International SEO Keyword Research

TL;DR — You can’t just translate your English keyword list and call it international keyword research. People in different countries search for the same thing using different words, the search volume and difficulty are different in every market, and the results Google shows can mean something different locally. The job is to research each market on its own terms.

You can’t just translate your keywords

The most common mistake I see is the easiest one to make: take your existing keyword list, run it through Google Translate, hand it to the translators, done. It feels efficient. It also routinely produces keywords that nobody actually searches for.

Here’s the thing — search is about how real people phrase things, not what a dictionary says is correct. A translation can be perfectly accurate and still be the wrong word, because it’s not the word locals reach for when they open a search box. The classic example: a hotel chain translated its budget-hotel keywords into French (“hôtel budget,” “hôtel bas coût”), and those terms had essentially zero search volume — even though the English originals pulled thousands of searches a month. The words were “correct.” Nobody searched them.

Three reasons international keyword research is different

1. Vocabulary diverges by market — even in the same language. This catches people off guard. English-speaking countries don’t search the same way:

• “vacation” (US) vs. “holiday” (UK)
• “petrol” (UK/AU) vs. “gasoline”/“gas” (US)
• “maths” (UK/AU) vs. “math” (US)
• “sneakers” (US) vs. “trainers” (UK)
• “color” (US) vs. “colour” (UK/AU)

Same language, different words. If your page is optimized for the US term, you can quietly miss most of the UK or Australian demand for the same product.

2. Search volume and difficulty are market-specific. A keyword with 50,000 monthly searches in the US might have almost none in Brazil — and a keyword that’s brutally competitive at home can be wide open somewhere else. You have to look at the numbers per country, not assume your home-market data carries over.

3. SERP intent shifts by locale. Even when a word translates cleanly, the results Google shows for it — and what searchers expect — can differ. “Football” in the US and the UK are two different sports and two different result sets. The only way to know is to look at the actual results in that market.

The high-level process

1. Pick your markets. You don’t research everywhere at once. Start with the countries or languages with the most potential and the least competition.
2. Find local terms. For each market, figure out how people actually phrase what you offer — not the translation, the real search term.
3. Check the local results. Search your target terms the way someone in that country would and see what comes up.
4. See who you’re really competing with. Your global competitors usually aren’t the ones ranking in a given local market.
5. Map it out. Decide which keyword goes on which page, market by market.

Want the full workflow — including exactly which Ahrefs reports to use and how keyword difficulty changes across markets? Switch to the Advanced tab.

International keyword research mistakes

Translating the home-market keyword list

Why it fails: literal equivalents miss local vocabulary, local demand, and the intent visible in that market’s results. Do instead: start with local seed terms, local competitors, and a country-specific SERP review.

Reusing one market’s volume and difficulty

Why it fails: both metrics change by country, so the apparent priority may not travel. Do instead: pull each metric with the target-country selector and treat thin data as directional rather than exact.

Mapping a keyword before checking the local SERP

Why it fails: the same phrase can trigger a different task or content format in another market. Do instead: inspect the live results and map the term only when your page matches the local intent.

Worked examples

Same product, different vocabulary

A US seed list built around “vacation” should not be mechanically reused for the UK. Research “holiday” and its related terms in the UK database, then compare the local results before deciding whether the same landing page concept fits.

A market-by-market mapping row

The mapping preserves the business concept while allowing the query and page format to change by market.

Test yourself: International keyword research

TL;DR — International keyword research is researching each market directly instead of translating your home list. The process: prioritize markets → identify local terminology → validate intent in the live local SERP → find the local competitors → map keywords per market, watching out for thin volume data in small markets. The reason translation fails is that search reflects culture, not dictionaries — and the reason this is worth the effort is that volume and KD are market-specific, so a keyword locked up at home can be wide open abroad. In Ahrefs the spine of this is the Keywords Explorer country selector, Site Explorer organic keywords filtered by country, Content Gap against local competitors, Organic Competitors per country, and Metrics by Country.

