is higher internationally because it is tempting to let one page loosely target several related local terms "since the market is smaller anyway" — which is exactly how two pages end up competing against each other for the same query six months later.
Build the map at the same time you decide country versus a single global site structure, because the two decisions inform each other: a market getting its own dedicated pages needs its own full keyword map, while a market sharing pages with a neighboring country needs its terms merged deliberately rather than split arbitrarily. Once the map and the structure agree, hreflang is what actually routes each searcher to the page you mapped their term to.
How often should this research get refreshed?
At least twice a year for any market you actively invest in, and quarterly for a market still young enough that search behavior is visibly shifting — new competitors entering, new product terminology becoming standard, or search habits maturing as more of the category moves online. A market you already rank well in and are not actively expanding can go longer between refreshes; one you are actively building content for cannot.
Treat a sudden shift in your own traffic-by-market data as a trigger to refresh early rather than waiting for the scheduled date — a real behavior change in a market you have not re-researched in months is easy to misread as a technical problem when it is actually a keyword list that has gone stale.
Keep a simple log of what changed at each refresh — new terms added, terms that dropped in volume, category names that shifted — rather than just overwriting the previous list. Six months from now, that log is what tells you whether a market's search behavior is genuinely evolving or whether last quarter's research was simply incomplete, and the two call for different fixes.
Questions people also ask
Can I just translate my English keyword list into other languages?
No. Translation gives you grammatically correct phrases; it does not tell you what people in that market actually type. The two lists overlap less than most teams expect, especially once you get past the most obvious, generic terms.
What do I do if a language has no search volume data at all?
Treat directional signals — autocomplete, related searches, competitor content, forum and marketplace language — as your primary evidence, and validate with a small paid test once real content exists, rather than waiting for volume data that may never arrive.
Should each country get its own keyword list, even if they share a language?
Yes, at least for your priority terms. Spelling, product names, units and search habits all vary between countries that share a language, and a shared list misses those differences even when the grammar is identical.
How is this different from just doing SEO in another language?
Doing SEO 'in another language' usually still starts from your home-market keyword list and translates it. International keyword research starts from the target market's actual search behavior and works backward to your content, which is a different and slower process.
How often should international keyword research be updated?
Revisit priority markets at least twice a year, and any market you are actively investing content into every quarter. Search behavior shifts faster in newer or smaller markets than in mature ones, so treat the refresh cadence as a floor, not a ceiling.
Sources
- Keyword Planner — Google Ads (checked 23 September 2026)
- Google Trends — Google (checked 23 September 2026)
- International SEO: how to reach a global audience — Semrush (checked 23 September 2026)
- International SEO: How to Optimize Your Site for Global Search — Ahrefs (checked 23 September 2026)
Get keyword research done market by market, not by translation
Book a free 30-minute consultation with Ibrahem and Jeremy. We look at your site while we talk, and you leave with a prioritised plan.
Who wrote this
Ibrahem HelalCo-founder
Leads strategy and delivery.
Works on: SEO, GEO
Reviewed before publishing by Jeremy Zava, Co-founder.
- Multilingual SEO: a complete guide for 2026How multilingual SEO works: URL structure, hreflang, translation, keyword research and redirects, with what each decision costs and how to check it is working.
- Single global site vs. country sitesOne global site works when pricing, legal terms and fulfillment don't change by market. Country sites earn their cost only once those genuinely differ.
- How to Implement Hreflang Tags CorrectlyHreflang tells search engines which language or country version of a page to show; how to add it, in the head or a sitemap, with a worked example.
- Geo IP Redirects: The SEO Risks and a Safer FixAutomatic geo IP redirects are a common, silent way to block search engines from crawling your own country pages. Here is what to do instead.