Why international SEO decides everything
Your business could be invisible to more than 130 million Arabic searchers — not because your product is wrong, but because Google is serving them the wrong version of your site, or no version at all. In Saudi Arabia, 99% of people are online, around 96% of them search on Google, and roughly 95% search in Arabic. Across the Gulf and Egypt, buyers are ready, connected and spending. The painful truth in 2026 is this: a flawless English page — or even a hastily “translated” Arabic one — can be perfectly indexed and still never be seen, because Google’s AI now decides which single answer to show, and it rarely picks a page that merely echoes English in another language.
That is the gap international and multilingual SEO closes. It is no longer about being on the map; it is about being the version Google chooses to put in front of the customer who is already looking for you. For a brand based in Cairo trying to win Riyadh, or a Gulf store trying to reach both Saudi and Egyptian buyers, the difference between “we have an Arabic page” and “we own Arabic search in our category” is the difference between a hobby and a market.
I have lived on both sides of this. I have watched perfectly good businesses leak their entire growth potential because Google was confused about who they were for — one ccTLD competing with its own subfolder, hreflang tags pointing in one direction only, an Arabic page that read like it was run through a machine. And I have rebuilt those same sites into the versions Google trusts, in the languages and countries that actually convert. This page is the playbook: what international SEO really is, how I engineer it, and the verified results it produces.
International vs multilingual vs local — the distinction that costs people money
Most Egypt and Gulf buyers I speak to use “international SEO”, “multilingual SEO” and “local SEO” interchangeably. They are three different jobs, and conflating them is the first expensive mistake.
International SEO (also called multi-regional SEO) optimises your site so search engines serve the correct country version to each user — the Saudi store to the Saudi shopper, the Egyptian catalogue to the Cairo buyer. Multilingual SEO is the language layer of that work — Arabic and English versions of the same content. Local SEO is something else entirely: it targets a single geographic area, one city or country, usually through Google Business Profile, maps, reviews and local citations.
The trap is that these overlap but are not the same. A business selling across Egypt, Saudi Arabia and the UAE needs international and multilingual SEO — country targeting and language targeting working together. It does not just need “local SEO done three times.” And critically, the same product is searched differently in Cairo, Riyadh, Jeddah and Dubai — not only in different dialects, but with different intent, price expectations, payment methods and seasonality. International SEO is the discipline of mapping all of that correctly before a single page is built.
| Discipline | What it optimises | Typical use | Core tools |
|---|---|---|---|
| International / multi-regional SEO | The right country version reaches the right user | A store selling in KSA, Egypt and UAE | hreflang, site architecture, Search Console country targeting |
| Multilingual SEO | The right language version reaches the right user | Arabic + English versions of one site | hreflang language codes, localisation, RTL |
| Local SEO | Visibility within a single area | One clinic or shop in Riyadh | Google Business Profile, maps, reviews, citations |
What Google and the experts actually say
I do not ask you to take my word for any of this. The technical backbone of international SEO comes straight from Google’s own documentation, and the strategic shifts are echoed by the most credible voices in the field.
Google is explicit about how to structure language versions. It wants a distinct, crawlable URL for each language — not content that quietly changes based on a cookie or a browser setting.
Google recommends using different URLs for each language version of a page rather than using cookies or browser settings to adjust the content language on the page.
Google is just as firm about what not to do — and this is one of the most common, most damaging mistakes I see on Gulf and Egyptian sites: forced redirects based on assumed language or IP address.
Avoid automatically redirecting users from one language version of a site to a different language version of a site. For example, don’t redirect based on what you think the user’s language may be.
Those two rules alone — distinct URLs, no assumption-based redirects — solve a startling share of international SEO failures. But the bigger 2026 story is how AI search has changed the stakes. Motoko Hunt, one of the most respected international SEO specialists in the industry, frames the shift precisely:
A market-specific page can be indexed, valid, and hreflang-correct and still never appear in AI Overviews or AI Mode. Visibility is now a selection problem, not a ranking problem.
That single sentence reframes the entire discipline, and I will return to it. For now, hold onto the consensus from the people who do this at the highest level: get the technical foundation perfect, then go far beyond it. As Hunt also notes, “translated pages that add no new intent, authority, or context are rarely retrieved. The most confident version of a concept — often English — wins globally.” The leading SEO teams agree. Ahrefs’s own growth lead, Erik Sarissky, puts the localisation principle bluntly: “If you’re just translating everything, you’re wasting time.”
