01International SEO
International SEO: a market is not a language, and both need their own answer
International SEO at Searlixa covers market and language targeting, hreflang architecture, localisation quality and locale-level AI visibility. The most common failure we see is a site that treated forty markets as a translation problem.
Quick answer
International SEO covers market and language targeting, hreflang architecture, localisation quality and locale-level AI visibility.
The distinction that fixes most problems: a market is not a language. Spanish for Mexico and Spanish for Spain share a language and almost nothing else commercially.
Key takeaways
- AI visibility barely transfers between markets. Roughly 21% of citations carry across on the most consistent engine.
- Prompt sets must be built in-market, in the market’s language. This is the largest cost driver in international work.
- Architecture is chosen against your resourcing, not a blog post’s preference. Subdirectories usually win on consolidated authority.
- Broken hreflang is common and quietly expensive.
- Machine translation at scale reads as foreign and rarely matches how people actually search.
- Script and RTL testing is a build requirement, not a content task. Allow 30 to 40 percent copy expansion.
Definition
International SEO is the practice of making the right version of a site visible to the right market and language, without the versions competing against each other. It combines URL architecture, hreflang signalling, content localisation and per-market measurement.
The distinction that fixes most problems
A market is a place with its own competitors, regulations, price expectations, buying norms and search behaviour. A language is how people phrase things. Spanish for Mexico and Spanish for Spain share a language and share almost nothing else commercially.
Sites that conflate the two end up with one Spanish page competing with itself across eight countries, or eight near-identical pages that differ only in currency symbol. Both underperform, and the second carries duplication risk.
What international SEO work covers
- Architecture decision. ccTLD, subdirectory or subdomain, chosen against your actual resourcing rather than a blog post's preference. Subdirectories usually win on consolidated authority; ccTLDs win where local trust or legal presence matters.
- Hreflang implementation. Correct return tags, self-referencing annotations, x-default handling. Broken hreflang is common and quietly expensive.
- Transcreation, not translation. Machine translation at scale produces pages that read as foreign to the market and rarely match how people actually search.
- Local keyword research per market. Done in-market, not translated from an English list.
- Locale-correct formatting. Dates, numbers, currency, addresses, phone formats. Small signals that carry disproportionate trust weight.
- Script and RTL testing. Arabic, Devanagari and CJK break layouts built only against Latin text. Allow 30–40% copy expansion in components.
AI visibility varies by market too
This is the part most international programmes have not caught up with. Assistant behaviour, the sources they draw on and even which assistants dominate all differ by country. A brand can be well cited in the United States and invisible in the UAE for the same query family, because the source landscape the model draws on is different.
That means a prompt set per market, in the market's language, sampled separately. It is not a translation exercise, which is also why it is the largest cost driver in international engagements. See pricing for how that scales.
| Market | Primary consideration | Common mistake |
|---|---|---|
| United States | Highest competition; AI answer penetration is furthest along | Treating it as the default all other markets copy |
| United Kingdom | Shares a language with the US, not a market | One English page serving both, ranking well in neither |
| UAE | Bilingual search behaviour; RTL layout requirements | Arabic added as a translation layer rather than a market |
| Australia | Distinct local competitor set despite shared language | Assuming US or UK content transfers unchanged |
| India | Multilingual, highly mobile, price-sensitive comparison behaviour | English-only targeting in a multilingual search market |
→What you are buying
International SEO services: what is included
International SEO agency work is decided before it is built. Getting the architecture wrong is the expensive mistake, and it is usually made in week one.
- Market and language architecture decided before anything is built
- Hreflang implemented and verified, not assumed
- Transcreation rather than machine translation, with local terminology researched
- Per-market technical review: indexation, canonicals, local hosting and performance
- Prompt sets rebuilt per market, because a US set measures nothing in the UK
- Reporting split by market, never averaged into one number
Who this is for
Businesses selling into more than one country, or planning to, where the same content is currently doing duty for every market.
Who it is not for
Businesses in one market that want to look international. That is a positioning exercise, not a search one.
From $2,730 a month. Two markets at this tier, up to five at Enterprise, each with its own prompt set rather than a translated one. SEO and AI visibility pricing is published in full.
The audit covers every market you name, separately. An average across markets hides the one that is broken. The $99 SEO and AI visibility audit comes back in three working days, and the fee comes off your first invoice if you continue.
Sources
Every figure on this page comes from one of the following. Where a claim is ours rather than a source’s, it carries an evidence label instead.
- Rankability — State of AI Search, 1,645 AI citation observations across 31 topics, 2026. Source for the 24.1% cross-platform overlap, the 73.5% ChatGPT exclusive citation rate, the 22.1% vs 8.7% first-100-words finding and the 43.3% vs 15.5% query-breadth finding. Independent study
- Google Search Central — Optimizing your website for generative AI features (published 15 May 2026, updated 10 July 2026). Source for the llms.txt, chunking and AI-schema positions, and for the statement that generative features run on core ranking systems. Primary
- Searlixa sampling across client engagements. Any figure drawn from this carries an Observed label with its sample size and date. Our own
Ask an assistant
Hand this page to an assistant and ask it to summarise or challenge the argument. If it gets something wrong about us, that is itself a useful finding.
04Common questions
Questions about multi-market work
Is a market the same as a language?
No, and conflating them is the most expensive mistake in international work. Spanish for Mexico and Spanish for Spain share a language and almost nothing commercially — different competitors, different sources, different buying norms.
How much AI visibility carries between markets?
About 21% on ChatGPT, the most globally consistent engine measured in 2026 research. Google AI Mode scored 14.6%. Roughly four fifths of your citation base does not transfer.
Should we use subdirectories, subdomains or ccTLDs?
Subdirectories usually win on consolidated authority and are cheapest to maintain. ccTLDs give the strongest local signal and cost the most. We choose against your resourcing, not against a blog post's preference.
Can we translate our existing content?
You can, and it will underperform. Machine translation at scale reads as foreign and rarely matches how people actually search. Transcreation — rewriting for the market — is the minimum that works.
What breaks technically when we add a language?
Layout, most often. Non-Latin scripts and RTL break components built only against English. Allow 30 to 40 percent copy expansion and test before the design system is signed off.
What is the largest cost driver?
Language scope, by a wide margin. Each additional language means a prompt set built in that language by someone working in the market, then sampled and coded separately.
How do we know if hreflang is working?
Check for reciprocal tags, self-referencing tags and correct region codes. Broken hreflang is common, quiet and expensive — it fails without any visible error.
Which market should we add first?
Wherever you already have revenue, not wherever is largest. Sequencing by market size and translating outward assumes a citation transfer that does not happen.