International SEO goes wrong in two predictable ways: a structure chosen before anyone worked out which markets justify a page, and hreflang annotations that do not reciprocate, which search engines respond to by ignoring the lot. Both are engineering problems, and both are cheaper to settle before you translate anything.
Thanks, that’s with us.
A senior engineer reads every enquiry and will reply as soon as they can.
Book a time that suits you, or send a message on WhatsApp.
15
Market pages we run
1,000+
Clients served
UK
Registered company
4.6/5
Rated on Clutch
Quick answer
Four decisions, taken in order. Which markets justify their own pages at all. What URL structure carries them. How you tell search engines which page serves which market. And how far you go beyond translating the words.
The structure question is usually settled badly because it is taken early and quietly. Country domains signal locality most strongly and split your authority into separate sites, each earning links from scratch. Subfolders keep one domain's authority intact and are the right answer far more often than the industry admits. Subdomains sit awkwardly between the two.
Hreflang is where most implementations fail, and the failure mode is unforgiving. The annotations have to reciprocate: if your UK page claims the German page serves de-DE, the German page must point back. When they disagree, search engines do not pick a winner, they discard the set, and you are left with the duplicate-content problem the annotations existed to solve.
Then localisation, which is not translation. Currency, units, payment methods, addresses and phone formats, local trust signals, and the fact that buyers in different markets search using different words for the same thing. A literal translation of a page that ranks in English frequently targets a phrase nobody in the target market types.
What we do
Eight areas. The audit says which ones apply to you rather than selling you all eight.
Country domains, subdomains or subfolders, decided on your authority, your resources and how distinct the markets really are. Subfolders are right more often than not, because splitting a domain means each new site starts earning authority from nothing.
Annotations that reciprocate, with x-default set deliberately. We verify both sides live after deployment rather than assuming the tags shipped correctly, because a one-sided annotation is discarded and silently achieves nothing.
Search Console settings, canonicals, internal linking and content signals all pointing the same way. Conflicting signals are common and produce the wrong page ranking in the wrong country, which looks like a hreflang fault but is not.
Researching how buyers in the target market actually search, rather than translating your English keyword list. The literal translation of a term that ranks here is frequently a phrase nobody there uses.
Separate pages for the UK, US, Australia and so on, all in English, differing by a few words and competing for the same queries. Often the right answer is one page with proper annotations rather than five near-identical ones.
Where the site is served from, how the CDN is configured, and what the experience is like at distance. A site that is quick in London can be slow in Sydney, and Core Web Vitals are measured where the user is.
Assistants answer differently depending on language and region. Being quotable in one market does not mean being quotable in another, and the entity work behind it has to hold up in each.
Moving from country domains to subfolders, or the reverse, without losing what each has accumulated. This is where the largest irreversible losses happen, and it needs mapping before anything moves.
The gap
The common sequence, and what we do instead.
Not sure whether a market justifies its own page?
That is the first thing the audit answers, and the answer is often no. A thin page for a market with little demand costs maintenance forever and competes with the page that was already ranking there. We would rather tell you to run three markets properly than fifteen badly.
In practice
Six recurring patterns. These describe what we encounter, not named client projects.
A UK page naming a German alternate that never points back, often because the two sites are managed by different teams or plugins. Search engines discard annotations that are not reciprocated, so the implementation achieves precisely nothing while appearing complete.
Separate UK, US, Australian, Canadian and Irish pages differing by spelling and a phone number, splitting links and leaving search engines to choose. Usually better consolidated into one strong page with annotations than maintained as five weak ones.
Pages translated faithfully from English, including the target phrases, so they optimise for terms nobody in that market searches. The page reads perfectly and ranks for nothing, which makes the problem hard to spot from the content alone.
A set of ccTLDs launched years ago, each needing its own links and content, none of which arrived. The combined result is weaker than one domain would have been, and consolidating them back is a careful, mappable job rather than a redirect sweep.
hreflang saying one thing while canonicals, internal links and Search Console targeting say another. The symptom looks like a hreflang bug and the cause is usually elsewhere, which is why these persist for months.
