How to Do International SEO: A Friendly, Honest Guide

Here’s the short version of how to do international SEO: expand only where you already see real demand, pick one URL structure (country-code domains, subdirectories, or subdomains) and stick with it, use hreflang to point search engines at the right language or regional version of each page, and localize your content properly instead of just translating it. Then measure each market separately so you know what’s actually working. That’s the whole system — the rest of this guide walks you through each piece so you can do it without the expensive mistakes.

Okay, let’s be honest for a second: international SEO has a reputation for being scary, and most of that fear comes from one word — hreflang. I promise it’s more tedious than hard. The genuinely risky part isn’t the technical setup at all. It’s expanding into markets that were never going to buy from you, with pages that read like a robot wrote them. So we’re going to cover the strategy first, the plumbing second, and I’ll flag every common trap along the way.

Should you actually go international yet?

Here’s the part nobody tells you: the most common international SEO mistake happens before a single hreflang tag is written. It’s expanding on hope instead of evidence. Translating a site into twelve languages feels like growth, but every language version you launch is a site you now have to maintain, update, and keep technically healthy — forever. Thin, neglected international sections can drag down the perceived quality of your whole site.

So before you build anything, look for real demand signals you already have:

• Existing international traffic. Open your analytics and Search Console, filter by country, and see who’s already finding you. If 8% of your organic traffic comes from Mexico and you’ve done nothing to earn it, that’s a signal worth taking seriously.
• International sales or signups. Orders shipping abroad, customers paying in your currency despite the friction, support tickets written in another language — these are people buying in spite of the experience you offer them. Imagine what happens when you actually meet them halfway.
• Search demand you can verify. Use a keyword tool set to the target market and check whether people there search for what you sell — in their own words, which (spoiler) are rarely a literal translation of yours.
• Operational readiness. Can you actually serve that market? Shipping, support hours, payment methods, legal requirements? SEO can bring visitors, but it can’t fix a checkout that rejects their cards.

If one or two markets light up across those signals, start there. One market done well beats six done badly, every single time. International SEO rewards depth, not breadth.

How to do international SEO: choosing your URL structure

This is the first big structural decision, and it’s genuinely hard to change later — so let’s get it right the first time. You have three main options for where your international content lives, and anyone who tells you one of them is always correct is selling something.

Structure	Example	Strengths	Trade-offs
ccTLDs (country-code domains)	yourbrand.de, yourbrand.fr	Strongest geo-targeting signal; local users often trust a local domain; clean separation per country	Each domain builds authority from scratch; you buy, manage, and defend multiple domains; most expensive to maintain
Subdirectories (subfolders)	yourbrand.com/de/, yourbrand.com/fr/	All markets share one domain’s authority; one site to maintain; cheapest to launch and scale	Weaker inherent country signal (you lean on hreflang and Search Console settings); one technical problem can affect every market
Subdomains	de.yourbrand.com, fr.yourbrand.com	Easier to host regionally or run on separate infrastructure; some organizational separation	Search engines may treat them as partially separate sites, so authority sharing is less predictable than subdirectories

My honest take: for most small and mid-sized teams, subdirectories are the pragmatic choice. You concentrate all your link equity on one domain, you maintain one site, and you can add markets without buying new domains. ccTLDs make sense when local trust is a major purchase factor, when legal or logistical reasons require separate entities per country, or when you’re a large brand with the budget to build authority in each market independently. Subdomains are usually chosen for infrastructure reasons more than SEO ones.

Whatever you pick, pick once. Migrating from one structure to another is a full-site migration with all the risk that implies. And if you’re not sure your overall technical foundation is solid enough to build on, run through a technical SEO audit before you multiply your site by three languages — every crawl issue you have today gets copied into every market you launch.

What does hreflang actually do (in plain words)?

Hreflang is a tag that tells search engines, “this page has sister versions in other languages or regions — here’s the full list, so please show each searcher the right one.” That’s it. If someone in Spain searches and you have a Spanish page, hreflang helps Google serve the Spanish version instead of the English one, even if the English page has more authority.

Each page in a language group lists every version of itself, including itself. So your English pricing page says “the English version is here, the German version is here, the French version is here” — and your German and French pricing pages each say the exact same thing. That self-reference and mutual listing is where most implementations fall apart.

