International SEO Checklist

• Rank target markets by demand, competition, and revenue potential.
• For each market, list the languages actually spoken — not just the official one. Minority-Language SEO is often the less contested entry.
• Decide whether you serve by language, by country, or both; they are different targeting models.
• Set a realistic order: one market at a time beats five half-built ones.
• Confirm you can support the market operationally — payments, support, legal, or delivery.

Skipping this stage is the most common failure: sites build hreflang for markets they cannot serve. The operational question comes first — structured pages for a country where you cannot accept payment are waste that drags site quality.

Before structure, settle the model; it determines everything downstream. The two are not interchangeable:

• Language targeting serves speakers of a language wherever they are — useful for content, tools, and products with no local presence.
• Country targeting serves a geographic market, with local pricing, currency, and legal differences — useful for e-commerce and services.

Many sites need both, which is why a code can be de (language) or de-AT (language and region). The mistake is picking a model implicitly, then finding the URL structure cannot express it.

Every added market multiplies content, support, legal, and measurement work. It is the wrong move when:

• The home market is not yet working — translation inherits the same weaknesses in a market where you have less context.
• You cannot support the buyer — no local payment method, support hours, or returns process produces refunds and churn.
• Demand is too thin to justify a market — serve it with one well-localized page rather than a full language section.
• The content depends on local trust signals you do not have — reviews and credentials that persuade at home may not exist in the new market.

The decision to add a market should be as evidence-driven as any keyword decision; the SEO targeting guide covers weighing opportunity against effort.

• Pick one structure and commit:
  • ccTLD (example.de) — strongest country signal, most expensive, hardest to scale.
  • Subdomain (de.example.com) — clear separation, moderate authority consolidation.
  • Subdirectory (example.com/de/) — easiest to scale, consolidates authority, the usual default.
• Keep one structure site-wide; mixing them complicates hreflang and reporting.
• Use language and/or country codes consistently (/de/ vs /de-de/).
• Avoid automatic redirects based on IP or browser language — let users and crawlers choose.
• Ensure each market has a distinct, crawlable URL.

The trade-off is authority versus clarity. Subdirectories pool link equity into one domain, which is why most sites default to them; ccTLDs give the strongest signal but split authority. Consistency matters more than the choice: mixed structures make hreflang sets hard to maintain. The redirect check is the one teams most often fail — IP-based redirects hide versions from crawlers and frustrate travelling or VPN users. Suggest a version with a dismissible banner instead.

• Add hreflang annotations for every language/country version of a page.
• Include self-referencing hreflang on each page.
• Make annotations reciprocal — A points to B, B points to A.
• Use x-default for the fallback page.
• Keep hreflang values valid (ISO language, optional ISO region).
• Keep the set identical across equivalent pages; a mismatch breaks the cluster.
• Verify implementation with a testing tool or the search console's international targeting report.

This is the stage most often done wrong; a broken cluster can suppress all versions. Three failure modes to memorise: a missing return link (A lists B, B omits A), a missing self-reference, and an inconsistent set (one page lists three markets, its sibling four). hreflang is a relationship, not a property, so any asymmetry degrades the group. Validate after every market launch and template change; the broader indexing mechanics live in the technical SEO audit checklist.

• Each market page self-canonicalizes.
• Do not canonical all languages to the primary-language page — that suppresses them.
• Confirm robots and sitemaps expose every market's URLs.
• Include all market variants in sitemaps (or use sitemap index files per market).
• Check for accidental duplicate content across near-identical translations.

hreflang and canonical answer different questions: hreflang says "same content, another audience," canonical says "this is the URL to index." They can conflict. The safe configuration is that each language version self-canonicalizes and uses hreflang to relate to its siblings. Pointing every market's canonical at the primary language overrides hreflang and tells search engines to drop the localized versions.

• Rewrite titles, descriptions, and headings for local search behavior — not literal translation.
• Redo Keyword Research per market; direct translations often miss what locals search.
• Adapt examples, currency, units, dates, and formats.
• Localize proof: reviews, case studies, and trust signals for that market (E-E-A-T Standards).
• Avoid mass machine translation of thin pages — low-quality translated pages drag down site quality.
• Give each language its own content plan, not a mirrored copy of the source language.

Localization decides whether a market feels native or like a translated brochure. The keyword layer is the clearest test: a literal translation of your best head term may describe a concept locals never search for, while the phrase they use sits in a keyword research pass you have not run. Treat each market's keyword map as a first-class asset and review it with a daily user of the language.

The vault's Canva analysis is the reference model: every language gets a designed landing page, its own URL, keyword strategy, and tone.

• Ensure the CDN or hosting serves each market with acceptable latency.
• Confirm fonts, scripts, and RTL layouts render correctly.
• Keep pages fast on the devices common in each market.
• Verify structured data and metadata localize correctly.
• Test in-market, not just from headquarters.

Latency affects experience a head-office test cannot reveal. A page fast near your data center may crawl on slow networks far from your servers, and mobile-first indexing means that market's mobile experience is what gets evaluated. RTL scripts and different character sets are rendering bugs, not cosmetic ones, when a broken layout hides content.

• Separate reporting per market (properties or views).
• Track impressions and clicks by country and language.
• Diagnose which market is underperforming and why — structure, content, or demand.
• Watch for one market cannibalizing another's impressions.
• Feed local findings back into each market's keyword map.

Blended global numbers hide everything useful. Segment first by market, then diagnose with a simple rule: impressions without clicks points at a title or relevance problem, clicks without conversions at an offer or trust problem, and no impressions at all at a structure or indexing problem.

• Redirecting by IP, which hides versions from crawlers and users.
• Translating instead of localizing, producing pages nobody searches for.
• Broken hreflang reciprocity, which fragments the cluster.
• Canonicalizing all languages to one URL, which suppresses the others.
• Launching ten markets at once, none of which get enough attention.
• Ignoring minority languages, leaving uncontested demand on the table.
• Mixing URL structures across markets, which breaks hreflang and reporting.
• Treating the second market as a one-time project rather than an ongoing plan.

• Multilingual SEO · Global Expansion SEO · International SEO
• Web Globalization · Minority-Language SEO · Independent Site Globalization
• AI Global Site Expansion · Keyword Research · E-E-A-T Standards

• Google Spam Policies for Web Search (official, updated 2026-08-28) — Google spam policies target scaled content abuse and doorway pages; genuine local pages are unaffected. (accessed 2026-09-09)

Going global multiplies every decision. The XKnow SEO Vault collects the globalization frameworks, hreflang guides, and per-market keyword assets behind this checklist in Markdown and Obsidian, so expansion follows a system instead of guesswork. Use the free checklist; run the full playbook from the Vault →.