How to audit international SEO without missing the basics

An international SEO audit is not just a hunt for hreflang errors. It is a way to prove whether a website’s search problem is genuinely international, meaning the wrong country or language pages are being shown, or whether the site has a more ordinary SEO problem such as weak indexing, thin content, poor internal linking or unclear page purpose.

That distinction matters because international setups add complexity very quickly. A site can have perfect country folders and still fail because the pages are barely localised. Equally, a site can have untidy hreflang and still perform acceptably if market targeting is otherwise clear. A good international SEO audit starts by asking what each section is for, who it serves, and what evidence shows search engines are understanding that intention.

What an international SEO audit is actually trying to prove

The first job is to define the decision the audit supports. Usually it is one of these:

• whether we need a true international architecture at all
• whether existing country or language sections are being interpreted correctly
• whether duplicate or near-duplicate pages are intentional localisation or accidental repetition
• whether visibility problems are caused by international targeting signals, or by broader SEO issues

That sounds basic, but it prevents a common mistake. People see several countries, several languages, or several domains and assume the international layer must be the problem. Often it is not.

If a UK business has one English site and a few pages mentioning overseas delivery, that is not automatically a multi-country SEO setup. If a company sells in Ireland and the UK with the same prices, same fulfilment, same legal terms and same audience language, separate sections may not be needed. On the other hand, if the site serves French speakers in Belgium and France with different products, currencies, delivery promises and regulatory wording, then the audit needs to test whether search engines can distinguish those versions.

So before opening any crawler or report, list the site’s intended markets in plain English:

• Which countries do we actively target?
• Which languages do we publish in?
• Are country and language paired, or independent?
• Which pages are meant to rank in which market?
• Which pages are global and should not be localised?

That list becomes the benchmark. Without it, you cannot tell whether a “duplicate” is a problem or a planned variation.

This is also where we separate search evidence from business assumptions. A stakeholder may say, “our German pages are for Germany”, but the page itself may use UK spelling, GBP pricing and a UK returns address. Search engines and users can only judge what is actually published.

If you need a simpler starting point before going international, our guide on checking a site before you buy an SEO audit helps frame the underlying diagnosis.

Start with market targeting and page purpose

Once we know the intended markets, we check whether each section has a clear purpose. In practice, that means reviewing a sample of pages from each country or language area and asking three questions.

First, what market is this page obviously for?

Look for direct signals:

• language of the main content
• country references in headings and body copy
• local currency
• delivery and returns information
• local phone numbers and addresses
• region-specific product availability
• legal or tax wording where relevant

For UK readers, this point often gets muddled with “EU targeting”. The EU is not a single country target in search terms. If a site says it serves “Europe” but all pages use only English, only GBP, and only a UK contact address, that is weak evidence of local relevance outside the UK. If there are separate pages for France, Germany and Spain, those should normally show genuine signs of serving those markets, not just translated headings.

Second, is the page genuinely localised or merely copied?

Strategic duplication can be fine. If we have an English page for the UK and another for the US, some overlap is expected. What matters is whether the version differences are meaningful. We would usually treat these as useful localisation signals:

• spelling and terminology adapted for the market
• pricing in local currency
• local shipping terms and delivery expectations
• local case studies or examples
• local contact details or service areas
• market-specific product constraints, compliance notes or FAQs

By contrast, changing only the country name in the title tag while leaving the rest untouched is usually a sign that the page exists for search mechanics rather than user need.

Third, is any duplication accidental?

This is where folder structures, CMS behaviour and faceted URLs create trouble. We often find:

• the same English page published under /uk/, /en-gb/ and parameterised URLs
• translated pages that still canonicalise to the source language
• multiple country pages generated from one template with no meaningful differences
• blog or resource content copied into every regional subfolder without a market reason

At this stage, we are not fixing technical tags yet. We are deciding whether the content model makes sense. If the page purpose is unclear, technical cleanup alone will not solve the problem.