A page per country because the list looked incomplete without them, each needing maintenance forever and several competing with the page that was already serving those visitors adequately in English.
How we work
The structure decision comes before translation, because reversing it afterwards is the expensive part.
Which markets show real demand, what your current structure is, whether your hreflang reciprocates, and whether your targeting signals agree with each other. We check both sides of every annotation rather than trusting the tags on one.
Which markets justify their own pages and which are better served by one page with proper annotations. This usually means recommending fewer markets than you asked about, and we will say which ones we think are not worth it.
Structure, annotations, targeting and templates, built by the people who diagnosed the problem. In-market keyword research before any translation, so pages are written for how buyers there actually search.
Every annotation checked live on both sides after deployment, not merely shipped. Then impressions and positions by country from your own Search Console, so you can see which markets are responding and which are not.
Clear scope · Regular demos · Full handover
Honest fit
Going multi-market multiplies maintenance. It pays when demand exists and you can serve it properly.
Structure options
The decision that is hardest to reverse, so it is worth taking on evidence rather than preference.
| Country domains | Subdomains | Subfolders | |
|---|---|---|---|
| Example | example.de | de.example.com | example.com/de/ |
| Authority | Each site earns its own from scratch | Partly shared, treated inconsistently | One domain's authority serves every market |
| Local signal | Strongest | Moderate | Needs hreflang and targeting to be explicit |
| Cost to run | Highest: separate sites, separate links | Moderate | Lowest |
| Right when | Markets are genuinely separate businesses | Rarely the best of the three | Most companies, most of the time |
| Hardest part | Building authority several times over | Ambiguity about how it is treated | Making the local signal explicit enough |
Sectors
Categories that export naturally, or where buying behaviour differs enough between markets to justify separate pages.
FAQ
What people ask before committing to a structure. If yours is not here, put it in the form.
Still have a question?
Ask us directly. A senior engineer reads every question.
Ask about your projectSubfolders suit most companies, because one domain's authority serves every market instead of each site earning its own from nothing. Country domains give the strongest local signal and suit businesses that are genuinely separate per market. Subdomains are rarely the best of the three.
It tells search engines which page serves which language or region. It fails because the annotations must reciprocate: if your page names an alternate, that alternate must point back. When the two sides disagree, search engines discard the set rather than choosing, so a half-done implementation achieves nothing.
Often not. Two English pages differing by spelling and a phone number tend to split links and compete with each other. Separate pages are justified when pricing, product, regulation or buying behaviour genuinely differ, not merely because the countries do.
Translation alone usually underperforms, because it carries your English keywords across literally. Buyers in another market often use a different phrase for the same thing, so the page reads well and targets a term nobody searches. Research the market first, then write.
Fewer than you expect. Each market needs content kept current, links earned locally and someone able to serve customers there. Three markets done properly beat twelve thin ones, and the audit will say which of yours show demand worth the maintenance.
It names the page shown when no other annotation matches the user's language or region, typically your main or global page. It is optional but worth setting deliberately rather than leaving search engines to infer a fallback from the rest of your set.
Usually consolidation onto one domain with proper annotations, which recovers the authority currently split across several weak sites. It is a careful mapped migration rather than a redirect sweep, and it should be planned before anything moves.
Yes. Assistants answer in the user's language and draw on sources available in that market, so being cited in one does not imply being cited in another. The entity and corroboration work behind it has to hold up per market, not just once in English.
Fixing broken annotations can show movement within weeks because you are removing an obstacle. Establishing a new market takes months, as you are building authority and relevance from a standing start. We report by country so you can see which markets are responding.
It depends mostly on how many markets you run and whether the structure needs changing. Maintaining several markets is a retainer, from £3,000 a month. A structural migration is a project, from £4,000. The free audit establishes how many markets genuinely justify a page before anything is quoted.
Start here
A senior engineer reads every enquiry and comes back with questions, a realistic view of the work, and an honest opinion on which markets are worth the maintenance.
Thanks, that’s with us.
A senior engineer reads every enquiry and will reply as soon as they can.
Book a time that suits you, or send a message on WhatsApp.