Site architecture: ccTLD vs subdomain vs subfolder
Before a word is written or a tag is added, one strategic decision shapes everything that follows: how you structure your international URLs. Get this wrong and you fight your own site for years. There are three options, and there is no universal “best” — only the right fit for your budget, your number of markets, and your existing authority.
Country-code top-level domains (ccTLDs) — example.sa, example.eg, example.ae — send the strongest possible geo-targeting signal. They scream “this site is for this country.” But each domain starts its authority from zero. You are effectively building, and earning links for, a brand-new site per market. Powerful for a well-funded brand committed to a handful of countries; punishing for a lean business spread across many.
Subdirectories (subfolders) — example.com/ar-sa/, example.com/ar-eg/ — keep every page on one domain, so all your authority compounds in one place. They are the cheapest to run, the easiest to manage, and let you set country targeting inside Google Search Console. For most businesses in our region, this is the recommended default, and it is what I steer the majority of clients toward.
Subdomains — sa.example.com — sit awkwardly in the middle. Google often treats them as separate sites, so link equity transfers inconsistently. They can make sense for genuinely distinct platforms or technical constraints, but they rarely beat a clean subfolder for SEO.
| Structure | Geo signal | Authority | Cost & effort | Best for |
|---|---|---|---|---|
| ccTLD (.sa, .eg, .ae) | Strongest | Built per domain, from zero | Highest | Funded brands, few committed markets |
| Subdirectory (/ar-sa/) | Moderate, set in GSC | Concentrated on one domain | Lowest | Most businesses — recommended default |
| Subdomain (sa.example.com) | Moderate | Inconsistent transfer | Medium | Distinct platforms or technical constraints |
Hreflang, done right — or not at all
Hreflang is the annotation that tells Google which language and region each URL targets, so it can serve ar-SA to a Saudi searcher and ar-EG to an Egyptian one instead of letting them compete or cannibalise each other. It is the heart of multilingual SEO — and it is also where most international sites quietly fall apart. Industry audits consistently find that a large share of international sites carry hreflang errors serious enough to fragment their rankings.
Hreflang is unforgiving because it has to be technically perfect to work at all. Three rules are non-negotiable:
- Self-referencing. Every page must include an hreflang tag pointing to itself, alongside tags for every other variant. Miss the self-reference and the whole cluster can be ignored.
- Reciprocal and symmetric. If page A points to page B, page B must point back to page A. One-directional hreflang — the most common error — is routinely discarded by Google. The return tags must match.
- Valid codes. Use ISO 639-1 language codes, optionally paired with ISO 3166-1 country codes:
ar-SA,ar-EG,en-US,en-SA. Invalid or invented codes simply fail silently.
And one more that protects you from chaos: an x-default annotation pointing to the version a user should see when no language or region matches. It is your safety net for the rest of the world.
| Hreflang rule | What it means | What happens if you skip it |
|---|---|---|
| Self-referencing tag | Each URL lists itself among the variants | The entire cluster may be ignored |
| Reciprocal / return tags | A→B requires B→A | Google discards the one-way annotations |
| Valid ISO codes | ar-SA, ar-EG, en-US only | Tag fails silently; wrong version ranks |
x-default | Fallback for unmatched users | International visitors land on the wrong page |
| Correct placement | In <head>, HTTP header, or sitemap — consistently | Mixed signals confuse crawlers |
When hreflang is right, the payoff is clean: each market sees its intended page, duplicate-content risk between near-identical language versions drops, and your Saudi page stops competing with your Egyptian one in the same SERP. When it is wrong, you get the worst outcome — the wrong country’s page ranking in the wrong country, or none ranking at all. After I deploy or repair hreflang, I verify it in Google Search Console’s International Targeting and performance reports two to four weeks later, because the proof has to be in your own dashboard, not my promise.