Why this isn’t the same job as domestic keyword research

Standard keyword research assumes one market. Cross a border and three things that you normally treat as fixed start moving:

• The words move. Even within one language, terminology diverges by country — “vacation” vs. “holiday,” “sneakers” vs. “trainers,” “petrol” vs. “gas.” Across languages it’s worse than divergence: a literal translation is frequently not the search term. Germans say “beamer” for a projector. A criminal-records query in France is “casier judiciaire,” not a word-for-word render of “background check.” Local market leaders also train vocabulary — Italians search “finanziamento auto” rather than “prestito auto” partly because that’s how the big lenders marketed it.
• The numbers move. Search volume is country-specific, and so is keyword difficulty. I’ll come back to this because it’s the most underrated opportunity in international SEOInternational SEO is the practice of optimizing a site so search engines understand which countries and/or languages it targets, and serve the right version to each user. It spans URL structure, hreflang, and on-page localization..
• The intent moves. The same string can pull a transactional SERP in one market and an informational one in another, with completely different result sets and competitors.

So you don’t translate a finished list. You run the research again, per market, and translation is at most one input — never the output.

The process

1. Prioritize markets first

You can’t do everything at once, and you shouldn’t. Before any keyword work, decide which markets earn the investment. Weigh three things: traffic potential (is there real demand there?), competition (can you realistically rank, or is the SERP locked up by entrenched local players?), and content investment (how much localized content and ongoing maintenance will this market actually require?).

A fast way to find your starting candidates is to look at where you already get traffic without trying. If you’re already picking up impressions in a market you never targeted, that’s demand asking to be served. (More on the tooling for this below — Metrics by Country and GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. filtered by country.)

2. Research local terminology per market

For each prioritized market, the question is not “what’s the translation” but “what do people there type.” That means idioms, local brand names that have become generic, regional vocabulary, and industry-specific local terms. This is where machine translation quietly fails — it gives you the linguistically correct word, not the one with the searches behind it. Use translation as a starting hypothesis and then validate it against real local search data before you commit.

3. Validate intent in the target market

Even when a keyword translates and has volume, check the live SERP in that market before you build for it. Pull up the actual results the way a local searcher would (see the Tools tab for doing this cleanly). Are the top results local brands or global ones? Is Google showing an informational SERP for what you assumed was a transactional term? Are there local SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. — local packs, regional news, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself. for local entities? An intent mismatch you catch here is far cheaper than one you catch after publishing a market’s worth of content.

4. Analyze the local competitors

This is the step people skip, and it poisons everything downstream. Your global competitors usually are not the ones ranking in a given local market — there are local brands and regional players you’d never think to research. If you run gap analysis against your home-market competitors for, say, German keywords, you get the wrong gap data. Identify the real local competitors first, then seed your research from their visibility in that country.

5. Account for data reliability in smaller markets

Keyword tools estimate volume from clickstream and panel data, and those panels are much thicker in big markets (US, UK, DE) than in small ones. The practical consequence: a keyword can show “0” volume in a smaller market and still have genuine traffic — that’s a data-collection limit, not an absence of searches. Ahrefs is explicit about the edges of this; for example, Burmese is supported only in Rank Tracker because there isn’t enough data for Keywords Explorer yet. Treat a zero in a thin market as “unconfirmed,” not “no demand,” and cross-check it (Google Trends for relative interest, GSC for proof you’re already getting impressions).

6. Map keywords per market — universal vs. local

You have three real options:

• Universal map with a localizationLocalization is adapting content for a specific target market — not just translating the words, but adjusting currency, formats, idioms, cultural references, local search terms, and trust signals so the experience feels native. layer — start from your primary-market keyword map, then audit each keyword for local equivalents and fill gaps with local-only research.
• Market-from-scratch — treat each market as a fresh project seeded from local sources (local sales/support teams, local competitor content, local SERPs).
• Hybrid — build the universal taxonomy first, then run a per-market gap analysis to surface what your structure misses. This is what I’d recommend for most teams.

7. The “search in your target market” method

The single most useful local-discovery habit is to actually search like a local: use Google with the correct country and language settings (or the right country domain), watch autocomplete in the target language, and read the SERP. Autocomplete and “people also search for” are a free window into real phrasing — they surface the local modifiers and colloquialisms a translation never would. Pair that with local forums, Q&A sites, and what your local sales and support teams hear customers say.

A caveat on all of this

None of the above guarantees indexingStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed., rankings, traffic, conversions, or AI citations in a market. Market-specific research improves query fit and lets you prioritize the right terms and markets — it doesn’t promise an outcome. Competition, content quality, technical execution, and ongoing maintenance in that market still decide what actually ranks or gets cited.