The classic hreflang mistakes, so you can skip the suffering:

• Missing return links. This is the big one. If page A points to page B, page B must point back to page A. One-way references get ignored. Every page in the group lists every page in the group — no exceptions.
• Wrong codes. Language codes follow ISO 639-1 and country codes follow ISO 3166-1 Alpha 2, and people mix them up constantly. The code for the United Kingdom is gb, not “uk.” There’s no valid “en-eu” because the EU isn’t a country. And the format is always language first, region second: “en-gb,” never “gb-en.”
• No x-default. The x-default value tells search engines which version to show when no other version matches the searcher — usually your global or English page, or a language selector. Without it, mismatched searchers get whatever Google guesses.
• Pointing at broken or redirected URLs. Every hreflang URL should return a clean 200 status and be the canonical version of that page. Tags pointing at redirects or 404s quietly stop working.

And here’s the thing that surprises people: hreflang is a hint, not a command. Search engines take it seriously, but they can and do override it when other signals disagree. If your “German” page is 80% English text, no tag will convince Google it’s the German version. Hreflang works best when it confirms what the content already makes obvious.

Language targeting vs. country targeting — which one do you need?

This distinction trips up almost everyone at first, so let’s make it simple. You can target a language (hreflang=”es” — Spanish speakers anywhere in the world) or a language plus a country (hreflang=”es-mx” — Spanish speakers in Mexico specifically). You cannot target a country alone; the language part is required.

When does the difference matter? Use plain language targeting when one version genuinely serves everyone who speaks that language — a blog in Spanish that works equally well for readers in Madrid, Mexico City, and Buenos Aires. Use language-plus-country when the markets actually need different content: different prices and currencies, different shipping policies, different legal terms, meaningfully different vocabulary. Spanish is a great example — a “computer” is commonly an ordenador in Spain and a computadora in much of Latin America. If your product pages mention currency, delivery, or local regulations, country-specific versions earn their keep. If they don’t, one well-written language version with hreflang=”es” is less work and less duplication.

A sane pattern many sites use: one general language version (es) as the default for that language, plus country versions (es-mx, es-ar) only for the markets where you truly operate differently. Don’t create es-cl, es-pe, and es-co pages that are word-for-word identical just to have them — that’s duplication with no user benefit, and we’ll talk about why that hurts in a minute.

Why localize instead of just translating?

Translation converts your words into another language. Localization converts your meaning into another market. The difference is the whole ballgame, and it starts with keywords.

Never literal-translate your keyword list. Search behavior differs between markets even when the language is technically the same. The word your translator chooses may be grammatically perfect and still not be the word people actually type into a search box. Do fresh keyword research in each target market, in the local language, using a tool set to that country — and ideally with a native speaker interpreting the results, because they’ll catch the colloquial term your tool surfaces that a textbook translation would never find. This is the same discipline you’d apply to any new content project: research what real people search for, then write for that, not for what you assumed.

Beyond keywords, real localization touches everything your visitor sees:

• Currency, units, and formats. Prices in the local currency, measurements in the local system, dates and phone formats that look native. Nothing says “we didn’t really build this for you” like a price in the wrong currency.
• Examples and references. Case studies, brand mentions, holidays, and cultural touchpoints should make sense locally. A Thanksgiving marketing example means little in markets that don’t celebrate it.
• Imagery and social proof. Testimonials from customers in that market, images that reflect the audience, trust signals that resonate locally.
• Native review, always. Before anything goes live, a native speaker of that market’s variant reads it. Not a fluent second-language speaker — a native. They catch tone, idiom, and the subtle wrongness that makes readers quietly lose trust.

This is also where cultural respect stops being a nice-to-have and becomes a ranking factor in disguise. Content that genuinely serves local readers earns engagement, links, and return visits. Content that’s obviously translated filler earns a back button. Search engines notice both.

Is Google the only search engine you need to think about?

Mostly, but not universally — and it’s worth being honest about this rather than pretending Google is the whole world. In many markets Google dominates search, and optimizing for it covers you. But some markets have significant local players: other search engines hold meaningful share in countries like South Korea, Russia, China, and the Czech Republic, among others. Market share shifts over time, so verify the current landscape for your specific target market before you invest — a quick check of recent, reputable market-share data for that country should be part of your market-entry research.