Translation is not localization
If you remember one thing from this page, make it this: translation is a step, localisation is a system. Word-for-word machine translation — the kind produced by auto-translate plugins — gives you pages that are technically in Arabic but read like a foreigner wrote them, miss the terms people actually search, and convert at a fraction of their potential. Backlinko’s Leigh McKenzie says it plainly: “Relying only on machine translations for your content is risky. At best, it can lead to awkward phrasing.” Ahrefs’s Mateusz Makosiewicz frames the discipline itself: “Translation is a step. Localization is a system.”
Localisation means rebuilding the page for the market, not just the language. It starts with native-speaker keyword research — discovering how Saudis versus Egyptians actually phrase a query, which synonyms and dialect words they use, and what intent sits behind each search. It extends to cultural nuance: examples, references, tone and imagery that feel native rather than imported. And it reaches into the commercial details that decide conversions — local currency (SAR vs EGP), payment methods (Mada, Tabby, Tamara in the Gulf), city-level intent (Riyadh, Jeddah, Cairo), and seasonality.
The business case is not soft. In a 2024 DeepL survey, 96% of marketers reported positive ROI from localisation efforts, and 65% saw at least 3x return. Localisation is one of the rare SEO investments where the numbers are this consistent — because you are matching real demand in the language it is expressed, instead of hoping a translated page is “close enough.”
If you’re just translating everything, you’re wasting time.
Arabic, dialects and RTL — the layer generic agencies skip
Arabic is where international SEO gets genuinely specialised, and where most global guides go quiet. There is no single “Arabic” to optimise for. There is Modern Standard Arabic (MSA) — formal, pan-regional, the language of official content — and there are dialects that differ sharply between the Gulf and Egypt in everyday speech and, crucially, in how people type their searches. A Saudi shopper and an Egyptian shopper looking for the same product will often use different words. Choosing the wrong register, or blending dialects on a single page, sends a muddled signal to both Google and the reader.
Then there is the technical reality of the script itself. Arabic is right-to-left (RTL), which affects layout, navigation, mixed Arabic-English strings, and the way numbers and brand terms render. Brand names frequently need careful transliteration decisions — do you keep the Latin spelling, write it in Arabic, or support both because people search both ways? These are not cosmetic choices; they determine whether your page matches the query at all.
And the prize for getting it right is enormous. Roughly 95% of all searches in Saudi Arabia are conducted in Arabic, and Google holds about 96% of the Saudi search market. An English-only site — however polished — is invisible to the overwhelming majority of that demand. Egypt, with around 96 million internet users at nearly 82% penetration, is the largest online audience in the Arab world. This is not a niche to dabble in; it is the core of the regional opportunity, and it rewards operators who treat Arabic as a first-class language rather than a translation afterthought.
| Factor | Wrong way (generic) | Right way (localised) |
|---|---|---|
| Keyword choice | One Arabic translation for all markets | Native research per market — Gulf vs Egyptian vs MSA |
| Dialect | Mixing registers on one page | Consistent register matched to the audience |
| Layout | LTR template with Arabic pasted in | True RTL design, mixed-string handling tested |
| Brand terms | Latin only, or one Arabic spelling | Deliberate transliteration, both spellings supported |
| Targeting | One “Arabic” version for everyone | ar-SA and ar-EG as distinct, intentful pages |
AI Overviews, semantic collapse and GEO
This is the headline change of 2026, and it rewrites the rules of international SEO. Google’s AI Overviews and AI Mode now appear on a large share of searches — roughly 48% of Google searches show an AI Overview as of March 2026, up from about 34.5% in December 2025 — and they are available in 200+ countries and 40+ languages. International markets are not adjacent to the AI shift; they are inside it.
What makes this so consequential for multilingual sites is how AI handles language. Instead of treating your Arabic page and your English page as separate documents to be ranked, AI often collapses multilingual content into shared semantic representations — it understands the meaning across languages and pulls the original facts (frequently English) to synthesise an answer in another language in real time. The result is exactly what Motoko Hunt warned about: a page can be indexed, valid and hreflang-correct, and still never be cited. “Translated pages that add no new intent, authority, or context are rarely retrieved,” she notes. “The most confident version of a concept — often English — wins globally.”
The click stakes are real. When an AI Overview appears, Ahrefs measured a 34.5% drop in position-one click-through rate for informational keywords, and Pew Research Center found click-through nearly halving — 8% with an AI Overview versus 15% without, a 46.7% fall. If your local page is correct but invisible, you are not losing a ranking; you are losing the customer to whichever source AI chose instead.