What this looks like at enterprise scale

I ran technical SEOTechnical SEO is the practice of making a site easy for search engines to crawl, render, index, and (now) be eligible for AI answers. It's the foundation that lets your content and links rank — not a ranking trick of its own. at IBM — a company operating in something like 170+ countries with a web presence big enough that people only half-joked it could rival Wikipedia in volume. At that scale, international keyword research stops being purely an analytical problem and becomes an organizational one.

The recurring issue is keyword ownership. A US product team targets “cloud computing,” the UK team targets its own variant, the German team targets theirs — and without coordination you get regions competing with each other for the same terms, or three teams independently doing (and re-doing) research for the same market. The fix that actually held up was a centralized keyword map with regional variant layers: one source of truth for the taxonomy, with local variants and local-only terms owned by the people in-market. It prevents internal cannibalization and stops the same research being paid for three times. The other enterprise reality is deployment — getting localized targeting and hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. right across multiple CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. platforms is its own headache, and it’s the reason the cleanest keyword map in the world still needs someone who can ship it.

The opportunity hiding in keyword difficulty

Here’s the part worth internalizing: KD is market-specific, and the gap is the opportunity. Ahrefs calculates KD from the backlink profiles of the pages currently ranking in that specific market. So a keyword that’s KD 70 in the US — because strong, well-linked US sites hold the top results — might be KD 30 in Brazil, or in another English market like Australia, where the ranking pages are weaker. A term you can’t realistically compete for at home can be genuinely attainable somewhere else. That’s not an edge case; it’s one of the main reasons international SEOInternational SEO is the practice of optimizing a site so search engines understand which countries and/or languages it targets, and serve the right version to each user. It spans URL structure, hreflang, and on-page localization. is worth doing. Always pull KD per market and look specifically for the markets where a valuable term is soft.

The Ahrefs workflow, report by report

This is the toolset I actually use for international keyword work. Each report is the country-level version of something you already know:

Keywords Explorer — country selector. Pick the target location from the dropdown (Keywords Explorer covers 230+ locations) and everything recalculates for that market: volume, KD, Matching terms (your seed term as it appears in local searches — the local modifiers and qualifiers), Related terms (the local terminology cluster around a concept), and per-country search suggestions. Set the country to the UK and “insurance” surfaces “buildings insurance,” “contents insurance,” “motor insurance” — UK terms, not the US “homeowners insurance” / “auto insurance.” That’s local terminology discovery without speaking the language.

Site Explorer → Organic Keywords → filter by country. Drop a local competitor’s domain in, open Organic Keywords, and filter to the target country. Now you see exactly what that competitor ranks for in that market — which you can cross-reference against your own rankings to find what you’re missing. This is also how you read a single domain’s market-by-market footprint.

Organic Competitors — per country. Run it filtered to the target market to find out who the real local SERP competitors are. These are frequently not your global competitors — they’re local brands and regional players you wouldn’t have known to research. This report seeds steps 3 and 4 above.

Metrics by Country. See where a domain’s traffic comes from across markets. Point it at your own site to spot which markets are already strong vs. untapped (untapped = demand you’re catching by accident, which is a great prioritization signal). Point it at a competitor to see which markets they’ve built up that you haven’t.

Content Gap (Site Explorer). Enter your local-market competitors and find the keywords they rank for that you don’t — filtered to the target country. Use 3–5 genuine local competitors (not global brands) and you’ll surface terms domestic players rank for that often have no clean English equivalent at all. This is the single most productive report for finding the keywords you’d never have thought to search for.

The throughline: do all of this per market. The same domain, the same seed term, and the same report tell you a different story in each country — and that difference is the entire point.

AI summary

A condensed take on the Advanced version:

• International keyword research = research each market directly, not translate your home-market list. Translation is at most one input.
• Three things move when you cross a border: the words (vocabulary diverges even within one language; literal translations often aren’t the search term), the numbers (volume and KD are market-specific), and the intent (the same string can pull different SERPs).
• The process: prioritize markets (traffic potential, competition, content cost) → research local terminology → validate intent in the live local SERP → find the local competitors → map keywords per market (universal + localizationLocalization is adapting content for a specific target market — not just translating the words, but adjusting currency, formats, idioms, cultural references, local search terms, and trust signals so the experience feels native. layer is the usual best).
• KD is the underrated opportunity: it’s computed from the pages ranking in that market, so a keyword locked up at home can be wide open abroad. Pull KD per market and hunt for soft markets.
• Watch thin data: small markets have sparse panel data, so “0 volume” can be a collection limit, not absence of demand. Cross-check with Google Trends and GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results..
• No guarantees: better research improves query fit and prioritization — it doesn’t guarantee indexingStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed., rankings, traffic, conversions, or AI citationsAn AI citation is the visible source link an AI answer engine shows next to its generated text — the clickable reference that credits the web page it used. A citation's presence is a separate thing from whether the cited page actually supports the statement, and from being retrieved (read behind the scenes) or merely mentioned (named without a link); citation is driven more by brand mentions and being retrievable than by traditional ranking.. Competition and execution in that market still decide the outcome.
• Ahrefs spine: Keywords Explorer country selector (volume/KD/matching/related per country), Site Explorer Organic Keywords filtered by country, Organic Competitors per country, Metrics by Country (untapped vs. strong markets), Content Gap against local competitors.
• Enterprise reality (IBM-scale): keyword research becomes an organizational problem — centralize the taxonomy, layer regional variants, and assign ownership to prevent internal cannibalization and duplicated work.