The good news: the fundamentals travel well. Crawlable architecture, fast pages, genuinely localized content, and clean technical health help you everywhere. If a local engine matters in your market, it usually has its own webmaster tools and documentation — register your site there just as you would with Google Search Console, and read its guidelines, because indexing requirements and ranking emphases can differ.

How do you avoid duplicate content across same-language regions?

Here’s a scenario that confuses a lot of smart people: you have a US page, a UK page, and an Australian page, all in English, all nearly identical except currency and a few spellings. Is that a duplicate content disaster?

Not if you set it up correctly — and the answer is canonicals and hreflang working together. Each regional version should have a self-referencing canonical tag (the US page canonicals to itself, the UK page to itself), and the hreflang group ties them together as intentional alternates (en-us, en-gb, en-au). That combination tells search engines, “yes, these pages are similar on purpose; they serve different audiences; show each audience its own version.” Handled this way, same-language regional pages are a normal, supported pattern.

What you must not do is canonical all the regional versions to one “main” page. That tells search engines to index only that one version, which quietly cancels your hreflang and removes the regional pages from their local results — the opposite of what you built them for. Canonical says “index this one”; hreflang says “serve the right one to each audience.” They only work when they agree.

And a friendly challenge: if your regional pages are so identical that this section made you nervous, ask whether they should exist at all. The strongest version of this pattern is pages that differ in ways users actually feel — pricing, availability, shipping, local examples. If the only difference is the hreflang tag, you’ve created maintenance burden without user value.

How do you measure international SEO market by market?

A single global traffic number will lie to you. Growth in your home market can completely mask a failing launch in a new one, and vice versa. So from day one, measure each market separately:

• Search Console, filtered by country. The performance report lets you segment clicks, impressions, and average position by the searcher’s country. This is your core view: is Germany actually seeing and clicking the German pages?
• Page-level views per market section. Filter by URL path (/de/, /fr/) to see how each language section performs as a unit — which pages index, which earn impressions, which convert impressions into clicks.
• Your own baselines, captured before launch. Record what traffic, rankings, and conversions from each target country look like before you launch its localized version. That’s your honest comparison point. Without it, you can’t tell whether the new section created demand or just absorbed traffic you were already getting.
• Conversions, not just clicks. Rankings in a market mean little if visitors don’t become customers. Track signups, orders, or leads per market, because that’s the number that justifies (or kills) the next market’s budget.

Set a realistic time horizon, too. New international sections start with little history and few local links; give a market several months of consistent effort before judging it, and judge it against its own baseline rather than against your mature home market.

One more measurement note: watch indexing per section in Search Console’s coverage reports. International builds multiply your page count fast, and if a whole language section isn’t getting indexed, that’s an architecture or quality problem to fix before you worry about rankings. A clear structure — shallow paths, strong internal links between hub and regional pages — matters even more at international scale, and the principles in our guide to improving site architecture for SEO apply doubly when every template exists in four languages.

What are the most common international SEO failure modes?

Let me save you from the greatest hits of international SEO pain, because every one of these is common and every one is avoidable:

• Auto-redirecting visitors by IP address. This feels helpful and it’s actually hostile. It traps travelers and expats on the “wrong” version, it can prevent search engine crawlers from ever seeing your regional pages, and it overrides user choice. The better pattern: let everyone access every version, show a polite, dismissible banner suggesting the local version, and remember their choice. Suggest; never force.
• Machine-translated pages published raw. Automated translation has gotten impressive, but publishing it unreviewed at scale is a quality gamble. Unedited machine output reads wrong to natives, erodes trust, and thin translated pages at scale can look like exactly the low-value content search engines work to filter out. If you use machine translation, treat it as a first draft that a native editor finishes — and if you can’t afford that review for a market, you can’t afford that market yet.
• Flag icons for language selection. Flags represent countries, not languages. Spanish is spoken in over twenty countries; Portuguese speakers in Brazil shouldn’t have to click Portugal’s flag. Use language names written in their own language — Español, Deutsch, Français — in your selector. Small detail, big respect.
• Slow pages for distant visitors. Your site might feel fast from your office and sluggish from the other side of the world. A CDN that serves assets from locations near your international visitors is one of the simplest wins in international SEO — speed is a user-experience factor everywhere, and distance is a real tax on it.
• Launching wide and maintaining nothing. Six language versions that go stale within a year serve no one. Match your footprint to your maintenance capacity.