A practical way to do this is to build a sheet with columns for URL, intended country, intended language, observed language, observed currency, local contact details, unique local content, and probable page type. After reviewing a modest sample, patterns usually appear very quickly.

Check crawling, indexing and international signals separately

This is the part people often compress into one vague technical bucket. It is better to separate it into five checks: crawling, indexing, canonicals, hreflang and sitemaps. They interact, but they are not the same thing.

1. Crawling

Crawling asks whether bots can fetch the page. Start with robots.txt, meta robots and any bot-specific rules.

Check whether important country or language folders are blocked in robots.txt. Also look for image, JavaScript or CSS blocking that might affect rendering. Google’s own guidance is clear that useful, accessible content and descriptive linking support discovery, but access still comes first in practice (Google SEO Starter Guide).

If the site is thinking about AI crawler controls, keep categories separate. Google Search access is not controlled through the same mechanism as model training access. OAI-SearchBot and GPTBot are not the same thing either. OpenAI documents search access and training controls separately (OpenAI crawler documentation). So in an international audit, note any bot rules, but do not confuse them with country targeting.

2. Indexing

Indexing asks whether the URL is eligible and actually included. Use Google Search Console to review page status by section, not just sitewide totals. The Page indexing report helps explain why URLs are excluded, and Google explicitly notes that not every excluded URL should be indexed (Google Page indexing report).

For international sections, compare:

• discovered but not indexed pages in each market folder
• alternate pages with canonical tags
• duplicate without user-selected canonical
• crawled but not indexed pages
• soft 404 patterns in thin regional pages

This is often where you discover the real issue. For example, if only one language version is indexed and the others are excluded as duplicates, the problem may be content similarity or canonical handling, not hreflang.

If you want a broader framework for evidence handling, our article on Search Console evidence and its limits is useful background.

3. Canonicals

Canonicals express preferred indexing versions. In international setups, a common error is cross-canonicalisation between market variants. If the French page canonicalises to the English page, you are effectively telling Google the French page is not the preferred version.

Audit canonical tags for:

• self-referencing canonicals on valid local pages
• accidental canonicals to another market or language
• parameter URLs canonicalising to clean URLs
• consistency between HTML canonicals and any HTTP header canonicals

Canonicals are not a language targeting system. They are a consolidation signal. If two pages are intended to coexist for different audiences, they usually should not canonicalise to each other.

4. Hreflang

Hreflang helps search engines understand alternate language and regional versions. It does not force indexing, and it does not fix weak content. Audit it on that basis.

Check:

• valid language and region codes
• reciprocal return links between alternates
• inclusion of self-references
• consistent alternate sets across all versions
• whether x-default is used sensibly, if at all

Also check whether the alternate set matches reality. If a page has hreflang references to six market versions but only three actually exist as useful pages, the annotation is over-claiming.

Where UK and EU practice diverge, the practical issue is usually business localisation rather than a special SEO rule. Search engines do not need a separate “EU” hreflang value because the EU is not a country code in this context. If the site truly targets multiple EU countries, those versions still need country or language logic that matches the published experience.

5. Sitemaps

XML sitemaps are discovery aids, not a substitute for internal linking or indexability. In an international audit, check whether:

• only canonical, indexable URLs are included
• market sections have complete coverage
• hreflang sitemap annotations, if used, match page-level annotations
• orphan local pages appear in sitemaps but nowhere in navigation

When sitemaps contain large numbers of blocked, redirected or canonicalised-away URLs, they become noisy evidence rather than helpful evidence.

Review templates, navigation and on-page localisation

After the technical signals, review what users actually see. This is where many international problems become obvious.