Official documentation

The search engines document site structureWebsite structure (site architecture) is a site's visible hierarchy, navigation, breadcrumbs, and URL organization — how pages relate and how people and search engines move between them. Internal linking is the primary signal Google reads to understand that structure, not URL folders. and targeting more than keyword methodology, but several pieces are directly relevant to international keyword strategy.

Google

• Managing multi-regional and multilingual sites — the geotargetingConfiguring a site or URL to target users in a specific country. signals Google uses (ccTLDsCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal., hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others., server location, local addresses/phone numbers, language, currency) and, critically for keyword work, the fact that Google reads the visible words on the page to determine language — not lang attributes or the URL. Get the local terms wrong and Google can infer the wrong market.
• International SEO overview — when and how to tell Google about language/region variations so it serves the right version per locale.
• Search Console — International Targeting (deprecated) — the dedicated report was retired (September 2022), but the Performance reportThe Google Search Console report that shows how your site actually performed in Google Search, built from real impressions and clicks. It reports four metrics — clicks, impressions, average CTR, and average position — and keeps the most recent 16 months of data. still filters by country, which is the natural starting point for finding markets where you already get impressions without targeting.

Bing / Microsoft

• Bing Webmaster Tools — Keyword Research — Bing’s built-in keyword tool offers country- and language-specific volume and competition data. Worth checking where Bing holds meaningful market share before dismissing its data.
• Bing also supports hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. for region/language targeting and sets geographic targeting in the site settings.

A note beyond Google and Bing: in some markets the dominant engine isn’t Google at all — Baidu in China, Naver in South Korea, Yandex historically in Russia, Seznam in the Czech Republic. Volume from Ahrefs or Google Keyword Planner won’t reflect demand in a Baidu-dominated market; confirm which engine your target audience actually uses first.

Quotes worth citing

Statements from practitioners and primary sources on why translation isn’t keyword research. Where a deep link to the exact passage is available, it jumps there.

On vocabulary diverging within the same language

• “most people in the US would use the word ‘vacation’ to describe taking a break from work to travel, whereas the more commonly used term in the UK is ‘holiday.’” — TransPerfect Travel. Jump to quote
• “Terminology varies from country to country, even if the language remains the same.” — Jamie Grant & Despina Gavoyannis, Ahrefs Blog. Jump to quote

On why translation fails

• “directly translating your keywords to the target language isn’t a complete international strategy.” — Elizabeth Pokorny, Weglot. Jump to quote
• “content may be linguistically correct but not contain a more popular phrase.” — Motoko Hunt, SEO consultant, in Search Engine Land. Jump to quote
• “each must adopt the target country’s spelling, commonly used words and writing style.” — Motoko Hunt, Search Engine Land. Jump to quote

On doing research before translating

• “Conducting keyword research before translation helps your team incorporate key terms from the start.” — Christina Spaulding, Search Engine Land. Jump to quote

The international keyword research framework

A repeatable, market-by-market process. Run it once per prioritized market — the inputs change every time even when the seed term doesn’t.

1. Prioritize markets. Score candidates on traffic potential, competition, and content investment. Surface candidates from where you already get impressions you never targeted (Metrics by Country, GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. by country).

2. Identify local terms per market. For each market, find how people actually search — idioms, regional vocabulary, local brand-driven terms, regulatory terms. Translation is a hypothesis to validate, not the answer.

3. Validate SERP intent in market. Pull the live local SERP. Confirm intent matches your content, note whether local or global brands dominate, and look for local SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic.. Kill or re-scope terms where the intent doesn’t match.