If you’re running an online store across markets, the stakes on several of these — currency handling, regional duplication, faceted URLs multiplied by language — get even higher, and our walkthrough of ecommerce SEO covers the product-page side of that puzzle in depth.

Your market-entry checklist

When you’re ready to enter a new market, work through this list in order. It’s deliberately demand-first:

1. Verify demand: existing traffic, sales, or inquiries from the market, plus keyword research confirming local search volume for what you offer.
2. Confirm you can serve it: shipping or service delivery, payments, support, and any legal requirements.
3. Choose scope: language-only targeting or language-plus-country, based on whether the market needs different prices, policies, or terms.
4. Confirm URL structure: your chosen pattern (for most teams, a subdirectory like /de/) and how this market fits it.
5. Research local keywords natively: fresh research in the local language and variant, never a literal translation of your English list.
6. Localize the launch set: start with your highest-value pages fully localized — currency, examples, imagery, legal bits — rather than the whole site shallowly.
7. Get native review: a native speaker of that market’s variant signs off before launch.
8. Implement hreflang: add the new versions to every relevant hreflang group, with return links and an x-default.
9. Capture baselines: record pre-launch traffic, rankings, and conversions from that country.
10. Register and monitor: Search Console (and any relevant local engine’s tools), indexing checks, and a recurring content maintenance plan.

The hreflang sanity checklist

Before and after every launch, run these checks — they catch the vast majority of hreflang problems:

• Every page in a group references every version, including itself.
• Every reference has a matching return link from the other page.
• Codes are valid: language first (ISO 639-1), country second (ISO 3166-1 Alpha 2) — en-gb, not en-uk, not gb-en.
• An x-default exists and points to your global page or language selector.
• Every URL in the tags returns 200 — no redirects, no 404s.
• Hreflang URLs match the canonical URL of each page, and every regional page canonicals to itself.
• Tags live in one consistent place (HTML head, HTTP headers, or XML sitemap — pick one method and audit it).
• Crawl the site with an SEO crawler after launch and re-check all of the above, because templates drift.

None of this guarantees a ranking — nothing does, and anyone promising otherwise should make you nervous. What it guarantees is that you’ve removed the self-inflicted problems, which honestly is most of the battle in international SEO.

One last practical note, because content rarely travels alone: if you’re entering new markets, your social presence usually goes with you — different time zones, different peak hours, sometimes different platforms entirely. Having one calendar where you can schedule posts per market and per audience, instead of juggling time-zone math at midnight, keeps the promotion side of your international launch as organized as the SEO side.

FAQ: how to do international SEO

Do I need hreflang if my site is only in one language?

If you have only one version of each page, no — hreflang exists to disambiguate between multiple language or regional versions. The moment you add a second language version, or a second regional version in the same language (like US and UK pages), hreflang becomes important so search engines serve each audience the right one.

Are subdirectories or ccTLDs better for international SEO?

There’s no universal winner. Subdirectories (yourbrand.com/de/) are the pragmatic choice for most teams because every market shares one domain’s authority and you maintain a single site. ccTLDs (yourbrand.de) send the strongest local signal and can earn more local trust, but each domain builds authority from scratch and costs more to maintain. Choose based on your budget, how much local trust matters in your industry, and how many markets you realistically plan to support.

Is it okay to use machine translation for international pages?

As a first draft, yes; as a publishing strategy, no. Unedited machine translation reads unnatural to native speakers and, published thin and at scale, risks being treated as low-quality content. Have a native speaker edit and localize every page before it goes live — including re-checking keywords, since the words people search differ from textbook translations.

Should I automatically redirect visitors to their country’s version?

No. Forced IP-based redirects frustrate travelers, override user choice, and can prevent crawlers from accessing your regional versions. Instead, let every visitor reach every version, show a dismissible banner suggesting their local version, and remember what they choose.

How long does international SEO take to show results?

There’s no honest universal number — it depends on the market’s competitiveness, your site’s existing authority, and how well you localize. New language sections start with little history and few local links, so plan in months, not weeks. Capture baseline metrics before launch, measure each market separately in Search Console, and judge progress against that market’s own starting point rather than your home market’s performance.