How to Build a Multilingual SEO Strategy

On this page

The strategic decisions come before any hreflang tag, and getting them right is most of the work. Multilingual SEO is a sequence of investment choices: which languages to build for, whether to target a language or a specific country, subfolder versus ccTLD, and how far to localize versus translate. Make those well and the implementation is mechanical. Make them on autopilot, defaulting to country-specific ccTLDs and word-for-word translation, and you scatter authority across zero-authority domains and publish content that misses how the market actually searches. One well-built Spanish asset can serve Spain, Mexico, and Argentina and concentrate authority; for most expanding companies, the right URL structure is subdirectories that inherit the main domain’s authority, not separate ccTLDs that start from zero.

This post owns the upstream strategy. The strategic choices below determine the hreflang you will later implement; the tag mechanics, self-reference, return tags, x-default, and language-region codes, are a separate implementation topic and are not taught here. Where country variants like es-ES versus es-MX come up, it is only as a strategy question, whether you need country variants at all, not as a coding tutorial.

Multilingual versus international: the distinction that drives everything

A multilingual site offers content in more than one language. A multi-regional (international) site explicitly targets users in different countries. The two are not the same, and conflating them is the first strategic error.

Targeting a language means one asset serves everyone who speaks it. A single Spanish version can reach Spain, Mexico, Argentina, and Colombia, and all the links, engagement, and authority earned by that asset concentrate in one place. Targeting a country means a separate asset per market, es-ES, es-MX, es-AR, each tuned to local pricing, currency, regulation, and search behavior, but each accumulating authority separately and competing for resources.

The decision is a trade-off between authority concentration and regional precision. Do you actually need to distinguish Mexican from Argentine Spanish, or does one well-localized Spanish asset serve the demand? Most companies start by targeting languages, because one strong asset beats several weak ones, and split into country variants only where a real market difference, pricing, legal requirements, materially different search patterns, justifies the cost of dividing the authority.

Localize versus translate

Translation and localization are not the same thing, and the right amount of localization varies by content type. Pure translation, word-for-word conversion, misses local search patterns: people in different markets phrase the same need differently, and a literal translation of your English keywords will not match what they type. At the other extreme, full re-creation of every page from scratch wastes the leverage of your existing structure and content.

Localization is the middle path, and usually the right one: keep the English page’s structure, do fresh keyword research in the target language, and culturally adapt where it matters while translating directly where it does not. A practical split:

  • Translate directly: product descriptions, technical specifications, and other content where the meaning is stable across markets.
  • Localize fully: anything where local context changes the substance, case studies and social proof, pricing and currency, examples, and compliance or legal content. These need local research and adaptation, not conversion.

The keyword research point is the one teams skip. Local search behavior is its own input; you discover the actual queries in the target language and build the localized page around them, rather than translating the keywords you already rank for and hoping they map.

AI translation: drafts, not publishing

Machine translation is good enough now to produce a usable first draft and to make localization faster. It is not good enough to publish unreviewed, especially on trust-sensitive pages. For anything that affects a purchase, a legal or compliance statement, or your credibility, a native speaker has to review before publication, because AI translation still produces phrasing that is technically correct but reads as foreign, misses idiom, and occasionally gets nuance wrong in ways that erode trust. Use AI to accelerate the draft; never let it be the final pass on pages where trust is the conversion.

Prioritize languages ruthlessly

You cannot build every language at once, and you should not try. SEO expands existing demand; it rarely creates demand from nothing. So prioritize languages by evidence of real opportunity, not by ambition:

  • Existing impressions from a market. Check Search Console for impressions your English content is already earning from a given country or language. Visibility you are getting without trying is the strongest signal that demand exists there.
  • Competitor investment. If competitors have built out a language, there is usually a market behind it; if none have, be skeptical there is easy demand to capture.
  • Existing sales traction. Markets where you already make sales, or get inbound interest, are markets where SEO investment compounds on real demand rather than gambling on creating it.

The sequencing discipline is to prove ROI on one language before scaling to the next. Build the highest-signal language fully, measure the return, and use that result to justify the next. This beats launching six half-built languages simultaneously, none of which gets the depth to rank.

URL architecture: subdirectory by default

The structural decision comes down to three options, and for most expanding companies the answer is the first.

Subdirectory (example.com/es/) is the default. All variants live on one domain, so every language inherits the main domain’s accumulated authority, and any links and authority the language versions earn flow back into the same domain. It is also the simplest to maintain. For a company expanding into new languages without a physical local presence in those markets, this authority inheritance is almost always the deciding advantage.

Subdomain (es.example.com) sits in the middle: cleaner separation than a subdirectory, but authority does not consolidate as cleanly across subdomains, so you give up some of the inheritance benefit without gaining a strong geographic signal.

Country-code TLD (example.es) sends the strongest geographic signal to users and search engines that the site is meant for that country, but each ccTLD is effectively a separate domain that builds authority from zero. Choose a ccTLD only when there are real reasons: an established local presence or brand in that country, regulatory requirements, or a genuine need for the strongest possible country signal. For an expanding company without local presence, the authority you forfeit by starting a fresh domain typically outweighs the geographic-signal gain.

The trade-off in one line: ccTLDs maximize the geographic signal, subdirectories maximize authority inheritance, and for most expansions authority inheritance wins. Note also that Google’s Search Console country-targeting (the International Targeting report) has been deprecated and removed, so you cannot lean on a Search Console setting to assign a generic domain to a country; the geographic signal now comes from your URL structure, hreflang, and on-page localization, which is another reason the structure decision matters.

Frequently Asked Questions

Should I use a ccTLD for each country to rank better locally?

Usually not, unless you have a real local presence, brand, or regulatory reason in that country. A ccTLD sends a strong country signal but starts with zero authority, while a subdirectory on your main domain inherits the authority you have already built. For an expanding company without local operations, the subdirectory’s authority inheritance generally outweighs the ccTLD’s geographic signal.

Do I need separate Spanish versions for Spain and Mexico?

Only if the markets differ in ways that change the content, pricing, currency, legal requirements, or materially different search behavior. If one well-localized Spanish asset serves the demand, build that first and concentrate the authority. Split into country variants when a concrete market difference justifies dividing the authority and the maintenance cost.

Sources