4. Analyze local competitors. Identify the real local-market competitors (not your global ones) via Organic Competitors per country, then seed your gap analysis from their local visibility.

5. Map keywords to existing or new pages. Decide universal vs. local, then assign each keyword to a page. For multi-team orgs, centralize the map and layer regional variants with clear ownership.

Decision tree: translate or localize a keyword?

Start: you have a keyword from your home market.
│
├─ Does a direct translation exist?
│   ├─ No  → LOCALIZE. Research the local term from scratch
│   │        (local SERP, autocomplete, local competitors, native speaker).
│   └─ Yes → continue ↓
│
├─ Does the translated term have real search volume in-market?
│   (check per-country volume; cross-check Trends/GSC if the tool shows ~0)
│   ├─ No  → LOCALIZE. The "correct" word isn't the searched word —
│   │        find what locals actually type.
│   └─ Yes → continue ↓
│
├─ Does the local SERP for it match your intent?
│   ├─ No  → RE-SCOPE. Right word, wrong intent — find the term that
│   │        pulls the SERP you can serve.
│   └─ Yes → continue ↓
│
└─ TRANSLATE (use the term) — but still confirm spelling/variant for the
   specific market (color/colour, -ise/-ize) and validate with a native speaker.

Default bias: when in doubt, localize. Translation is the exception that has to earn its place, not the starting assumption.

International keyword research checklist

Market selection

• Listed candidate markets and scored each on traffic potential, competition, and content investment.
• Checked where you already get impressions you never targeted (Metrics by Country / GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. Performance filtered by country).
• Confirmed which search engine dominates each candidate market (not always Google — Baidu, Naver, Yandex, Seznam).
• Decided how many markets you can realistically resource right now (don’t boil the ocean).

Keyword discovery (per market)

• Set Keywords Explorer to the target country and language before reading any metric.
• Pulled Matching terms and Related terms in-market for local modifiers and terminology clusters.
• Identified 3–5 genuine local competitors (not global brands) via Organic Competitors per country.
• Ran Content Gap against those local competitors, filtered to the country.
• Watched local autocomplete / “people also search for” in the target language.
• Pulled seed terms from local sources (sales/support, forums, Q&A, local competitor content).

Validation (per keyword)

• Confirmed per-country volume and KD — not home-market figures.
• Checked the live local SERP for intent match and local vs. global brand dominance.
• Cross-checked any “0 volume” against Google Trends and GSC before writing the market off.
• Had a native speaker / localizationLocalization is adapting content for a specific target market — not just translating the words, but adjusting currency, formats, idioms, cultural references, local search terms, and trust signals so the experience feels native. reviewer sanity-check the final list for cultural fit.
• Mapped each keyword to a page and (for multi-team orgs) assigned ownership to prevent regional cannibalization.

Tools for international keyword research

Ahrefs Keywords Explorer (primary). The country selector is the workhorse — pick from 230+ locations and every metric recalculates for that market: search volume, KD (which is computed from the pages ranking in that country, so it varies by market), Matching terms, Related terms, and search suggestions, all local. It’s how you do real local-terminology discovery in a language you may not speak. Pair it with Site Explorer (Organic Keywords filtered by country, Organic Competitors per country, Metrics by Country) and Content Gap against local competitors — covered in detail on the Advanced tab.

Google Keyword Planner (with country targeting). A useful second source, especially in smaller markets — its volume comes straight from Google Search data, so it’s a good cross-check when a third-party tool shows a suspicious zero. Set the country/language target explicitly.

Google Trends (by country/region). Trends gives relative interest, not absolute volume, which makes it perfect for two jobs: comparing demand between markets, and confirming a term has genuine interest even when a keyword tool reports near-zero volume in a thin market.

Google Search from the target country. The “search like a local” method — use Google with the correct country and language settings (or the country-specific domain, e.g. google.co.uk, google.com.au), optionally via VPN, to see the real local SERP and harvest autocomplete suggestions in the target language. Free and high-signal.

DeepL (for terminology discovery, not just translation). Use it to understand how locals phrase something, not to generate your final keywords — read its alternatives and context, then validate the candidates against real search data. Treat all machine translation as a hypothesis, never the answer.

Native speaker review. The last and non-negotiable step for any serious market. Data tells you what has volume; a native speaker tells you whether it reads naturally, carries the right connotation, and matches how people actually talk. Engage one to validate the final list for cultural fit before you build.

Changelog