Start with templates. If all country pages share one template, inspect whether the template itself supports localisation or undermines it. Common issues include:

• a global header that always links users back to the wrong country section
• forms that submit to a single generic market page
• footer addresses showing only the head office regardless of region
• trust signals or delivery notes that contradict the page’s market

Then review navigation. Internal links tell both users and crawlers how the site is organised. If the UK homepage links prominently to UK categories while the German section is buried behind a country selector and receives almost no internal links, that weakens discoverability. Google’s guidance on descriptive links is relevant here too, because navigation labels help explain destination purpose (Google SEO Starter Guide).

On-page localisation checks should be practical and visible. We usually review:

• title tags and meta descriptions for market clarity
• H1 and subheadings for language consistency
• currency symbols and checkout wording
• units of measure, such as cm versus inches, kg versus lb
• date formats and phone number formats
• local addresses, service areas and opening details
• delivery, returns and tax information
• image text embedded in the wrong language
• breadcrumbs and related links pointing to the correct section

This is also where we assess user-facing cues. Can someone land on the page and immediately tell whether it is for them? If not, search engines may have the same problem.

Accessibility overlaps with this. Clear controls, understandable navigation and readable design help users interpret market intent, especially where language selectors or country selectors are involved (W3C accessibility design tips). That does not make accessibility a country-targeting signal by itself, but poor usability often hides localisation choices.

Be cautious with machine-translated or lightly edited pages. The question is not whether a page was translated with software. The question is whether the published result is clear, accurate and fit for the target audience. If category pages, support articles or policy pages read unnaturally, users may bounce back to search, and editors may misdiagnose that as a technical international issue.

For online shops, many of these checks overlap with product and category architecture. Our guide on auditing an online shop without missing the basics covers the underlying retail checks that often sit underneath an international problem.

Decide what to fix first and what can wait

By this point, an international SEO audit should produce a list of findings. The useful next step is not “fix everything”. It is to sort issues by likely impact on discoverability and user clarity.

Usually, we prioritise in this order.

First, fix anything that prevents crawling or indexing of the intended pages. That includes blocked sections, noindex mistakes, broken canonicals and severe duplication that causes the wrong version to be indexed.

Second, fix contradictions between signals. If hreflang says a page is for Germany, but the canonical points to the UK page and the page itself shows GBP and a UK address, that confusion matters more than polishing metadata.

Third, improve user-visible localisation on important templates and landing pages. If high-value pages are technically indexable but unclear about country or language, rankings alone are not the only concern. Users need to recognise relevance quickly.

Fourth, tidy lower-impact issues such as incomplete x-default usage, minor sitemap noise or small metadata inconsistencies, unless those patterns are widespread enough to suggest a larger template fault.

It also helps to label findings by confidence level.

Established guidance includes things like:

• blocked pages cannot be crawled
• canonical tags influence preferred indexing
• hreflang is for alternate language or regional versions
• not every excluded page in Google Search Console is an error
• no special AI schema or text file is required for Google’s AI search features (Google AI features)

Interpretation includes things like:

• whether two market pages are too similar to justify separate indexing
• whether a country selector is clear enough for users
• whether a local page is strong enough to deserve its own search presence

That distinction keeps the audit honest. Some conclusions can be proven directly from source code or reports. Others require judgement.

Finally, connect each finding to the next check. If one market section is underperforming, ask whether the issue is indexing, weak localisation or poor internal prominence. If a whole language set is excluded as duplicate, inspect canonicals and content differentiation before touching hreflang. If traffic is flat but impressions are healthy, the issue may be snippet clarity or page relevance rather than international setup.

A good international SEO audit is not a list of tags to validate. It is a structured test of whether the site’s market logic is visible, consistent and indexable. When you separate page purpose, technical access, indexing evidence and localisation quality, the basics stop being easy to miss.

If you want a general method for staying evidence-led before buying software or a service, our piece on how to audit your site before you buy SEO tools is the next sensible read.

The SEO Blog is published by Fleeta Limited in Ipswich. Where practical implementation work is needed, Suffolk Web shares the same owner, but this publication’s job is to help you assess the evidence before you buy the wrong fix.