So what wins? Generative Engine Optimization (GEO) — also called answer-engine optimisation — is now a core part of international SEO. The work is to become the version AI selects and cites: genuine, in-market authority; opinionated, expert content written by people who actually know the market, not auto-translations; clean structured data; and a real brand presence in each country and language. Hreflang still earns you eligibility and controls duplication. GEO earns you the citation.
Translated pages that add no new intent, authority, or context are rarely retrieved. The most confident version of a concept – often English – wins globally.
This is also where local authority and E-E-A-T per market become decisive. Links, mentions and citations from in-market, in-language sources strengthen the regional version of your site and improve your odds of being the page AI cites. Generic global backlinks do not transfer that geo-relevance — a strong link from a US publication does little to convince Google you are the authority in Riyadh. Building real, local credibility, market by market, is the work that AI rewards. As Bauer Media’s Stuart Forrest put it on adapting to AI search: “We need to make sure that it’s us being cited and not our rivals. Things like writing good quality content… it’s amazing the number of publishers that just give up on that.”
My international SEO process
I do not bolt a translation plugin onto your site and call it international SEO. I run a deliberate, market-by-market process designed to validate demand first, build the technical foundation flawlessly, and then earn the local authority that AI and classic search both reward.
- Validate demand, market by market. Before localising anything, I confirm there is real, financially viable search demand in each target country — and map how intent differs between Cairo, Riyadh, Jeddah and Dubai. The same product is rarely searched the same way twice. No business should pay to localise a page nobody in that market is searching for.
- Choose the architecture. ccTLD, subdomain or subfolder — decided against your budget, your markets and your existing authority, not a template. For most regional businesses this means a clean subfolder strategy with Search Console country targeting.
- Engineer the technical layer. Distinct crawlable URLs per language, perfect reciprocal hreflang with valid ISO codes and
x-default, no assumption-based redirects, and a structure Google can read — exactly to Google’s own guidance. - Localise, don’t translate. Native-speaker keyword research per market, culturally adapted content, correct RTL handling, and local commercial details — currency, payment methods, city intent. Real Arabic, written for Saudis and Egyptians as distinct audiences.
- Build local authority and GEO. In-market, in-language links and mentions; structured data; and distinctive, expert content so your page is the one AI selects and cites — not just one that is technically valid.
- Measure what matters, then scale. Start with one market, get it right, then expand. I check Search Console’s International Targeting two to four weeks after hreflang deployment, and track per-country organic visibility, AI Overview citations and conversions — not vanity global rankings. Every number comes from your own dashboards.
The results international SEO unlocks
International and multilingual SEO is rarely the flashy headline — but when it is done properly, it is the work that opens entire markets. These are verified results from my own projects, every number checkable in Google Search Console, Moz and Semrush.
- Roseberry (Saudi Arabia) climbed from roughly 25 impressions a day to 51.5M impressions and 545K clicks, ranking for 2,855 keywords over 16 months. Winning the Saudi market in Arabic — with the right architecture, intent and authority — is what made that scale possible.
- A niche store was driven to #1 in Saudi Arabia in 166 days, by treating the Saudi audience as its own market rather than an afterthought translated from elsewhere.
- Conscent grew from 61K to 1.2M impressions in six months once the content and structure matched real local demand.
- Oxford (Egypt) reached 70.6K impressions, capturing Egyptian search intent in its own right rather than borrowing a generic global page.
The pattern across all of them is the same. We did not “add Arabic.” We won a market — by validating demand, building an architecture Google trusts, localising for how people there actually search, and earning the local authority that makes a page the one chosen, not just one that is correct.
If your business is selling into Egypt, Saudi Arabia or the Gulf — or you are an English-speaking decision-maker who needs to reach Arabic audiences — the opportunity is enormous and the field is wide open. Most competitors are still relying on auto-translate plugins, broken hreflang and English-only thinking, and almost none have adapted to the 2026 AI-search reality. That gap is exactly where I build advantage.
Let’s find the markets you are missing, fix the version of your site Google should be trusting, and make you the answer it puts in front of the customer already searching for you.